Release [run release-20261111T213646Z]

#27
by Battam - opened
This view is limited to 50 files because it contains too many changes. See the raw diff here.
Files changed (50) hide show
  1. CHANGELOG.md +18 -0
  2. README.md +46 -46
  3. README.zh.md +46 -46
  4. campaign_files/2026-10-a-spreadsheets-with-answers-2/packt-textbook-companion-workbooks/README.md +0 -2
  5. campaign_files/2026-10-b-disk-images-3/opentoall-2015-catz/README.md +0 -3
  6. campaign_files/2026-10-b-disk-images-3/opentoall-2015-gone/README.md +0 -3
  7. campaign_files/2026-10-c-repros-2/overset-grid-subcell-repro/README.md +0 -2
  8. campaign_files/2026-10-g-solver-instances-2/or-library-common-due-date/README.md +0 -6
  9. campaign_files/2026-10-g-solver-instances-2/or-library-hdtt/README.md +0 -12
  10. campaign_files/2026-10-g-solver-instances-2/or-library-sch/README.md +0 -4
  11. campaign_files/2026-10-g-solver-instances-2/or-library-wt/README.md +0 -6
  12. campaign_files/2026-10-g-solver-instances-2/or-library/README.md +0 -4
  13. campaign_files/2026-10-g-solver-instances-2/ors-machine-scheduling/README.md +0 -4
  14. campaign_files/2026-10-g-solver-instances-2/pace-2017-treewidth-bonus/README.md +0 -12
  15. campaign_files/2026-10-g-solver-instances-2/pace-2017-treewidth-exact/README.md +0 -22
  16. campaign_files/2026-10-g-solver-instances-2/pace-2024-ocm/README.md +0 -4
  17. campaign_files/2026-10-gt-browser-tests-2/openscad-cmdline-tests/README.md +0 -22
  18. campaign_files/2026-10-gt-browser-tests-2/pacparser-tests/README.md +0 -4
  19. campaign_files/2026-10-hw-eda-tests-2/openroad-src-dpl-test/README.md +0 -4
  20. campaign_files/2026-10-hw-eda-tests-2/openroad-src-gpl-test/README.md +0 -4
  21. campaign_files/2026-10-hw-eda-tests-2/openroad-src-ppl-test/README.md +0 -4
  22. campaign_files/2026-10-hw-eda-tests-2/openroad-src-tap-test/README.md +0 -4
  23. campaign_files/2026-10-m-issue-trackers-2/openldap-technical/README.md +0 -4
  24. campaign_files/2026-10-m-issue-trackers-2/opensips-users/README.md +0 -4
  25. campaign_files/2026-10-m-issue-trackers-2/openssl-users/README.md +0 -4
  26. campaign_files/2026-10-m-issue-trackers-2/openvswitch-ovs-discuss/README.md +0 -4
  27. campaign_files/2026-10-m-issue-trackers-2/oran-sc-gerrit-changes/README.md +0 -4
  28. campaign_files/2026-10-m-issue-trackers-2/osgeo-gdal-dev/README.md +0 -4
  29. campaign_files/2026-10-m-issue-trackers-2/osmocom-gerrit-changes/README.md +0 -4
  30. campaign_files/2026-10-m-issue-trackers-2/osmocom-openbsc/README.md +0 -4
  31. campaign_files/2026-10-m-issue-trackers-2/oxfordbus-passenger-gtfs/README.md +0 -4
  32. campaign_files/2026-10-m-issue-trackers-2/ozlabs-openbmc/README.md +0 -4
  33. campaign_files/2026-10-m-issue-trackers-2/pace-gtfs/README.md +0 -4
  34. campaign_files/2026-10-m-issue-trackers-2/path-gtfs/README.md +0 -4
  35. campaign_files/2026-10-md-media-and-documents/openslr-70/README.md +0 -4
  36. campaign_files/2026-10-md-media-and-documents/openslr-83/README.md +0 -6
  37. campaign_files/2026-10-md-media-and-documents/org-ruby-html-examples-tables/README.md +0 -4
  38. campaign_files/2026-10-md-media-and-documents/ox-hugo/README.md +0 -4
  39. campaign_files/2026-10-md-media-and-documents/pandoc-crossref-m2m-divtable/README.md +0 -4
  40. campaign_files/2026-10-md-media-and-documents/pandoc-ext-abstract-section/README.md +0 -4
  41. campaign_files/2026-10-md-media-and-documents/pandoc-ext-diagram/README.md +0 -4
  42. campaign_files/2026-10-md-media-and-documents/pandoc-ext-list-table/README.md +0 -4
  43. campaign_files/2026-10-md-media-and-documents/pandoc-ext-multibib/README.md +0 -4
  44. campaign_files/2026-10-md-media-and-documents/pandoc-ext-pagebreak/README.md +0 -4
  45. campaign_files/2026-10-md-media-and-documents/pandoc-ext-section-bibliographies/README.md +0 -4
  46. campaign_files/2026-10-md-media-and-documents/parsoid-parsertests-definition-lists/README.md +0 -4
  47. campaign_files/2026-10-ml-serving-configs-2/openvino-op-reference-tests/README.md +0 -6
  48. campaign_files/2026-10-ml-serving-configs-2/openvinotoolkit-nncf/README.md +0 -24
  49. campaign_files/2026-10-ml-serving-configs-2/outlines-core-json-schema-regex-tests/README.md +0 -4
  50. campaign_files/2026-10-op-rule-calculations-2/openpeppol-peppol-bis-invoice-3/README.md +0 -22
CHANGELOG.md CHANGED
@@ -3285,3 +3285,21 @@ Changes to the content of this index, oldest first. There is one entry for each
3285
  - hf-training-records: refused
3286
  - npm: refused
3287
  - stackexchange-api-posts: refused
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3285
  - hf-training-records: refused
3286
  - npm: refused
3287
  - stackexchange-api-posts: refused
3288
+
3289
+ ## 2026-11-11: sources held back
3290
+
3291
+ - Each source below keeps its newest version that passed every check; its newer batch waits until it passes them:
3292
+ - borealis-dataverse-workbooks-cc-by: changed pointer group file data/borealis-dataverse-workbooks-cc-by/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 2 key(s) of the old version are in no data file of source borealis-dataverse-workbooks-cc-by after the commit (row 4 from batch 2026-10-a-spreadsheets-with-answers, row 5 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
3293
+ - borealis-dataverse-workbooks-cc0: changed pointer group file data/borealis-dataverse-workbooks-cc0/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 1 key(s) of the old version are in no data file of source borealis-dataverse-workbooks-cc0 after the commit (row 7 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
3294
+ - cfpb-complaints: refused
3295
+ - data-gov-uk-workbooks: changed pointer group file data/data-gov-uk-workbooks/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 4 key(s) of the old version are in no data file of source data-gov-uk-workbooks after the commit (row 5 from batch 2026-10-a-spreadsheets-with-answers, row 6 from batch 2026-10-a-spreadsheets-with-answers, row 12 from batch 2026-10-a-spreadsheets-with-answers, 1 more); the removal record holds none of them, and none fails a publish check now
3296
+ - data-gov-uk-workbooks-noassertion: changed pointer group file data/data-gov-uk-workbooks-noassertion/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 1 key(s) of the old version are in no data file of source data-gov-uk-workbooks-noassertion after the commit (row 6 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
3297
+ - dataversenl-workbooks-cc-by: changed pointer group file data/dataversenl-workbooks-cc-by/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 1 key(s) of the old version are in no data file of source dataversenl-workbooks-cc-by after the commit (row 17 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
3298
+ - dataversenl-workbooks-cc0: changed pointer group file data/dataversenl-workbooks-cc0/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 3 key(s) of the old version are in no data file of source dataversenl-workbooks-cc0 after the commit (row 7 from batch 2026-10-a-spreadsheets-with-answers, row 13 from batch 2026-10-a-spreadsheets-with-answers, row 16 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
3299
+ - dvn-fvhlgg-repro: changed pointer group file data/dvn-fvhlgg-repro/2026-10-c-repros/part-00000.parquet: 3 key(s) of the old version are in no data file of source dvn-fvhlgg-repro after the commit (row 5 from batch 2026-10-c-repros, row 7 from batch 2026-10-c-repros, row 16 from batch 2026-10-c-repros); the removal record holds none of them, and none fails a publish check now
3300
+ - eu-institution-workbooks-cc-by: changed pointer group file data/eu-institution-workbooks-cc-by/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 3 key(s) of the old version are in no data file of source eu-institution-workbooks-cc-by after the commit (row 2 from batch 2026-10-a-spreadsheets-with-answers, row 4 from batch 2026-10-a-spreadsheets-with-answers, row 5 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
3301
+ - hdx-workbooks-cc0: changed pointer group file data/hdx-workbooks-cc0/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 2 key(s) of the old version are in no data file of source hdx-workbooks-cc0 after the commit (row 4 from batch 2026-10-a-spreadsheets-with-answers, row 5 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
3302
+ - hdx-workbooks-odbl: changed pointer group file data/hdx-workbooks-odbl/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 1 key(s) of the old version are in no data file of source hdx-workbooks-odbl after the commit (row 4 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
3303
+ - hf-training-records: refused
3304
+ - npm: refused
3305
+ - stackexchange-api-posts: refused
README.md CHANGED
@@ -24,25 +24,25 @@ configs:
24
  <!-- artifact-pointer-layout: v2 -->
25
  <!-- Generated by artpool.index_projection from meta/, cards/*.yaml, the field registry and its templates. Every release regenerates it, so edits to this file do not last. -->
26
 
27
- <img class="not-prose" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/banner.svg" alt="Public web, in two layers (task material in collections; corpora), one row per file (address, size, SHA-256, license), your selection, fetched and verified">
28
 
29
  This dataset is the public index of the **Training Data Foundation**: find public files to build agent tasks and environments from, and fetch them with one command.
30
 
31
  English | [简体中文](README.zh.md) · [Data page](https://huggingface.co/spaces/Battam/training-data-foundation) · [Usage video](https://huggingface.co/spaces/Battam/training-data-foundation#usage-video)
32
 
33
  <picture class="not-prose">
34
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/numbers-narrow-light.svg">
35
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/numbers-wide-light.svg" alt="1,808 material sets with 4,195 units in 14 collections; 139 corpora with 1,216,362,128 rows; 4.45 PB pointed to, as the index holds addresses, not copies; 99.6% of the 33,826 rows compared match their recorded size.">
36
  </picture>
37
  <picture class="not-prose">
38
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/numbers-narrow-dark.svg">
39
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/numbers-wide-dark.svg" alt="1,808 material sets with 4,195 units in 14 collections; 139 corpora with 1,216,362,128 rows; 4.45 PB pointed to, as the index holds addresses, not copies; 99.6% of the 33,826 rows compared match their recorded size.">
40
  </picture>
41
 
42
  ## Get data in one command
43
 
44
  ```sh
45
- uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/tools/fetch.py --form spreadsheets --sample 3 --out ./data
46
  ```
47
 
48
  This fetches 3 units of every spreadsheet material set into `./data`, one folder per unit, and checks each file against the size and hashes its row records.
@@ -52,7 +52,7 @@ the same command takes the same units of this release.
52
  <details>
53
  <summary>What the command takes, and in which order</summary>
54
 
55
- The command and the queries below are pinned to the release tag `release-2026-11-11.19`.
56
  [uv](https://docs.astral.sh/uv/) reads the script's dependencies, at the versions published before the day of this release.
57
 
58
  The command returns units: 3 units from each material set whose form is `spreadsheets`, each in a folder of its own named after its `unit_id`.
@@ -80,7 +80,7 @@ Choose material sets from the catalog with [DuckDB](https://duckdb.org/):
80
 
81
  ```sql
82
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
83
- FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-11.19/meta/catalog.parquet'
84
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
85
  ORDER BY collection, units DESC;
86
  ```
@@ -91,7 +91,7 @@ with `kind = 'corpus'` it returns corpora instead, counted in rows (`files`).
91
  Choose files in one corpus:
92
 
93
  ```sql
94
- SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-11.19/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-11.19/meta/catalog.parquet' WHERE source = 'if-archive'));
95
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
96
  FROM read_parquet(getvariable('files'))
97
  WHERE source = 'if-archive' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
@@ -113,12 +113,12 @@ A data file may hold the rows of several sources, so the query keeps this one's
113
  ### Task material: units by collection
114
 
115
  <picture class="not-prose">
116
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/material-narrow-light.svg">
117
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/material-wide-light.svg" alt="Units by collection, longest bar first: Spreadsheets 536; Browser tests 439; Tests and exercises 425; Model serving configurations 398; Media and documents 391; Science reference data 342; Rule calculations 329; Security detection 328; Hardware design (EDA) tests 258; Issue trackers 258; Solver instances 197; Disk images 179; Reproduction packages and runs 106; In no collection 9; SEC filing reviews 0.">
118
  </picture>
119
  <picture class="not-prose">
120
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/material-narrow-dark.svg">
121
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/material-wide-dark.svg" alt="Units by collection, longest bar first: Spreadsheets 536; Browser tests 439; Tests and exercises 425; Model serving configurations 398; Media and documents 391; Science reference data 342; Rule calculations 329; Security detection 328; Hardware design (EDA) tests 258; Issue trackers 258; Solver instances 197; Disk images 179; Reproduction packages and runs 106; In no collection 9; SEC filing reviews 0.">
122
  </picture>
123
 
124
  1,808 material sets with 4,195 units, in 14 collections (1 of them with no material set yet); 1 of the material sets are in no collection.
@@ -151,12 +151,12 @@ Upstreams are the upstream works the material sets of a collection come from.
151
  ### Corpora: rows by main domain
152
 
153
  <picture class="not-prose">
154
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/corpora-narrow-light.svg">
155
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/corpora-wide-light.svg" alt="Corpus rows by main domain, longest bar first: Data science 750,815,212; System administration 185,826,217; Many domains 109,262,742; Software engineering 45,267,701; Science 44,633,675; Operations 39,404,642; Media 12,751,612; Machine learning 9,636,889; Hardware 8,716,520; Unclassified 5,284,162; Security 4,521,631; Misc 241,125.">
156
  </picture>
157
  <picture class="not-prose">
158
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/corpora-narrow-dark.svg">
159
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/corpora-wide-dark.svg" alt="Corpus rows by main domain, longest bar first: Data science 750,815,212; System administration 185,826,217; Many domains 109,262,742; Software engineering 45,267,701; Science 44,633,675; Operations 39,404,642; Media 12,751,612; Machine learning 9,636,889; Hardware 8,716,520; Unclassified 5,284,162; Security 4,521,631; Misc 241,125.">
160
  </picture>
161
 
162
  139 corpora with 1,216,362,128 rows.
@@ -189,12 +189,12 @@ Every material set and every corpus, with its counts, size, license tier and log
189
  ## How it is made
190
 
191
  <picture class="not-prose">
192
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/steps-narrow-light.svg">
193
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/steps-wide-light.svg" alt="How the index is made, in four steps. 1, collect: collectors read whole upstreams into corpora, one row per file or record. 2, admit: works the search found upstream and groups of corpus rows become material sets. 3, check: licenses, personal data, overlap with benchmarks and sizes are checked before release. 4, release: one commit with its own tag, release-2026-11-11.19.">
194
  </picture>
195
  <picture class="not-prose">
196
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/steps-narrow-dark.svg">
197
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/steps-wide-dark.svg" alt="How the index is made, in four steps. 1, collect: collectors read whole upstreams into corpora, one row per file or record. 2, admit: works the search found upstream and groups of corpus rows become material sets. 3, check: licenses, personal data, overlap with benchmarks and sizes are checked before release. 4, release: one commit with its own tag, release-2026-11-11.19.">
198
  </picture>
199
 
200
  The terms, each step and how every number is made are folded below.
@@ -248,9 +248,9 @@ only their sizes add up to the size of the whole index.
248
 
249
  | Level | Recorded size | Deduplicated size | Rows read | Rows compared | Rows that match | Rows read, not compared | Rows without a recorded size | Their estimated size |
250
  |---|---:|---:|---:|---:|---:|---:|---:|---:|
251
- | Whole index | 4.45 PB | 4.11 PB | 34,704 | 33,826 | 33,697 (99.6%) | 878: address unreachable 449, file gone 80, login needed 329, no size from the server 19, other error 1 | 312,300,338 | 202 TB, an estimate (95% range 171 TB to 245 TB), for 215,326,513 of them, from 14,161 rows read |
252
  | Corpora | 4.45 PB | 4.11 PB | 31,594 | 30,723 | 30,610 (99.6%) | 871: address unreachable 449, file gone 80, login needed 322, no size from the server 19, other error 1 | 312,300,338 | 202 TB, an estimate (95% range 171 TB to 245 TB), for 215,326,513 of them, from 14,161 rows read |
253
- | Task material | 2.88 GB | 2.82 GB | 3,110 | 3,103 | 3,087 (99.5%) | 7: login needed 7 | 0 | none |
254
 
255
  Rows read are the sample of rows the check reads in `meta/sizes.parquet`;
256
  a row compared matched its recorded size or did not, and a row read but not compared never reached its file.
@@ -259,7 +259,7 @@ above one source its range is the sum of its sources' ranges, wider than a joint
259
 
260
  The recorded size is the sum of `size_bytes` over the rows that have one (catalog column `bytes`). The check reads a sample of rows of every data file (`meta/sizes.parquet`): 20 of its rows with a recorded size, and of its rows without one its share of its source's sample, sized for their mean size within 25% at 95% confidence by the spread of the sizes read so far (at most 400 rows per source per round, 20 for a source with none read yet); its share is the rows that match their recorded size out of the rows it compared (status `match`, `size_mismatch` or `hash_mismatch`), and the rows it could not compare, which never reached their file, are counted beside it by status (`unreachable`, `gone`, `needs_auth`, `no_length`, `error`; catalog column `size_check`). The estimate for a source's rows without a recorded size (catalog columns `unsized_rows` and `unsized_estimate`) is their number times the mean measured size of such rows read, with a 95% BCa bootstrap range (`scipy.stats.bootstrap`, its default resample count); a source with fewer than 10 such rows read has none. Above one source, the estimate is the sum of its sources' estimates and its range the sum of their ranges, which is wider than a joint interval. The figures of every level are in `meta/size_figures.json`.
261
 
262
- Distinct content is the sum of `size_bytes` over the rows minus the sizes of the rows that repeat another row's content (the same sha256, else sha1, else md5 and size, within the same hash scope; one copy of each content is kept, in the first source by registry order), counted after each release over the data files it published, per source in the catalog column `distinct_bytes`; archive members whose archive the same source also lists whole have other hashes, so they are not subtracted. `tools/fetch.py` removes duplicates over your own selection when it reads it. The deduplicated sizes here were counted over the data files of release `release-2026-11-11.15`.
263
 
264
  </details>
265
 
@@ -273,7 +273,7 @@ The training permit and the personal-data mark are labels of every row, neither
273
  Each table below gives what a training stage can take, counted before the rules and after each of them, in their order, down to the stage's count;
274
  its last two rows count how much of the stage's count is allowed for training (every file's training permit yes or conditional) and how much is marked as carrying personal data.
275
 
276
- This release has not counted its stage fields (`meta/stage_fields.parquet`) yet, so every count below says not counted yet.
277
 
278
  #### Post-training (in use)
279
 
@@ -281,12 +281,12 @@ Material set units for building tasks: every file of the unit in English or no n
281
 
282
  | Step | Material sets | Units | Upstream sources |
283
  |---|---:|---:|---:|
284
- | In the stage's condition, before the rules | not counted yet | not counted yet | not counted yet |
285
- | After the benchmark rule | not counted yet | not counted yet | not counted yet |
286
- | After the rule on rows not compared with the benchmark list | not counted yet | not counted yet | not counted yet |
287
- | After deduplication: the stage's count | not counted yet | not counted yet | not counted yet |
288
- | Of the stage's count, allowed for training (yes or conditional) | not counted yet | not counted yet | not counted yet |
289
- | Of the stage's count, marked as carrying personal data | not counted yet | not counted yet | not counted yet |
290
 
291
  #### Mid-training (provisional: conditions and counts only)
292
 
@@ -294,12 +294,12 @@ Corpus rows of code, documentation and the terminal side: conditions and counts
294
 
295
  | Step | Rows | Size |
296
  |---|---:|---:|
297
- | In the stage's condition, before the rules | not counted yet | not counted yet |
298
- | After the benchmark rule | not counted yet | not counted yet |
299
- | After the rule on rows not compared with the benchmark list | not counted yet | not counted yet |
300
- | After deduplication: the stage's count | not counted yet | not counted yet |
301
- | Of the stage's count, allowed for training (yes or conditional) | not counted yet | not counted yet |
302
- | Of the stage's count, marked as carrying personal data | not counted yet | not counted yet |
303
 
304
  #### Pre-training (provisional: conditions and counts only)
305
 
@@ -307,12 +307,12 @@ Corpus rows of plain text, the mid-training types not counted here: conditions a
307
 
308
  | Step | Rows | Size |
309
  |---|---:|---:|
310
- | In the stage's condition, before the rules | not counted yet | not counted yet |
311
- | After the benchmark rule | not counted yet | not counted yet |
312
- | After the rule on rows not compared with the benchmark list | not counted yet | not counted yet |
313
- | After deduplication: the stage's count | not counted yet | not counted yet |
314
- | Of the stage's count, allowed for training (yes or conditional) | not counted yet | not counted yet |
315
- | Of the stage's count, marked as carrying personal data | not counted yet | not counted yet |
316
 
317
  </details>
318
 
@@ -449,9 +449,9 @@ Every row field and every column of the catalog: [`docs/fields.md`](docs/fields.
449
  <details>
450
  <summary>Releases</summary>
451
 
452
- - Latest release: `release-2026-11-11.19` (its newest CHANGELOG entry is dated 2026-11-11).
453
  - Releases are listed in [CHANGELOG.md](CHANGELOG.md);
454
- each one is a commit with its own release tag (this one is [`release-2026-11-11.19`](https://huggingface.co/datasets/Battam/training-data-foundation/tree/release-2026-11-11.19)),
455
  so any earlier state stays readable at its tag.
456
  - `meta/projection.json` records the release tag and the version of the program and templates that generated this card.
457
  - Croissant metadata: [generated by the Hub](https://huggingface.co/api/datasets/Battam/training-data-foundation/croissant);
@@ -484,7 +484,7 @@ A removal covers the current version of this repository (the main branch) and th
484
  @misc{training_data_foundation,
485
  title = {training-data-foundation: an index of public materials for building agent environments},
486
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
487
- note = {Release release-2026-11-11.19},
488
  year = {2026}
489
  }
490
  ```
 
24
  <!-- artifact-pointer-layout: v2 -->
25
  <!-- Generated by artpool.index_projection from meta/, cards/*.yaml, the field registry and its templates. Every release regenerates it, so edits to this file do not last. -->
26
 
27
+ <img class="not-prose" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/banner.svg" alt="Public web, in two layers (task material in collections; corpora), one row per file (address, size, SHA-256, license), your selection, fetched and verified">
28
 
29
  This dataset is the public index of the **Training Data Foundation**: find public files to build agent tasks and environments from, and fetch them with one command.
30
 
31
  English | [简体中文](README.zh.md) · [Data page](https://huggingface.co/spaces/Battam/training-data-foundation) · [Usage video](https://huggingface.co/spaces/Battam/training-data-foundation#usage-video)
32
 
33
  <picture class="not-prose">
34
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/numbers-narrow-light.svg">
35
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/numbers-wide-light.svg" alt="1,808 material sets with 4,195 units in 14 collections; 139 corpora with 1,216,362,128 rows; 4.45 PB pointed to, as the index holds addresses, not copies; 99.6% of the 33,506 rows compared match their recorded size.">
36
  </picture>
37
  <picture class="not-prose">
38
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/numbers-narrow-dark.svg">
39
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/numbers-wide-dark.svg" alt="1,808 material sets with 4,195 units in 14 collections; 139 corpora with 1,216,362,128 rows; 4.45 PB pointed to, as the index holds addresses, not copies; 99.6% of the 33,506 rows compared match their recorded size.">
40
  </picture>
41
 
42
  ## Get data in one command
43
 
44
  ```sh
45
+ uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/tools/fetch.py --form spreadsheets --sample 3 --out ./data
46
  ```
47
 
48
  This fetches 3 units of every spreadsheet material set into `./data`, one folder per unit, and checks each file against the size and hashes its row records.
 
52
  <details>
53
  <summary>What the command takes, and in which order</summary>
54
 
55
+ The command and the queries below are pinned to the release tag `release-2026-11-11.20`.
56
  [uv](https://docs.astral.sh/uv/) reads the script's dependencies, at the versions published before the day of this release.
57
 
58
  The command returns units: 3 units from each material set whose form is `spreadsheets`, each in a folder of its own named after its `unit_id`.
 
80
 
81
  ```sql
82
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
83
+ FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-11.20/meta/catalog.parquet'
84
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
85
  ORDER BY collection, units DESC;
86
  ```
 
91
  Choose files in one corpus:
92
 
93
  ```sql
94
+ SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-11.20/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-11.20/meta/catalog.parquet' WHERE source = 'if-archive'));
95
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
96
  FROM read_parquet(getvariable('files'))
97
  WHERE source = 'if-archive' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
 
113
  ### Task material: units by collection
114
 
115
  <picture class="not-prose">
116
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/material-narrow-light.svg">
117
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/material-wide-light.svg" alt="Units by collection, longest bar first: Spreadsheets 536; Browser tests 439; Tests and exercises 425; Model serving configurations 398; Media and documents 391; Science reference data 342; Rule calculations 329; Security detection 328; Hardware design (EDA) tests 258; Issue trackers 258; Solver instances 197; Disk images 179; Reproduction packages and runs 106; In no collection 9; SEC filing reviews 0.">
118
  </picture>
119
  <picture class="not-prose">
120
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/material-narrow-dark.svg">
121
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/material-wide-dark.svg" alt="Units by collection, longest bar first: Spreadsheets 536; Browser tests 439; Tests and exercises 425; Model serving configurations 398; Media and documents 391; Science reference data 342; Rule calculations 329; Security detection 328; Hardware design (EDA) tests 258; Issue trackers 258; Solver instances 197; Disk images 179; Reproduction packages and runs 106; In no collection 9; SEC filing reviews 0.">
122
  </picture>
123
 
124
  1,808 material sets with 4,195 units, in 14 collections (1 of them with no material set yet); 1 of the material sets are in no collection.
 
151
  ### Corpora: rows by main domain
152
 
153
  <picture class="not-prose">
154
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/corpora-narrow-light.svg">
155
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/corpora-wide-light.svg" alt="Corpus rows by main domain, longest bar first: Data science 750,815,212; System administration 185,826,217; Many domains 109,262,742; Software engineering 45,267,701; Science 44,633,675; Operations 39,404,642; Media 12,751,612; Machine learning 9,636,889; Hardware 8,716,520; Unclassified 5,284,162; Security 4,521,631; Misc 241,125.">
156
  </picture>
157
  <picture class="not-prose">
158
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/corpora-narrow-dark.svg">
159
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/corpora-wide-dark.svg" alt="Corpus rows by main domain, longest bar first: Data science 750,815,212; System administration 185,826,217; Many domains 109,262,742; Software engineering 45,267,701; Science 44,633,675; Operations 39,404,642; Media 12,751,612; Machine learning 9,636,889; Hardware 8,716,520; Unclassified 5,284,162; Security 4,521,631; Misc 241,125.">
160
  </picture>
161
 
162
  139 corpora with 1,216,362,128 rows.
 
189
  ## How it is made
190
 
191
  <picture class="not-prose">
192
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/steps-narrow-light.svg">
193
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/steps-wide-light.svg" alt="How the index is made, in four steps. 1, collect: collectors read whole upstreams into corpora, one row per file or record. 2, admit: works the search found upstream and groups of corpus rows become material sets. 3, check: licenses, personal data, overlap with benchmarks and sizes are checked before release. 4, release: one commit with its own tag, release-2026-11-11.20.">
194
  </picture>
195
  <picture class="not-prose">
196
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/steps-narrow-dark.svg">
197
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/steps-wide-dark.svg" alt="How the index is made, in four steps. 1, collect: collectors read whole upstreams into corpora, one row per file or record. 2, admit: works the search found upstream and groups of corpus rows become material sets. 3, check: licenses, personal data, overlap with benchmarks and sizes are checked before release. 4, release: one commit with its own tag, release-2026-11-11.20.">
198
  </picture>
199
 
200
  The terms, each step and how every number is made are folded below.
 
248
 
249
  | Level | Recorded size | Deduplicated size | Rows read | Rows compared | Rows that match | Rows read, not compared | Rows without a recorded size | Their estimated size |
250
  |---|---:|---:|---:|---:|---:|---:|---:|---:|
251
+ | Whole index | 4.45 PB | 4.11 PB | 34,384 | 33,506 | 33,377 (99.6%) | 878: address unreachable 449, file gone 80, login needed 329, no size from the server 19, other error 1 | 312,300,338 | 202 TB, an estimate (95% range 171 TB to 245 TB), for 215,326,513 of them, from 14,161 rows read |
252
  | Corpora | 4.45 PB | 4.11 PB | 31,594 | 30,723 | 30,610 (99.6%) | 871: address unreachable 449, file gone 80, login needed 322, no size from the server 19, other error 1 | 312,300,338 | 202 TB, an estimate (95% range 171 TB to 245 TB), for 215,326,513 of them, from 14,161 rows read |
253
+ | Task material | 2.88 GB | 2.82 GB | 2,790 | 2,783 | 2,767 (99.4%) | 7: login needed 7 | 0 | none |
254
 
255
  Rows read are the sample of rows the check reads in `meta/sizes.parquet`;
256
  a row compared matched its recorded size or did not, and a row read but not compared never reached its file.
 
259
 
260
  The recorded size is the sum of `size_bytes` over the rows that have one (catalog column `bytes`). The check reads a sample of rows of every data file (`meta/sizes.parquet`): 20 of its rows with a recorded size, and of its rows without one its share of its source's sample, sized for their mean size within 25% at 95% confidence by the spread of the sizes read so far (at most 400 rows per source per round, 20 for a source with none read yet); its share is the rows that match their recorded size out of the rows it compared (status `match`, `size_mismatch` or `hash_mismatch`), and the rows it could not compare, which never reached their file, are counted beside it by status (`unreachable`, `gone`, `needs_auth`, `no_length`, `error`; catalog column `size_check`). The estimate for a source's rows without a recorded size (catalog columns `unsized_rows` and `unsized_estimate`) is their number times the mean measured size of such rows read, with a 95% BCa bootstrap range (`scipy.stats.bootstrap`, its default resample count); a source with fewer than 10 such rows read has none. Above one source, the estimate is the sum of its sources' estimates and its range the sum of their ranges, which is wider than a joint interval. The figures of every level are in `meta/size_figures.json`.
261
 
262
+ Distinct content is the sum of `size_bytes` over the rows minus the sizes of the rows that repeat another row's content (the same sha256, else sha1, else md5 and size, within the same hash scope; one copy of each content is kept, in the first source by registry order), counted after each release over the data files it published, per source in the catalog column `distinct_bytes`; archive members whose archive the same source also lists whole have other hashes, so they are not subtracted. `tools/fetch.py` removes duplicates over your own selection when it reads it. The deduplicated sizes here were counted over the data files of release `release-2026-11-11.17`.
263
 
264
  </details>
265
 
 
273
  Each table below gives what a training stage can take, counted before the rules and after each of them, in their order, down to the stage's count;
274
  its last two rows count how much of the stage's count is allowed for training (every file's training permit yes or conditional) and how much is marked as carrying personal data.
275
 
276
+ The stage fields (`meta/stage_fields.parquet`) were counted over the data files of release `release-2026-11-11.17`, after it was published.
277
 
278
  #### Post-training (in use)
279
 
 
281
 
282
  | Step | Material sets | Units | Upstream sources |
283
  |---|---:|---:|---:|
284
+ | In the stage's condition, before the rules | 1,788 | 4,150 | 1,843 |
285
+ | After the benchmark rule | 1,787 | 4,149 | 1,842 |
286
+ | After the rule on rows not compared with the benchmark list | 1,787 | 4,149 | 1,842 |
287
+ | After deduplication: the stage's count | 1,758 | 4,040 | 1,817 |
288
+ | Of the stage's count, allowed for training (yes or conditional) | 1,405 | 3,397 | 1,333 |
289
+ | Of the stage's count, marked as carrying personal data | 357 | 631 | 400 |
290
 
291
  #### Mid-training (provisional: conditions and counts only)
292
 
 
294
 
295
  | Step | Rows | Size |
296
  |---|---:|---:|
297
+ | In the stage's condition, before the rules | 34,477,690 | 4.00 TB |
298
+ | After the benchmark rule | 34,468,408 | 4.00 TB |
299
+ | After the rule on rows not compared with the benchmark list | 34,468,408 | 4.00 TB |
300
+ | After deduplication: the stage's count | 34,079,767 | 3.86 TB |
301
+ | Of the stage's count, allowed for training (yes or conditional) | 9,890,653 | 1.42 TB |
302
+ | Of the stage's count, marked as carrying personal data | 6,102,344 | 341 GB |
303
 
304
  #### Pre-training (provisional: conditions and counts only)
305
 
 
307
 
308
  | Step | Rows | Size |
309
  |---|---:|---:|
310
+ | In the stage's condition, before the rules | 154,035,344 | 30.7 TB |
311
+ | After the benchmark rule | 154,034,327 | 30.7 TB |
312
+ | After the rule on rows not compared with the benchmark list | 154,034,327 | 30.7 TB |
313
+ | After deduplication: the stage's count | 149,118,607 | 30.6 TB |
314
+ | Of the stage's count, allowed for training (yes or conditional) | 117,929,843 | 29.9 TB |
315
+ | Of the stage's count, marked as carrying personal data | 40,455,308 | 676 GB |
316
 
317
  </details>
318
 
 
449
  <details>
450
  <summary>Releases</summary>
451
 
452
+ - Latest release: `release-2026-11-11.20` (its newest CHANGELOG entry is dated 2026-11-11).
453
  - Releases are listed in [CHANGELOG.md](CHANGELOG.md);
454
+ each one is a commit with its own release tag (this one is [`release-2026-11-11.20`](https://huggingface.co/datasets/Battam/training-data-foundation/tree/release-2026-11-11.20)),
455
  so any earlier state stays readable at its tag.
456
  - `meta/projection.json` records the release tag and the version of the program and templates that generated this card.
457
  - Croissant metadata: [generated by the Hub](https://huggingface.co/api/datasets/Battam/training-data-foundation/croissant);
 
484
  @misc{training_data_foundation,
485
  title = {training-data-foundation: an index of public materials for building agent environments},
486
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
487
+ note = {Release release-2026-11-11.20},
488
  year = {2026}
489
  }
490
  ```
README.zh.md CHANGED
@@ -11,25 +11,25 @@ tags:
11
  ---
12
  <!-- Generated by artpool.index_projection from meta/, cards/*.yaml, the field registry and its templates. Every release regenerates it, so edits to this file do not last. -->
13
 
14
- <img class="not-prose" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/banner.zh.svg" alt="公开网络,分两层(按合集的任务材料;语料库),每个文件一行(地址、大小、SHA-256、许可证),你的选择,下载并校验">
15
 
16
  本数据集是**数据基座**的公开索引:在这里找到可以用来构建智能体任务和环境的公开文件,并用一条命令把它们取下来。
17
 
18
  [English](README.md) | 简体中文 · [数据页](https://huggingface.co/spaces/Battam/training-data-foundation) · [使用示例视频](https://huggingface.co/spaces/Battam/training-data-foundation#usage-video)
19
 
20
  <picture class="not-prose">
21
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/numbers-narrow-light.zh.svg">
22
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/numbers-wide-light.zh.svg" alt="1,808 个材料集,共 4,195 个单元,分在 14 个合集;139 个语料库,共 1,216,362,128 行;指向的大小共 4.45 PB,索引存的是地址,不是文件副本;比对过的 33,826 行中,99.6% 与记录的大小一致。">
23
  </picture>
24
  <picture class="not-prose">
25
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/numbers-narrow-dark.zh.svg">
26
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/numbers-wide-dark.zh.svg" alt="1,808 个材料集,共 4,195 个单元,分在 14 个合集;139 个语料库,共 1,216,362,128 行;指向的大小共 4.45 PB,索引存的是地址,不是文件副本;比对过的 33,826 行中,99.6% 与记录的大小一致。">
27
  </picture>
28
 
29
  ## 一条命令取数据
30
 
31
  ```sh
32
- uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/tools/fetch.py --form spreadsheets --sample 3 --out ./data
33
  ```
34
 
35
  这条命令从每个电子表格材料集各取 3 个单元放进 `./data`,每个单元一个文件夹,并按每个文件所在行记录的大小和哈希值校验它。
@@ -39,7 +39,7 @@ uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/r
39
  <details>
40
  <summary>命令取什么、按什么顺序取</summary>
41
 
42
- 这条命令和下面的查询都固定在发布标签 `release-2026-11-11.19` 上。
43
  [uv](https://docs.astral.sh/uv/) 会读取脚本声明的依赖,取这个发布当天之前发布的版本。
44
 
45
  命令返回单元:形式为 `spreadsheets` 的每个材料集各 3 个单元,每个单元放在以它的 `unit_id` 命名的单独文件夹里。
@@ -67,7 +67,7 @@ uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/r
67
 
68
  ```sql
69
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
70
- FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-11.19/meta/catalog.parquet'
71
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
72
  ORDER BY collection, units DESC;
73
  ```
@@ -78,7 +78,7 @@ ORDER BY collection, units DESC;
78
  在一个语料库里挑选文件:
79
 
80
  ```sql
81
- SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-11.19/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-11.19/meta/catalog.parquet' WHERE source = 'if-archive'));
82
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
83
  FROM read_parquet(getvariable('files'))
84
  WHERE source = 'if-archive' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
@@ -100,12 +100,12 @@ ORDER BY size_bytes LIMIT 20;
100
  ### 任务材料:按合集的单元数
101
 
102
  <picture class="not-prose">
103
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/material-narrow-light.zh.svg">
104
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/material-wide-light.zh.svg" alt="按合集的单元数,条形由长到短:电子表格 536;浏览器测试 439;测试和练习 425;模型服务配置 398;媒体和文档 391;科学参考数据 342;规则计算 329;安全检测 328;硬件设计(EDA)测试 258;问题跟踪系统 258;求解器实例 197;磁盘镜像 179;复现包和运行 106;不属于任何合集 9;SEC 申报审查 0。">
105
  </picture>
106
  <picture class="not-prose">
107
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/material-narrow-dark.zh.svg">
108
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/material-wide-dark.zh.svg" alt="按合集的单元数,条形由长到短:电子表格 536;浏览器测试 439;测试和练习 425;模型服务配置 398;媒体和文档 391;科学参考数据 342;规则计算 329;安全检测 328;硬件设计(EDA)测试 258;问题跟踪系统 258;求解器实例 197;磁盘镜像 179;复现包和运行 106;不属于任何合集 9;SEC 申报审查 0。">
109
  </picture>
110
 
111
  1,808 个材料集,共 4,195 个单元,分在 14 个合集中(其中 1 个合集还没有材料集);其中 1 个材料集不属于任何合集。
@@ -138,12 +138,12 @@ ORDER BY size_bytes LIMIT 20;
138
  ### 语料库:按主要领域的行数
139
 
140
  <picture class="not-prose">
141
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/corpora-narrow-light.zh.svg">
142
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/corpora-wide-light.zh.svg" alt="按主要领域的语料库行数,条形由长到短:数据科学 750,815,212;系统管理 185,826,217;多领域 109,262,742;软件工程 45,267,701;科学 44,633,675;运营 39,404,642;媒体 12,751,612;机器学习 9,636,889;硬件 8,716,520;未分类 5,284,162;安全 4,521,631;杂项 241,125。">
143
  </picture>
144
  <picture class="not-prose">
145
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/corpora-narrow-dark.zh.svg">
146
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/corpora-wide-dark.zh.svg" alt="按主要领域的语料库行数,条形由长到短:数据科学 750,815,212;系统管理 185,826,217;多领域 109,262,742;软件工程 45,267,701;科学 44,633,675;运营 39,404,642;媒体 12,751,612;机器学习 9,636,889;硬件 8,716,520;未分类 5,284,162;安全 4,521,631;杂项 241,125。">
147
  </picture>
148
 
149
  139 个语料库,共 1,216,362,128 行。
@@ -176,12 +176,12 @@ ORDER BY size_bytes LIMIT 20;
176
  ## 怎样做成
177
 
178
  <picture class="not-prose">
179
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/steps-narrow-light.zh.svg">
180
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/steps-wide-light.zh.svg" alt="索引分四步做成。1,采集:采集器整批读取完整上游,每个文件或每条记录记一行,成为语料库。2,收录:检索在上游找到的作品和成组的语料库行,收录为材料集。3,检查:发布前核对许可证、个人信息、与基准测试的重叠和大小。4,发布:一个带有自己标签的提交,release-2026-11-11.19。">
181
  </picture>
182
  <picture class="not-prose">
183
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/steps-narrow-dark.zh.svg">
184
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.19/docs/figures/steps-wide-dark.zh.svg" alt="索引分四步做成。1,采集:采集器整批读取完整上游,每个文件或每条记录记一行,成为语料库。2,收录:检索在上游找到的作品和成组的语料库行,收录为材料集。3,检查:发布前核对许可证、个人信息、与基准测试的重叠和大小。4,发布:一个带有自己标签的提交,release-2026-11-11.19。">
185
  </picture>
186
 
187
  术语、每一步,以及每个数字怎样得出,都折叠在下面。
@@ -235,9 +235,9 @@ ORDER BY size_bytes LIMIT 20;
235
 
236
  | 层级 | 记录的大小 | 去重后大小 | 读过的行 | 比对过的行 | 一致的行 | 读过但未比对的行 | 没有记录大小的行 | 这些行的估计大小 |
237
  |---|---:|---:|---:|---:|---:|---:|---:|---:|
238
- | 整个索引 | 4.45 PB | 4.11 PB | 34,704 | 33,826 | 33,697(99.6%) | 878:地址无法访问 449 行、文件已不存在 80 行、需要登录 329 行、服务器未给出大小 19 行、其他错误 1 行 | 312,300,338 | 202 TB,这是估计值(95% 范围 171 TB 到 245 TB),针对其中 215,326,513 行,依据读过的 14,161 行 |
239
  | 语料库 | 4.45 PB | 4.11 PB | 31,594 | 30,723 | 30,610(99.6%) | 871:地址无法访问 449 行、文件已不存在 80 行、需要登录 322 行、服务器未给出大小 19 行、其他错误 1 行 | 312,300,338 | 202 TB,这是估计值(95% 范围 171 TB 到 245 TB),针对其中 215,326,513 行,依据读过的 14,161 行 |
240
- | 任务材料 | 2.88 GB | 2.82 GB | 3,110 | 3,103 | 3,087(99.5%) | 7:需要登录 7 行 | 0 | 无 |
241
 
242
  读过的行是核对在 `meta/sizes.parquet` 中读取的抽样行;
243
  比对过的行与记录的大小一致或不一致,读过但未比对的行没有取到它的文件。
@@ -246,7 +246,7 @@ ORDER BY size_bytes LIMIT 20;
246
 
247
  记录的大小是有大小的各行 `size_bytes` 之和(目录列 `bytes`)。核对读取每个数据文件中抽取的一组行(`meta/sizes.parquet`):有记录大小的行抽 20 行;没有记录大小的行按其所在来源的样本分摊,样本量按目前读到的大小的离散程度定,使其平均大小在 95% 置信度下误差不超过 25%(每个来源每轮最多 400 行,还没有读数的来源取 20 行);其比例是比对过的行(状态 `match`、`size_mismatch` 或 `hash_mismatch`)中与记录的大小一致的行所占的比例,没能比对的行(没有取到文件)按状态另行计数(`unreachable`、`gone`、`needs_auth`、`no_length`、`error`;目录列 `size_check`)。一个来源没有记录大小的行的估计(目录列 `unsized_rows` 和 `unsized_estimate`)是这些行的行数乘以读过的这类行的平均实测大小,并给出 95% BCa 自助法范围(`scipy.stats.bootstrap`,用其默认重抽样次数);读过的这类行少于 10 行的来源没有估计。在来源之上的层级,估计是其各来源估计之和,范围是各来源范围之和,比联合区间更宽。每个层级的数字都在 `meta/size_figures.json` 中。
248
 
249
- 去重后内容是各行 `size_bytes` 之和,减去与另一行内容重复的那些行的大小(sha256 相同;没有 sha256 时 sha1 相同;再没有时 md5 和大小都相同;并且哈希范围相同;每种内容只保留一份,留在按注册顺序排在最前的来源中),在每次发布之后按它发布的数据文件重新计算,每个来源的值在目录列 `distinct_bytes` 中;同一来源也整个列出的存档,其成员的哈希与存档不同,所以不扣除。`tools/fetch.py` 读取你自己的选择时会去掉其中的重复。 这里的去重后大小是按发布 `release-2026-11-11.15` 的数据文件算出的。
250
 
251
  </details>
252
 
@@ -260,7 +260,7 @@ ORDER BY size_bytes LIMIT 20;
260
  下面每张表给出一个训练阶段能用的材料,在各条规则之前和按顺序每条规则之后计数,直到该阶段的计数;
261
  每张表的最后两行统计该阶段的计数中有��少允许训练(每个文件的训练许可为 yes 或 conditional),有多少标为带有个人数据。
262
 
263
- 本次发布还没有计算它的阶段字段(`meta/stage_fields.parquet`),所以下面每个计数都写作尚未计数。
264
 
265
  #### 后训练(在用)
266
 
@@ -268,12 +268,12 @@ ORDER BY size_bytes LIMIT 20;
268
 
269
  | 步骤 | 材料集数 | 单元数 | 上游来源数 |
270
  |---|---:|---:|---:|
271
- | 符合该阶段的条件,在各条规则之前 | 尚未计数 | 尚未计数 | 尚未计数 |
272
- | 基准测试规则之后 | 尚未计数 | 尚未计数 | 尚未计数 |
273
- | 未与基准测试清单比对的行的规则之后 | 尚未计数 | 尚未计数 | 尚未计数 |
274
- | 去重之后:该阶段的计数 | 尚未计数 | 尚未计数 | 尚未计数 |
275
- | 该阶段的计数中,允许训练的(yes 或 conditional) | 尚未计数 | 尚未计数 | 尚未计数 |
276
- | 该阶段的计数中,标为带有个人数据的 | 尚未计数 | 尚未计数 | 尚未计数 |
277
 
278
  #### 中训练(暂定:只给条件和计数)
279
 
@@ -281,12 +281,12 @@ ORDER BY size_bytes LIMIT 20;
281
 
282
  | 步骤 | 行数 | 大小 |
283
  |---|---:|---:|
284
- | 符合该阶段的条件,在各条规则之前 | 尚未计数 | 尚未计数 |
285
- | 基准测试规则之后 | 尚未计数 | 尚未计数 |
286
- | 未与基准测试清单比对的行的规则之后 | 尚未计数 | 尚未计数 |
287
- | 去重之后:该阶段的计数 | 尚未计数 | 尚未计数 |
288
- | 该阶段的计数中,允许训练的(yes 或 conditional) | 尚未计数 | 尚未计数 |
289
- | 该阶段的计数中,标为带有个人数据的 | 尚未计数 | 尚未计数 |
290
 
291
  #### 预训练(暂定:只给条件和计数)
292
 
@@ -294,12 +294,12 @@ ORDER BY size_bytes LIMIT 20;
294
 
295
  | 步骤 | 行数 | 大小 |
296
  |---|---:|---:|
297
- | 符合该阶段的条件,在各条规则之前 | 尚未计数 | 尚未计数 |
298
- | 基准测试规则之后 | 尚未计数 | 尚未计数 |
299
- | 未与基准测试清单比对的行的规则之后 | 尚未计数 | 尚未计数 |
300
- | 去重之后:该阶段的计数 | 尚未计数 | 尚未计数 |
301
- | 该阶段的计数中,允许训练的(yes 或 conditional) | 尚未计数 | 尚未计数 |
302
- | 该阶段的计数中,标为带有个人数据的 | 尚未计数 | 尚未计数 |
303
 
304
  </details>
305
 
@@ -438,9 +438,9 @@ ORDER BY size_bytes LIMIT 20;
438
  <details>
439
  <summary>发布</summary>
440
 
441
- - 最新发布:`release-2026-11-11.19`(其最新的 CHANGELOG 条目日期为 2026-11-11)。
442
  - 各次发布列在 [CHANGELOG.md](CHANGELOG.md) 中;
443
- 每次发布是一个带有自己发布标签的提交(本次为 [`release-2026-11-11.19`](https://huggingface.co/datasets/Battam/training-data-foundation/tree/release-2026-11-11.19)),
444
  因此任何较早的状态都可以在它的标签处读到。
445
  - `meta/projection.json` 记录发布标签,以及生成本卡片的程序和模板的版本。
446
  - Croissant 元数据:[由 Hub 生成](https://huggingface.co/api/datasets/Battam/training-data-foundation/croissant);
@@ -481,7 +481,7 @@ ORDER BY size_bytes LIMIT 20;
481
  @misc{training_data_foundation,
482
  title = {training-data-foundation: an index of public materials for building agent environments},
483
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
484
- note = {Release release-2026-11-11.19},
485
  year = {2026}
486
  }
487
  ```
 
11
  ---
12
  <!-- Generated by artpool.index_projection from meta/, cards/*.yaml, the field registry and its templates. Every release regenerates it, so edits to this file do not last. -->
13
 
14
+ <img class="not-prose" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/banner.zh.svg" alt="公开网络,分两层(按合集的任务材料;语料库),每个文件一行(地址、大小、SHA-256、许可证),你的选择,下载并校验">
15
 
16
  本数据集是**数据基座**的公开索引:在这里找到可以用来构建智能体任务和环境的公开文件,并用一条命令把它们取下来。
17
 
18
  [English](README.md) | 简体中文 · [数据页](https://huggingface.co/spaces/Battam/training-data-foundation) · [使用示例视频](https://huggingface.co/spaces/Battam/training-data-foundation#usage-video)
19
 
20
  <picture class="not-prose">
21
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/numbers-narrow-light.zh.svg">
22
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/numbers-wide-light.zh.svg" alt="1,808 个材料集,共 4,195 个单元,分在 14 个合集;139 个语料库,共 1,216,362,128 行;指向的大小共 4.45 PB,索引存的是地址,不是文件副本;比对过的 33,506 行中,99.6% 与记录的大小一致。">
23
  </picture>
24
  <picture class="not-prose">
25
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/numbers-narrow-dark.zh.svg">
26
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/numbers-wide-dark.zh.svg" alt="1,808 个材料集,共 4,195 个单元,分在 14 个合集;139 个语料库,共 1,216,362,128 行;指向的大小共 4.45 PB,索引存的是地址,不是文件副本;比对过的 33,506 行中,99.6% 与记录的大小一致。">
27
  </picture>
28
 
29
  ## 一条命令取数据
30
 
31
  ```sh
32
+ uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/tools/fetch.py --form spreadsheets --sample 3 --out ./data
33
  ```
34
 
35
  这条命令从每个电子表格材料集各取 3 个单元放进 `./data`,每个单元一个文件夹,并按每个文件所在行记录的大小和哈希值校验它。
 
39
  <details>
40
  <summary>命令取什么、按什么顺序取</summary>
41
 
42
+ 这条命令和下面的查询都固定在发布标签 `release-2026-11-11.20` 上。
43
  [uv](https://docs.astral.sh/uv/) 会读取脚本声明的依赖,取这个发布当天之前发布的版本。
44
 
45
  命令返回单元:形式为 `spreadsheets` 的每个材料集各 3 个单元,每个单元放在以它的 `unit_id` 命名的单独文件夹里。
 
67
 
68
  ```sql
69
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
70
+ FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-11.20/meta/catalog.parquet'
71
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
72
  ORDER BY collection, units DESC;
73
  ```
 
78
  在一个语料库里挑选文件:
79
 
80
  ```sql
81
+ SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-11.20/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-11.20/meta/catalog.parquet' WHERE source = 'if-archive'));
82
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
83
  FROM read_parquet(getvariable('files'))
84
  WHERE source = 'if-archive' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
 
100
  ### 任务材料:按合集的单元数
101
 
102
  <picture class="not-prose">
103
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/material-narrow-light.zh.svg">
104
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/material-wide-light.zh.svg" alt="按合集的单元数,条形由长到短:电子表格 536;浏览器测试 439;测试和练习 425;模型服务配置 398;媒体和文档 391;科学参考数据 342;规则计算 329;安全检测 328;硬件设计(EDA)测试 258;问题跟踪系统 258;求解器实例 197;磁盘镜像 179;复现包和运行 106;不属于任何合集 9;SEC 申报审查 0。">
105
  </picture>
106
  <picture class="not-prose">
107
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/material-narrow-dark.zh.svg">
108
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/material-wide-dark.zh.svg" alt="按合集的单元数,条形由长到短:电子表格 536;浏览器测试 439;测试和练习 425;模型服务配置 398;媒体和文档 391;科学参考数据 342;规则计算 329;安全检测 328;硬件设计(EDA)测试 258;问题跟踪系统 258;求解器实例 197;磁盘镜像 179;复现包和运行 106;不属于任何合集 9;SEC 申报审查 0。">
109
  </picture>
110
 
111
  1,808 个材料集,共 4,195 个单元,分在 14 个合集中(其中 1 个合集还没有材料集);其中 1 个材料集不属于任何合集。
 
138
  ### 语料库:按主要领域的行数
139
 
140
  <picture class="not-prose">
141
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/corpora-narrow-light.zh.svg">
142
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/corpora-wide-light.zh.svg" alt="按主要领域的语料库行数,条形由长到短:数据科学 750,815,212;系统管理 185,826,217;多领域 109,262,742;软件工程 45,267,701;科学 44,633,675;运营 39,404,642;媒体 12,751,612;机器学习 9,636,889;硬件 8,716,520;未分类 5,284,162;安全 4,521,631;杂项 241,125。">
143
  </picture>
144
  <picture class="not-prose">
145
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/corpora-narrow-dark.zh.svg">
146
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/corpora-wide-dark.zh.svg" alt="按主要领域的语料库行数,条形由长到短:数据科学 750,815,212;系统管理 185,826,217;多领域 109,262,742;软件工程 45,267,701;科学 44,633,675;运营 39,404,642;媒体 12,751,612;机器学习 9,636,889;硬件 8,716,520;未分类 5,284,162;安全 4,521,631;杂项 241,125。">
147
  </picture>
148
 
149
  139 个语料库,共 1,216,362,128 行。
 
176
  ## 怎样做成
177
 
178
  <picture class="not-prose">
179
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/steps-narrow-light.zh.svg">
180
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/steps-wide-light.zh.svg" alt="索引分四步做成。1,采集:采集器整批读取完整上游,每个文件或每条记录记一行,成为语料库。2,收录:检索在上游找到的作品和成组的语料库行,收录为材料集。3,检查:发布前核对许可证、个人信息、与基准测试的重叠和大小。4,发布:一个带有自己标签的提交,release-2026-11-11.20。">
181
  </picture>
182
  <picture class="not-prose">
183
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/steps-narrow-dark.zh.svg">
184
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-11.20/docs/figures/steps-wide-dark.zh.svg" alt="索引分四步做成。1,采集:采集器整批读取完整上游,每个文件或每条记录记一行,成为语料库。2,收录:检索在上游找到的作品和成组的语料库行,收录为材料集。3,检查:发布前核对许可证、个人信息、与基准测试的重叠和大小。4,发布:一个带有自己标签的提交,release-2026-11-11.20。">
185
  </picture>
186
 
187
  术语、每一步,以及每个数字怎样得出,都折叠在下面。
 
235
 
236
  | 层级 | 记录的大小 | 去重后大小 | 读过的行 | 比对过的行 | 一致的行 | 读过但未比对的行 | 没有记录大小的行 | 这些行的估计大小 |
237
  |---|---:|---:|---:|---:|---:|---:|---:|---:|
238
+ | 整个索引 | 4.45 PB | 4.11 PB | 34,384 | 33,506 | 33,377(99.6%) | 878:地址无法访问 449 行、文件已不存在 80 行、需要登录 329 行、服务器未给出大小 19 行、其他错误 1 行 | 312,300,338 | 202 TB,这是估计值(95% 范围 171 TB 到 245 TB),针对其中 215,326,513 行,依据读过的 14,161 行 |
239
  | 语料库 | 4.45 PB | 4.11 PB | 31,594 | 30,723 | 30,610(99.6%) | 871:地址无法访问 449 行、文件已不存在 80 行、需要登录 322 行、服务器未给出大小 19 行、其他错误 1 行 | 312,300,338 | 202 TB,这是估计值(95% 范围 171 TB 到 245 TB),针对其中 215,326,513 行,依据读过的 14,161 行 |
240
+ | 任务材料 | 2.88 GB | 2.82 GB | 2,790 | 2,783 | 2,767(99.4%) | 7:需要登录 7 行 | 0 | 无 |
241
 
242
  读过的行是核对在 `meta/sizes.parquet` 中读取的抽样行;
243
  比对过的行与记录的大小一致或不一致,读过但未比对的行没有取到它的文件。
 
246
 
247
  记录的大小是有大小的各行 `size_bytes` 之和(目录列 `bytes`)。核对读取每个数据文件中抽取的一组行(`meta/sizes.parquet`):有记录大小的行抽 20 行;没有记录大小的行按其所在来源的样本分摊,样本量按目前读到的大小的离散程度定,使其平均大小在 95% 置信度下误差不超过 25%(每个来源每轮最多 400 行,还没有读数的来源取 20 行);其比例是比对过的行(状态 `match`、`size_mismatch` 或 `hash_mismatch`)中与记录的大小一致的行所占的比例,没能比对的行(没有取到文件)按状态另行计数(`unreachable`、`gone`、`needs_auth`、`no_length`、`error`;目录列 `size_check`)。一个来源没有记录大小的行的估计(目录列 `unsized_rows` 和 `unsized_estimate`)是这些行的行数乘以读过的这类行的平均实测大小,并给出 95% BCa 自助法范围(`scipy.stats.bootstrap`,用其默认重抽样次数);读过的这类行少于 10 行的来源没有估计。在来源之上的层级,估计是其各来源估计之和,范围是各来源范围之和,比联合区间更宽。每个层级的数字都在 `meta/size_figures.json` 中。
248
 
249
+ 去重后内容是各行 `size_bytes` 之和,减去与另一行内容重复的那些行的大小(sha256 相同;没有 sha256 时 sha1 相同;再没有时 md5 和大小都相同;并且哈希范围相同;每种内容只保留一份,留在按注册顺序排在最前的来源中),在每次发布之后按它发布的数据文件重新计算,每个来源的值在目录列 `distinct_bytes` 中;同一来源也整个列出的存档,其成员的哈希与存档不同,所以不扣除。`tools/fetch.py` 读取你自己的选择时会去掉其中的重复。 这里的去重后大小是按发布 `release-2026-11-11.17` 的数据文件算出的。
250
 
251
  </details>
252
 
 
260
  下面每张表给出一个训练阶段能用的材料,在各条规则之前和按顺序每条规则之后计数,直到该阶段的计数;
261
  每张表的最后两行统计该阶段的计数中有��少允许训练(每个文件的训练许可为 yes 或 conditional),有多少标为带有个人数据。
262
 
263
+ 阶段字段(`meta/stage_fields.parquet`)是按发布 `release-2026-11-11.17` 的数据文件、在它发布之后算出的。
264
 
265
  #### 后训练(在用)
266
 
 
268
 
269
  | 步骤 | 材料集数 | 单元数 | 上游来源数 |
270
  |---|---:|---:|---:|
271
+ | 符合该阶段的条件,在各条规则之前 | 1,788 | 4,150 | 1,843 |
272
+ | 基准测试规则之后 | 1,787 | 4,149 | 1,842 |
273
+ | 未与基准测试清单比对的行的规则之后 | 1,787 | 4,149 | 1,842 |
274
+ | 去重之后:该阶段的计数 | 1,758 | 4,040 | 1,817 |
275
+ | 该阶段的计数中,允许训练的(yes 或 conditional) | 1,405 | 3,397 | 1,333 |
276
+ | 该阶段的计数中,标为带有个人数据的 | 357 | 631 | 400 |
277
 
278
  #### 中训练(暂定:只给条件和计数)
279
 
 
281
 
282
  | 步骤 | 行数 | 大小 |
283
  |---|---:|---:|
284
+ | 符合该阶段的条件,在各条规则之前 | 34,477,690 | 4.00 TB |
285
+ | 基准测试规则之后 | 34,468,408 | 4.00 TB |
286
+ | 未与基准测试清单比对的行的规则之后 | 34,468,408 | 4.00 TB |
287
+ | 去重之后:该阶段的计数 | 34,079,767 | 3.86 TB |
288
+ | 该阶段的计数中,允许训练的(yes 或 conditional) | 9,890,653 | 1.42 TB |
289
+ | 该阶段的计数中,标为带有个人数据的 | 6,102,344 | 341 GB |
290
 
291
  #### 预训练(暂定:只给条件和计数)
292
 
 
294
 
295
  | 步骤 | 行数 | 大小 |
296
  |---|---:|---:|
297
+ | 符合该阶段的条件,在各条规则之前 | 154,035,344 | 30.7 TB |
298
+ | 基准测试规则之后 | 154,034,327 | 30.7 TB |
299
+ | 未与基准测试清单比对的行的规则之后 | 154,034,327 | 30.7 TB |
300
+ | 去重之后:该阶段的计数 | 149,118,607 | 30.6 TB |
301
+ | 该阶段的计数中,允许训练的(yes 或 conditional) | 117,929,843 | 29.9 TB |
302
+ | 该阶段的计数中,标为带有个人数据的 | 40,455,308 | 676 GB |
303
 
304
  </details>
305
 
 
438
  <details>
439
  <summary>发布</summary>
440
 
441
+ - 最新发布:`release-2026-11-11.20`(其最新的 CHANGELOG 条目日期为 2026-11-11)。
442
  - 各次发布列在 [CHANGELOG.md](CHANGELOG.md) 中;
443
+ 每次发布是一个带有自己发布标签的提交(本次为 [`release-2026-11-11.20`](https://huggingface.co/datasets/Battam/training-data-foundation/tree/release-2026-11-11.20)),
444
  因此任何较早的状态都可以在它的标签处读到。
445
  - `meta/projection.json` 记录发布标签,以及生成本卡片的程序和模板的版本。
446
  - Croissant 元数据:[由 Hub 生成](https://huggingface.co/api/datasets/Battam/training-data-foundation/croissant);
 
481
  @misc{training_data_foundation,
482
  title = {training-data-foundation: an index of public materials for building agent environments},
483
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
484
+ note = {Release release-2026-11-11.20},
485
  year = {2026}
486
  }
487
  ```
campaign_files/2026-10-a-spreadsheets-with-answers-2/packt-textbook-companion-workbooks/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained.
10
 
11
- **Answers:** Each file with an answer has one answer file, at answer/<unit path>.json (the unit path plus .json). Reference result kind: key_formula_values.
12
-
13
  **Tools to open and check:** openpyxl (formulas and cached values); LibreOffice headless recalculation (soffice --headless --convert-to xlsx)
14
 
15
  **Licence (SPDX):** MIT
 
8
 
9
  **Self-contained:** Each file is self-contained.
10
 
 
 
11
  **Tools to open and check:** openpyxl (formulas and cached values); LibreOffice headless recalculation (soffice --headless --convert-to xlsx)
12
 
13
  **Licence (SPDX):** MIT
campaign_files/2026-10-b-disk-images-3/opentoall-2015-catz/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether the file needs other upstream files is not stated. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** Each file with an answer has one answer file, at answer/<unit path>.json (the unit path plus .json). Reference result kind: deleted_files.
12
-
13
  **Tools to open and check:** Open with tools for these file types: .bz2.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -20,5 +18,4 @@ Files:
20
 
21
  - `opentoall-ctf-2015/forensics/catz/d17c5ed999bc505ec80e64c714c6d965.tar.bz2`: ext4 image with one deleted hidden file /.cat.jpg (OpenToAll CTF 2015 'Catz').
22
  Origin: Organisers' challenge file, mirrored by the write-up archive.
23
- Reference result: deleted_files
24
  Dependencies: bunzip2 and tar or Python bz2 and tarfile to unpack; only needed to rerun the exercise, not to grade the answer
 
8
 
9
  **Self-contained:** Whether the file needs other upstream files is not stated. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Open with tools for these file types: .bz2.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
18
 
19
  - `opentoall-ctf-2015/forensics/catz/d17c5ed999bc505ec80e64c714c6d965.tar.bz2`: ext4 image with one deleted hidden file /.cat.jpg (OpenToAll CTF 2015 'Catz').
20
  Origin: Organisers' challenge file, mirrored by the write-up archive.
 
21
  Dependencies: bunzip2 and tar or Python bz2 and tarfile to unpack; only needed to rerun the exercise, not to grade the answer
campaign_files/2026-10-b-disk-images-3/opentoall-2015-gone/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether the file needs other upstream files is not stated. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** Each file with an answer has one answer file, at answer/<unit path>.json (the unit path plus .json). Reference result kind: corruption_desc.
12
-
13
  **Tools to open and check:** Open with tools for these file types: .bz2.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -20,5 +18,4 @@ Files:
20
 
21
  - `opentoall-ctf-2015/forensics/gone/gone.img.tar.bz2`: ext4 image with a corrupt primary superblock, backup intact (OpenToAll CTF 2015 'Gone').
22
  Origin: Organisers' challenge file, mirrored by the write-up archive.
23
- Reference result: corruption_desc
24
  Dependencies: bunzip2 and tar or Python bz2 and tarfile to unpack; only needed to rerun the exercise, not to grade the answer
 
8
 
9
  **Self-contained:** Whether the file needs other upstream files is not stated. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Open with tools for these file types: .bz2.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
18
 
19
  - `opentoall-ctf-2015/forensics/gone/gone.img.tar.bz2`: ext4 image with a corrupt primary superblock, backup intact (OpenToAll CTF 2015 'Gone').
20
  Origin: Organisers' challenge file, mirrored by the write-up archive.
 
21
  Dependencies: bunzip2 and tar or Python bz2 and tarfile to unpack; only needed to rerun the exercise, not to grade the answer
campaign_files/2026-10-c-repros-2/overset-grid-subcell-repro/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each unit needs other upstream files is not stated. 2 of 2 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** Each unit with an answer has one answer file, at answer/<package>/reported.json.
12
-
13
  **Tools to open and check:** Open with tools for these file types: .jl; .toml.
14
 
15
  **Licence (SPDX):** MIT
 
8
 
9
  **Self-contained:** Whether each unit needs other upstream files is not stated. 2 of 2 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Open with tools for these file types: .jl; .toml.
12
 
13
  **Licence (SPDX):** MIT
campaign_files/2026-10-g-solver-instances-2/or-library-common-due-date/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Proven optimum (objective value); best known upper bound (objective value), not marked optimal. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Program: parse the agent's job sequence (and optional start idle), schedule the jobs without overlap, compute sum of a(i) times earliness plus b(i) times tardiness against d = 69, require it at most 841; Program: parse the agent's job sequence (and optional start idle), schedule the jobs without overlap, compute sum of a(i) times earliness plus b(i) times tardiness against d = 130, require it at most 2,986 (the published upper bound; a lower value also passes).
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,14 +19,10 @@ Files:
21
  - `files/sch10.txt`: Common due date single machine earliness-tardiness, n=10, problem k=1 with h=0.6 (optimum 841).
22
  Format: plain text: number of problems (10), then per problem n and n lines of processing time p, earliness penalty a, tardiness penalty b (schinfo.html lines 43 to 48); the case is problem 1 with due date d = floor(116 x 0.6) = 69
23
  Origin: Biskup and Feldmann (2001) common due date benchmark, schinfo.html lines 30 to 60
24
- Reference result: proven optimum (objective value)
25
- Reference source: schinfo.html line 69: 'k = 1 116 1,936 1,025 841* 818*'; the asterisk marks optimal values; h = 0.6 column gives 841
26
  Check: Program: parse the agent's job sequence (and optional start idle), schedule the jobs without overlap, compute sum of a(i) times earliness plus b(i) times tardiness against d = 69, require it at most 841
27
  Companions (not included): schinfo.html
28
  - `files/sch20.txt`: Common due date single machine earliness-tardiness, n=20, problem k=1 with h=0.6 (best known upper bound 2,986).
29
  Format: plain text: number of problems (10), then per problem n and n lines of processing time p, earliness penalty a, tardiness penalty b (schinfo.html lines 43 to 48); the case is problem 1, SUM_P = 217 (computed from the file), due date d = floor(217 x 0.6) = 130
30
  Origin: Biskup and Feldmann (2001) common due date benchmark, schinfo.html lines 30 to 60
31
- Reference result: best known upper bound (objective value), not marked optimal
32
- Reference source: schinfo.html line 85: 'k = 1 4,431 3,066 2,986 2,986' under the header of line 83 (n=20, h = 0.2, 0.4, 0.6, 0.8); the 20-job table carries no optimality asterisk, so 2,986 is the best known upper bound
33
  Check: Program: parse the agent's job sequence (and optional start idle), schedule the jobs without overlap, compute sum of a(i) times earliness plus b(i) times tardiness against d = 130, require it at most 2,986 (the published upper bound; a lower value also passes)
34
  Companions (not included): schinfo.html
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Program: parse the agent's job sequence (and optional start idle), schedule the jobs without overlap, compute sum of a(i) times earliness plus b(i) times tardiness against d = 69, require it at most 841; Program: parse the agent's job sequence (and optional start idle), schedule the jobs without overlap, compute sum of a(i) times earliness plus b(i) times tardiness against d = 130, require it at most 2,986 (the published upper bound; a lower value also passes).
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `files/sch10.txt`: Common due date single machine earliness-tardiness, n=10, problem k=1 with h=0.6 (optimum 841).
20
  Format: plain text: number of problems (10), then per problem n and n lines of processing time p, earliness penalty a, tardiness penalty b (schinfo.html lines 43 to 48); the case is problem 1 with due date d = floor(116 x 0.6) = 69
21
  Origin: Biskup and Feldmann (2001) common due date benchmark, schinfo.html lines 30 to 60
 
 
22
  Check: Program: parse the agent's job sequence (and optional start idle), schedule the jobs without overlap, compute sum of a(i) times earliness plus b(i) times tardiness against d = 69, require it at most 841
23
  Companions (not included): schinfo.html
24
  - `files/sch20.txt`: Common due date single machine earliness-tardiness, n=20, problem k=1 with h=0.6 (best known upper bound 2,986).
25
  Format: plain text: number of problems (10), then per problem n and n lines of processing time p, earliness penalty a, tardiness penalty b (schinfo.html lines 43 to 48); the case is problem 1, SUM_P = 217 (computed from the file), due date d = floor(217 x 0.6) = 130
26
  Origin: Biskup and Feldmann (2001) common due date benchmark, schinfo.html lines 30 to 60
 
 
27
  Check: Program: parse the agent's job sequence (and optional start idle), schedule the jobs without overlap, compute sum of a(i) times earliness plus b(i) times tardiness against d = 130, require it at most 2,986 (the published upper bound; a lower value also passes)
28
  Companions (not included): schinfo.html
campaign_files/2026-10-g-solver-instances-2/or-library-hdtt/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 5 of 5 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Known optimum (zero clashes). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Program: parse the agent's period for each meeting in hdtt4list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0; Program: parse the agent's period for each meeting in hdtt5list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0; Program: parse the agent's period for each meeting in hdtt6list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0; Program: parse the agent's period for each meeting in hdtt7list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0; Program: parse the agent's period for each meeting in hdtt8list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,35 +19,25 @@ Files:
21
  - `files/hdtt4list.txt`: Hard timetabling hdtt4: 4 classes, 4 teachers, 4 rooms, 30 periods, 120 meetings; optimum zero clashes.
22
  Format: three plain text files: hdtt4list (requirements as English statements), hdtt4note (counts), hdtt4req (class by teacher requirement matrix), described in tableinfo.html lines 58 to 80
23
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
24
- Reference result: known optimum (zero clashes)
25
- Reference source: tableinfo.html lines 55 to 56: 'The optimal objective function for each of these problems is zero clashes'
26
  Check: Program: parse the agent's period for each meeting in hdtt4list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
27
  Companions (not included): files/hdtt4note.txt; files/hdtt4req.txt; tableinfo.html
28
  - `files/hdtt5list.txt`: Hard timetabling hdtt5: 5 classes, 5 teachers, 5 rooms, 30 periods, 150 meetings; optimum zero clashes.
29
  Format: three plain text files: hdtt5list (requirements as English statements), hdtt5note (counts), hdtt5req (class by teacher requirement matrix), format described in tableinfo.html lines 58 to 80
30
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
31
- Reference result: known optimum (zero clashes)
32
- Reference source: tableinfo.html lines 55 to 56: 'The optimal objective function for each of these problems is zero clashes' (line 51 names hdtt5 among them)
33
  Check: Program: parse the agent's period for each meeting in hdtt5list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
34
  Companions (not included): files/hdtt5note.txt; files/hdtt5req.txt; tableinfo.html
35
  - `files/hdtt6list.txt`: Hard timetabling hdtt6: 6 classes, 6 teachers, 6 rooms, 30 periods, 180 meetings; optimum zero clashes.
36
  Format: three plain text files: hdtt6list (requirements as English statements), hdtt6note (counts), hdtt6req (class by teacher requirement matrix per room), described in tableinfo.html lines 58 to 80
37
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
38
- Reference result: known optimum (zero clashes)
39
- Reference source: tableinfo.html lines 55 to 56: 'The optimal objective function for each of these problems is zero clashes'
40
  Check: Program: parse the agent's period for each meeting in hdtt6list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
41
  Companions (not included): files/hdtt6note.txt; files/hdtt6req.txt; tableinfo.html
42
  - `files/hdtt7list.txt`: Hard timetabling hdtt7: 7 classes, 7 teachers, 7 rooms, 30 periods, 210 meetings; optimum zero clashes.
43
  Format: three plain text files: hdtt7list (requirements as English statements), hdtt7note (counts), hdtt7req (class by teacher requirement matrix per room), described in tableinfo.html lines 58 to 80
44
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
45
- Reference result: known optimum (zero clashes)
46
- Reference source: tableinfo.html lines 55 to 56: 'The optimal objective function for each of these problems is zero clashes'
47
  Check: Program: parse the agent's period for each meeting in hdtt7list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
48
  Companions (not included): files/hdtt7note.txt; files/hdtt7req.txt; tableinfo.html
49
  - `files/hdtt8list.txt`: Hard timetabling hdtt8: 8 classes, 8 teachers, 8 rooms, 30 periods, 240 meetings; optimum zero clashes.
50
  Format: three plain text files: hdtt8list (requirements as English statements), hdtt8note (counts), hdtt8req (class by teacher requirement matrix per room), described in tableinfo.html lines 58 to 80
51
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
52
- Reference result: known optimum (zero clashes)
53
- Reference source: tableinfo.html lines 55 to 56: 'The optimal objective function for each of these problems is zero clashes'
54
  Check: Program: parse the agent's period for each meeting in hdtt8list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
55
  Companions (not included): files/hdtt8note.txt; files/hdtt8req.txt; tableinfo.html
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 5 of 5 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Program: parse the agent's period for each meeting in hdtt4list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0; Program: parse the agent's period for each meeting in hdtt5list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0; Program: parse the agent's period for each meeting in hdtt6list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0; Program: parse the agent's period for each meeting in hdtt7list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0; Program: parse the agent's period for each meeting in hdtt8list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `files/hdtt4list.txt`: Hard timetabling hdtt4: 4 classes, 4 teachers, 4 rooms, 30 periods, 120 meetings; optimum zero clashes.
20
  Format: three plain text files: hdtt4list (requirements as English statements), hdtt4note (counts), hdtt4req (class by teacher requirement matrix), described in tableinfo.html lines 58 to 80
21
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
 
 
22
  Check: Program: parse the agent's period for each meeting in hdtt4list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
23
  Companions (not included): files/hdtt4note.txt; files/hdtt4req.txt; tableinfo.html
24
  - `files/hdtt5list.txt`: Hard timetabling hdtt5: 5 classes, 5 teachers, 5 rooms, 30 periods, 150 meetings; optimum zero clashes.
25
  Format: three plain text files: hdtt5list (requirements as English statements), hdtt5note (counts), hdtt5req (class by teacher requirement matrix), format described in tableinfo.html lines 58 to 80
26
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
 
 
27
  Check: Program: parse the agent's period for each meeting in hdtt5list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
28
  Companions (not included): files/hdtt5note.txt; files/hdtt5req.txt; tableinfo.html
29
  - `files/hdtt6list.txt`: Hard timetabling hdtt6: 6 classes, 6 teachers, 6 rooms, 30 periods, 180 meetings; optimum zero clashes.
30
  Format: three plain text files: hdtt6list (requirements as English statements), hdtt6note (counts), hdtt6req (class by teacher requirement matrix per room), described in tableinfo.html lines 58 to 80
31
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
 
 
32
  Check: Program: parse the agent's period for each meeting in hdtt6list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
33
  Companions (not included): files/hdtt6note.txt; files/hdtt6req.txt; tableinfo.html
34
  - `files/hdtt7list.txt`: Hard timetabling hdtt7: 7 classes, 7 teachers, 7 rooms, 30 periods, 210 meetings; optimum zero clashes.
35
  Format: three plain text files: hdtt7list (requirements as English statements), hdtt7note (counts), hdtt7req (class by teacher requirement matrix per room), described in tableinfo.html lines 58 to 80
36
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
 
 
37
  Check: Program: parse the agent's period for each meeting in hdtt7list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
38
  Companions (not included): files/hdtt7note.txt; files/hdtt7req.txt; tableinfo.html
39
  - `files/hdtt8list.txt`: Hard timetabling hdtt8: 8 classes, 8 teachers, 8 rooms, 30 periods, 240 meetings; optimum zero clashes.
40
  Format: three plain text files: hdtt8list (requirements as English statements), hdtt8note (counts), hdtt8req (class by teacher requirement matrix per room), described in tableinfo.html lines 58 to 80
41
  Origin: Abramson and Dang hard timetabling problems, OR-Library tableinfo
 
 
42
  Check: Program: parse the agent's period for each meeting in hdtt8list order, require every meeting placed in a period 1 to 30 and zero clashes (no class, teacher or room twice in a period); pass if clashes = 0
43
  Companions (not included): files/hdtt8note.txt; files/hdtt8req.txt; tableinfo.html
campaign_files/2026-10-g-solver-instances-2/or-library-sch/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Best known upper bound. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Program: read the agent's start times (or sequence plus start of the first job) for problem 1, require no overlap, compute d = floor(SUM_P * 0.6) = 329 and the weighted earliness plus tardiness with the a and b of each job; pass when the value is at most 17990.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `files/sch50.txt`: Common due date single machine earliness and tardiness, 50 jobs, problem k = 1 at h = 0.6; best known 17990.
22
  Format: plain integers: number of problems, then per problem n and n lines of (p, a, b) as described in schinfo.html lines 43 to 54; the common due date is d = floor(SUM_P * h)
23
  Origin: Biskup and Feldmann common due date instances, OR-Library schinfo
24
- Reference result: best known upper bound
25
- Reference source: schinfo.html line 101, row 'k = 1', column h = 0.6 (header line 99): 17,990
26
  Check: Program: read the agent's start times (or sequence plus start of the first job) for problem 1, require no overlap, compute d = floor(SUM_P * 0.6) = 329 and the weighted earliness plus tardiness with the a and b of each job; pass when the value is at most 17990
27
  Companions (not included): schinfo.html
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Program: read the agent's start times (or sequence plus start of the first job) for problem 1, require no overlap, compute d = floor(SUM_P * 0.6) = 329 and the weighted earliness plus tardiness with the a and b of each job; pass when the value is at most 17990.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `files/sch50.txt`: Common due date single machine earliness and tardiness, 50 jobs, problem k = 1 at h = 0.6; best known 17990.
20
  Format: plain integers: number of problems, then per problem n and n lines of (p, a, b) as described in schinfo.html lines 43 to 54; the common due date is d = floor(SUM_P * h)
21
  Origin: Biskup and Feldmann common due date instances, OR-Library schinfo
 
 
22
  Check: Program: read the agent's start times (or sequence plus start of the first job) for problem 1, require no overlap, compute d = floor(SUM_P * 0.6) = 329 and the weighted earliness plus tardiness with the a and b of each job; pass when the value is at most 17990
23
  Companions (not included): schinfo.html
campaign_files/2026-10-g-solver-instances-2/or-library-wt/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Optimum; best known value. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Program: read the agent's job sequence for instance 1, require a permutation of the 50 jobs, compute completion times without idle time and the total weighted tardiness from the p, w, d of instance 1; pass when the value is at most 2134; Program: read the agent's job sequence for instance 1, require a permutation of the 100 jobs, compute completion times without idle time and the total weighted tardiness from the p, w, d of instance 1; pass when the value is at most 5988.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,14 +19,10 @@ Files:
21
  - `files/wt50.txt`: Single machine total weighted tardiness, 50 jobs, instance 1 of 125; optimum 2134.
22
  Format: plain integers, 125 instances one after another, each listing 50 processing times, then 50 weights, then 50 due dates (wtinfo.html lines 57 to 61); the task uses instance 1 (the first 150 integers)
23
  Origin: Crauwels, Potts and Van Wassenhove weighted tardiness instances, OR-Library wtinfo
24
- Reference result: optimum
25
- Reference source: files/wtopt50.txt line 1 value 2134 for instance 1; wtinfo.html lines 69 to 72: optima known for 115 of the 50-job instances, unsolved are 11, 12, 14, 19, 36, 44, 66, 87, 88 and 111, so instance 1 is a proven optimum
26
  Check: Program: read the agent's job sequence for instance 1, require a permutation of the 50 jobs, compute completion times without idle time and the total weighted tardiness from the p, w, d of instance 1; pass when the value is at most 2134
27
  Companions (not included): files/wtopt50.txt; wtinfo.html
28
  - `files/wt100.txt`: Single machine total weighted tardiness, 100 jobs, instance 1 of 125; best known value 5988.
29
  Format: plain integers, 125 instances one after another, each listing 100 processing times, then 100 weights, then 100 due dates (wtinfo.html lines 57 to 61); the task uses instance 1 (the first 300 integers)
30
  Origin: Crauwels, Potts and Van Wassenhove weighted tardiness instances, OR-Library wtinfo
31
- Reference result: best known value
32
- Reference source: files/wtbest100b.txt line 1 value 5988 for instance 1; wtinfo.html line 92: 'to date are given in the file wtbest100b'
33
  Check: Program: read the agent's job sequence for instance 1, require a permutation of the 100 jobs, compute completion times without idle time and the total weighted tardiness from the p, w, d of instance 1; pass when the value is at most 5988
34
  Companions (not included): files/wtbest100b.txt; wtinfo.html
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Program: read the agent's job sequence for instance 1, require a permutation of the 50 jobs, compute completion times without idle time and the total weighted tardiness from the p, w, d of instance 1; pass when the value is at most 2134; Program: read the agent's job sequence for instance 1, require a permutation of the 100 jobs, compute completion times without idle time and the total weighted tardiness from the p, w, d of instance 1; pass when the value is at most 5988.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `files/wt50.txt`: Single machine total weighted tardiness, 50 jobs, instance 1 of 125; optimum 2134.
20
  Format: plain integers, 125 instances one after another, each listing 50 processing times, then 50 weights, then 50 due dates (wtinfo.html lines 57 to 61); the task uses instance 1 (the first 150 integers)
21
  Origin: Crauwels, Potts and Van Wassenhove weighted tardiness instances, OR-Library wtinfo
 
 
22
  Check: Program: read the agent's job sequence for instance 1, require a permutation of the 50 jobs, compute completion times without idle time and the total weighted tardiness from the p, w, d of instance 1; pass when the value is at most 2134
23
  Companions (not included): files/wtopt50.txt; wtinfo.html
24
  - `files/wt100.txt`: Single machine total weighted tardiness, 100 jobs, instance 1 of 125; best known value 5988.
25
  Format: plain integers, 125 instances one after another, each listing 100 processing times, then 100 weights, then 100 due dates (wtinfo.html lines 57 to 61); the task uses instance 1 (the first 300 integers)
26
  Origin: Crauwels, Potts and Van Wassenhove weighted tardiness instances, OR-Library wtinfo
 
 
27
  Check: Program: read the agent's job sequence for instance 1, require a permutation of the 100 jobs, compute completion times without idle time and the total weighted tardiness from the p, w, d of instance 1; pass when the value is at most 5988
28
  Companions (not included): files/wtbest100b.txt; wtinfo.html
campaign_files/2026-10-g-solver-instances-2/or-library/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Optimum for 124 instances, best known for instance 19. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Program: the agent returns a job sequence; compute the total weighted tardiness and require it at most the wtopt40.txt value for that instance.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `orlib/files/wt40.txt`: Single machine total weighted tardiness, 125 instances with 40 jobs, with optimal or best values.
22
  Format: whitespace separated integers: per instance 40 processing times, 40 weights, 40 due dates (wtinfo.html line 63); 125 cases in one file; wtopt40.txt one value per line in the same order
23
  Origin: Crauwels, Potts and Van Wassenhove instances distributed by OR-Library
24
- Reference result: optimum for 124 instances, best known for instance 19
25
- Reference source: wtopt40.txt line k gives instance k (line 1: 913); wtinfo.html line 69 "Optimal values of solutions are available for 124 and 115 of the 40 and 50" and line 71 "The unsolved 40 job problem is number 19"
26
  Check: Program: the agent returns a job sequence; compute the total weighted tardiness and require it at most the wtopt40.txt value for that instance
27
  Companions (not included): orlib/files/wtopt40.txt
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Program: the agent returns a job sequence; compute the total weighted tardiness and require it at most the wtopt40.txt value for that instance.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `orlib/files/wt40.txt`: Single machine total weighted tardiness, 125 instances with 40 jobs, with optimal or best values.
20
  Format: whitespace separated integers: per instance 40 processing times, 40 weights, 40 due dates (wtinfo.html line 63); 125 cases in one file; wtopt40.txt one value per line in the same order
21
  Origin: Crauwels, Potts and Van Wassenhove instances distributed by OR-Library
 
 
22
  Check: Program: the agent returns a job sequence; compute the total weighted tardiness and require it at most the wtopt40.txt value for that instance
23
  Companions (not included): orlib/files/wtopt40.txt
campaign_files/2026-10-g-solver-instances-2/ors-machine-scheduling/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Optimal maximum lateness (lower bound equals best found). Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Check: Program: simulate the agent's sequence on one machine, no job starts before its release time, compute max over jobs of completion minus due date; pass when it equals 30.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -22,7 +20,5 @@ Units:
22
  Read out of the archive: https://www.projectmanagement.ugent.be/sites/default/files/files/machinescheduling/Set1.zip
23
  Format: text: first line job count, then one line of three integers per job (processing time, release date, due date; column order as the OR&S papers document it, to be confirmed by the checker author); one case = one instance; answer = a job sequence and its maximum lateness
24
  Origin: generated by Sels and Vanhoucke (2011, 2012) under a controlled design. Standard benchmark suite or competition instance as published by the library, recorded tool_fixture.
25
- Reference result: optimal maximum lateness (lower bound equals best found)
26
- Reference source: BKR_SET_I.xlsx sheet "SET I - FULL" row 872: A872 "100_Dependent0870", G872 LB 30, N872 GALS 30, P872 MultiS 30
27
  Check: Program: simulate the agent's sequence on one machine, no job starts before its release time, compute max over jobs of completion minus due date; pass when it equals 30
28
  Companions (not included): machinescheduling/SET_I/BKR_SET_I.xlsx
 
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Program: simulate the agent's sequence on one machine, no job starts before its release time, compute max over jobs of completion minus due date; pass when it equals 30.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
20
  Read out of the archive: https://www.projectmanagement.ugent.be/sites/default/files/files/machinescheduling/Set1.zip
21
  Format: text: first line job count, then one line of three integers per job (processing time, release date, due date; column order as the OR&S papers document it, to be confirmed by the checker author); one case = one instance; answer = a job sequence and its maximum lateness
22
  Origin: generated by Sels and Vanhoucke (2011, 2012) under a controlled design. Standard benchmark suite or competition instance as published by the library, recorded tool_fixture.
 
 
23
  Check: Program: simulate the agent's sequence on one machine, no job starts before its release time, compute max over jobs of completion minus due date; pass when it equals 30
24
  Companions (not included): machinescheduling/SET_I/BKR_SET_I.xlsx
campaign_files/2026-10-g-solver-instances-2/pace-2017-treewidth-bonus/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 5 of 5 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Published optimal value (treewidth) with a witness tree decomposition. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Check: Width = (largest bag size) - 1; it passes when feasible and width <= 13. The published value is optimal, so a pass means width equals 13; width = (largest bag size) - 1; it passes when feasible and width <= 11. The published value is optimal, so a pass means width equals 11; width = (largest bag size) - 1; it passes when feasible and width <= 14. The published value is optimal, so a pass means width equals 14; width = (largest bag size) - 1; it passes when feasible and width <= 22. The published value is optimal, so a pass means width equals 22.
14
 
15
  **Licence (SPDX):** CC0-1.0
@@ -22,8 +20,6 @@ Units:
22
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/GTFS_VBB_EndeApr_Dez2016.zip_train+metro+tram_9.gr.xz
23
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/GTFS_VBB_EndeApr_Dez2016.zip_train+metro+tram_9.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
24
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
25
- Reference result: published optimal value (treewidth) with a witness tree decomposition
26
- Reference source: instances.csv line 88: "GTFS_VBB_EndeApr_Dez2016.zip_train+metro+tram_9.gr.xz , 13 , 314.76" (columns filename, treewidth, time; header line 1); td/GTFS_VBB_EndeApr_Dez2016.zip_train+metro+tram_9.td.xz unpacked line 1: "s td 126 14 143" (max bag size 14, width 13, agrees with the csv); README.md: "[td/](td/) contains the optimal tree decompositions (if known)." and "[instances.csv](instances.csv) lists the instances, the treewidth (if known), and the time required to compute the treewidth." (9 of 100 instances have "?" and no td; none of those is taken).
27
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 13. The published value is optimal, so a pass means width equals 13
28
  Companions (not included): td/GTFS_VBB_EndeApr_Dez2016.zip_train+metro+tram_9.td.xz; instances.csv; README.md; LICENSE
29
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -31,8 +27,6 @@ Units:
31
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/Promedas_21_9.gr.xz
32
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/Promedas_21_9.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
33
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
34
- Reference result: published optimal value (treewidth) with a witness tree decomposition
35
- Reference source: instances.csv line 61: "Promedas_21_9.gr.xz , 11 , 3059.90" (columns filename, treewidth, time; header line 1); td/Promedas_21_9.td.xz unpacked line 1: "s td 209 12 253" (max bag size 12, width 11, agrees with the csv); README.md: "[td/](td/) contains the optimal tree decompositions (if known)." and "[instances.csv](instances.csv) lists the instances, the treewidth (if known), and the time required to compute the treewidth." (9 of 100 instances have "?" and no td; none of those is taken).
36
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 11. The published value is optimal, so a pass means width equals 11
37
  Companions (not included): td/Promedas_21_9.td.xz; instances.csv; README.md; LICENSE
38
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -40,8 +34,6 @@ Units:
40
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/6s151.gaifman_3.gr.xz
41
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/6s151.gaifman_3.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
42
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
43
- Reference result: published optimal value (treewidth) with a witness tree decomposition
44
- Reference source: instances.csv line 65: "6s151.gaifman_3.gr.xz , 14 , 2209.71" (columns filename, treewidth, time; header line 1); td/6s151.gaifman_3.td.xz unpacked line 1: "s td 228 15 253" (max bag size 15, width 14, agrees with the csv); README.md: "[td/](td/) contains the optimal tree decompositions (if known)." and "[instances.csv](instances.csv) lists the instances, the treewidth (if known), and the time required to compute the treewidth." (9 of 100 instances have "?" and no td; none of those is taken).
45
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 14. The published value is optimal, so a pass means width equals 14
46
  Companions (not included): td/6s151.gaifman_3.td.xz; instances.csv; README.md; LICENSE
47
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -49,8 +41,6 @@ Units:
49
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/smtlib-qfbv-aigs-lfsr_004_127_112-tseitin.gaifman_6.gr.xz
50
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/smtlib-qfbv-aigs-lfsr_004_127_112-tseitin.gaifman_6.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
51
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
52
- Reference result: published optimal value (treewidth) with a witness tree decomposition
53
- Reference source: instances.csv line 45: "smtlib-qfbv-aigs-lfsr_004_127_112-tseitin.gaifman_6.gr.xz , 13 , 6839.28" (columns filename, treewidth, time; header line 1); td/smtlib-qfbv-aigs-lfsr_004_127_112-tseitin.gaifman_6.td.xz unpacked line 1: "s td 220 14 316" (max bag size 14, width 13, agrees with the csv); README.md: "[td/](td/) contains the optimal tree decompositions (if known)." and "[instances.csv](instances.csv) lists the instances, the treewidth (if known), and the time required to compute the treewidth." (9 of 100 instances have "?" and no td; none of those is taken).
54
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 13. The published value is optimal, so a pass means width equals 13
55
  Companions (not included): td/smtlib-qfbv-aigs-lfsr_004_127_112-tseitin.gaifman_6.td.xz; instances.csv; README.md; LICENSE
56
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -58,8 +48,6 @@ Units:
58
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/SAT_dat.k80-24_1_rule_1.gaifman_3.gr.xz
59
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/SAT_dat.k80-24_1_rule_1.gaifman_3.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
60
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
61
- Reference result: published optimal value (treewidth) with a witness tree decomposition
62
- Reference source: instances.csv line 47: "SAT_dat.k80-24_1_rule_1.gaifman_3.gr.xz , 22 , 6615.95" (columns filename, treewidth, time; header line 1); td/SAT_dat.k80-24_1_rule_1.gaifman_3.td.xz unpacked line 1: "s td 99 23 130" (max bag size 23, width 22, agrees with the csv); README.md: "[td/](td/) contains the optimal tree decompositions (if known)." and "[instances.csv](instances.csv) lists the instances, the treewidth (if known), and the time required to compute the treewidth." (9 of 100 instances have "?" and no td; none of those is taken).
63
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 22. The published value is optimal, so a pass means width equals 22
64
  Companions (not included): td/SAT_dat.k80-24_1_rule_1.gaifman_3.td.xz; instances.csv; README.md; LICENSE
65
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 5 of 5 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Width = (largest bag size) - 1; it passes when feasible and width <= 13. The published value is optimal, so a pass means width equals 13; width = (largest bag size) - 1; it passes when feasible and width <= 11. The published value is optimal, so a pass means width equals 11; width = (largest bag size) - 1; it passes when feasible and width <= 14. The published value is optimal, so a pass means width equals 14; width = (largest bag size) - 1; it passes when feasible and width <= 22. The published value is optimal, so a pass means width equals 22.
12
 
13
  **Licence (SPDX):** CC0-1.0
 
20
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/GTFS_VBB_EndeApr_Dez2016.zip_train+metro+tram_9.gr.xz
21
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/GTFS_VBB_EndeApr_Dez2016.zip_train+metro+tram_9.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
22
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
 
 
23
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 13. The published value is optimal, so a pass means width equals 13
24
  Companions (not included): td/GTFS_VBB_EndeApr_Dez2016.zip_train+metro+tram_9.td.xz; instances.csv; README.md; LICENSE
25
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
27
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/Promedas_21_9.gr.xz
28
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/Promedas_21_9.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
29
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
 
 
30
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 11. The published value is optimal, so a pass means width equals 11
31
  Companions (not included): td/Promedas_21_9.td.xz; instances.csv; README.md; LICENSE
32
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
34
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/6s151.gaifman_3.gr.xz
35
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/6s151.gaifman_3.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
36
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
 
 
37
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 14. The published value is optimal, so a pass means width equals 14
38
  Companions (not included): td/6s151.gaifman_3.td.xz; instances.csv; README.md; LICENSE
39
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
41
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/smtlib-qfbv-aigs-lfsr_004_127_112-tseitin.gaifman_6.gr.xz
42
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/smtlib-qfbv-aigs-lfsr_004_127_112-tseitin.gaifman_6.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
43
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
 
 
44
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 13. The published value is optimal, so a pass means width equals 13
45
  Companions (not included): td/smtlib-qfbv-aigs-lfsr_004_127_112-tseitin.gaifman_6.td.xz; instances.csv; README.md; LICENSE
46
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
48
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-bonus-instances/a6cd8149a0cdba96f99bc578c99cd0459f6e4ec8/gr/SAT_dat.k80-24_1_rule_1.gaifman_3.gr.xz
49
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line); unpacks with Python lzma; companion td/SAT_dat.k80-24_1_rule_1.gaifman_3.td.xz holds the published optimal tree decomposition in .td format and instances.csv gives the treewidth; one file is one case
50
  Origin: PACE 2017 bonus treewidth instance built by the organisers after the challenge to be harder than the exact set (README line 3); the file name names the original collection (GTFS public transport graphs of Berlin VBB, UAI Promedas medical diagnosis networks, Gaifman graphs of SAT and SMT-LIB instances), preprocessed by the organisers
 
 
51
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 22. The published value is optimal, so a pass means width equals 22
52
  Companions (not included): td/SAT_dat.k80-24_1_rule_1.gaifman_3.td.xz; instances.csv; README.md; LICENSE
53
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
campaign_files/2026-10-g-solver-instances-2/pace-2017-treewidth-exact/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 10 of 10 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Published optimal value (treewidth) with a witness tree decomposition. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Check: Width = (largest bag size) - 1; it passes when feasible and width <= 10. Since the published decomposition is optimal, a pass means width equals 10; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 9. Since the published decomposition is optimal, a pass means width equals 9; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 22. Since the published decomposition is optimal, a pass means width equals 22; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 6. Since the published decomposition is optimal, a pass means width equals 6; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 540. Since the published decomposition is optimal, a pass means width equals 540; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 34. Since the published decomposition is optimal, a pass means width equals 34; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 12. Since the published decomposition is optimal, a pass means width equals 12; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 18. Since the published decomposition is optimal, a pass means width equals 18; a lower width would signal a checker or data error.
14
 
15
  **Licence (SPDX):** CC0-1.0
@@ -22,8 +20,6 @@ Units:
22
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex001.gr.xz
23
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex001.td.xz holds the published optimal tree decomposition in .td format; one file is one case
24
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 1: "ex001.gr task2.5_output_BFS2-7-8-9_clean/SAT-instances-gaifman/app16-gaifman/countbitsarray04_32.gaifman_8.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
25
- Reference result: published optimal value (treewidth) with a witness tree decomposition
26
- Reference source: td/exact/ex001.td.xz, unpacked line 1: "s td 188 11 262" (max bag size 11, so width 10); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
27
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 10. Since the published decomposition is optimal, a pass means width equals 10; a lower width would signal a checker or data error
28
  Companions (not included): td/exact/ex001.td.xz; README.md; LICENSE
29
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -31,8 +27,6 @@ Units:
31
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex021.gr.xz
32
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex021.td.xz holds the published optimal tree decomposition in .td format; one file is one case
33
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 21: "ex021.gr task2.5_output_BFS-1-2-3-4-5-6_clean/task2_output/MMAP_UAI2014/Promedas_35_6.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
34
- Reference result: published optimal value (treewidth) with a witness tree decomposition
35
- Reference source: td/exact/ex021.td.xz, unpacked line 1: "s td 258 10 318" (max bag size 10, so width 9); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
36
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 9. Since the published decomposition is optimal, a pass means width equals 9; a lower width would signal a checker or data error
37
  Companions (not included): td/exact/ex021.td.xz; README.md; LICENSE
38
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -40,8 +34,6 @@ Units:
40
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex041.gr.xz
41
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex041.td.xz holds the published optimal tree decomposition in .td format; one file is one case
42
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 41: "ex041.gr task2.5_output_BFS-7-8-9_clean/task2_output/MAR_prob/Promedus_27_9.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
43
- Reference result: published optimal value (treewidth) with a witness tree decomposition
44
- Reference source: td/exact/ex041.td.xz, unpacked line 1: "s td 169 10 205" (max bag size 10, so width 9); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
45
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 9. Since the published decomposition is optimal, a pass means width equals 9; a lower width would signal a checker or data error
46
  Companions (not included): td/exact/ex041.td.xz; README.md; LICENSE
47
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -49,8 +41,6 @@ Units:
49
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex061.gr.xz
50
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex061.td.xz holds the published optimal tree decomposition in .td format; one file is one case
51
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 61: "ex061.gr task2.5_output_BFS-1-2-3-4-5-6_clean/task2_output/PR_prob/logistics.a.cnf_2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
52
- Reference result: published optimal value (treewidth) with a witness tree decomposition
53
- Reference source: td/exact/ex061.td.xz, unpacked line 1: "s td 115 23 158" (max bag size 23, so width 22); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
54
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 22. Since the published decomposition is optimal, a pass means width equals 22; a lower width would signal a checker or data error
55
  Companions (not included): td/exact/ex061.td.xz; README.md; LICENSE
56
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -58,8 +48,6 @@ Units:
58
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex081.gr.xz
59
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex081.td.xz holds the published optimal tree decomposition in .td format; one file is one case
60
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 81: "ex081.gr task2.5_output_BFS-1-2-3-4-5-6_clean/SAT-instances-gaifman/app16-gaifman/div2.c.20.smt2-stp212.gaifman_2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
61
- Reference result: published optimal value (treewidth) with a witness tree decomposition
62
- Reference source: td/exact/ex081.td.xz, unpacked line 1: "s td 169 7 188" (max bag size 7, so width 6); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
63
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 6. Since the published decomposition is optimal, a pass means width equals 6; a lower width would signal a checker or data error
64
  Companions (not included): td/exact/ex081.td.xz; README.md; LICENSE
65
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -67,8 +55,6 @@ Units:
67
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex101.gr.xz
68
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex101.td.xz holds the published optimal tree decomposition in .td format; one file is one case
69
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 101: "ex101.gr task2.5_output_BFS-1-2-3-4-5-6_clean/SAT-instances-gaifman/sc14-app-gaifman/grid-strips-grid-y-3.035-NOTKNOWN.gaifman_1.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
70
- Reference result: published optimal value (treewidth) with a witness tree decomposition
71
- Reference source: td/exact/ex101.td.xz, unpacked line 1: "s td 10 541 1038" (max bag size 541, so width 540); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
72
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 540. Since the published decomposition is optimal, a pass means width equals 540; a lower width would signal a checker or data error
73
  Companions (not included): td/exact/ex101.td.xz; README.md; LICENSE
74
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -76,8 +62,6 @@ Units:
76
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex121.gr.xz
77
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex121.td.xz holds the published optimal tree decomposition in .td format; one file is one case
78
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 121: "ex121.gr selected_C2_C3/jgiraldezlevy.2200.9086.08.40.83.gaifman_2_C2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
79
- Reference result: published optimal value (treewidth) with a witness tree decomposition
80
- Reference source: td/exact/ex121.td.xz, unpacked line 1: "s td 129 35 204" (max bag size 35, so width 34); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
81
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 34. Since the published decomposition is optimal, a pass means width equals 34; a lower width would signal a checker or data error
82
  Companions (not included): td/exact/ex121.td.xz; README.md; LICENSE
83
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -85,8 +69,6 @@ Units:
85
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex141.gr.xz
86
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex141.td.xz holds the published optimal tree decomposition in .td format; one file is one case
87
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 141: "ex141.gr selected_C2_C3/modgen-n200-m90860q08c40-14808.gaifman_2_C2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
88
- Reference result: published optimal value (treewidth) with a witness tree decomposition
89
- Reference source: td/exact/ex141.td.xz, unpacked line 1: "s td 144 35 226" (max bag size 35, so width 34); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
90
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 34. Since the published decomposition is optimal, a pass means width equals 34; a lower width would signal a checker or data error
91
  Companions (not included): td/exact/ex141.td.xz; README.md; LICENSE
92
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -94,8 +76,6 @@ Units:
94
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex161.gr.xz
95
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex161.td.xz holds the published optimal tree decomposition in .td format; one file is one case
96
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 161: "ex161.gr selected_C2_C3/aaai10-planning-ipc5-TPP-21-step11.gaifman_1_C29_C3.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
97
- Reference result: published optimal value (treewidth) with a witness tree decomposition
98
- Reference source: td/exact/ex161.td.xz, unpacked line 1: "s td 672 13 1046" (max bag size 13, so width 12); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
99
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 12. Since the published decomposition is optimal, a pass means width equals 12; a lower width would signal a checker or data error
100
  Companions (not included): td/exact/ex161.td.xz; README.md; LICENSE
101
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
@@ -103,8 +83,6 @@ Units:
103
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex181.gr.xz
104
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex181.td.xz holds the published optimal tree decomposition in .td format; one file is one case
105
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 181: "ex181.gr task2.5_output_BFS-1-2-3-4-5-6_clean/SAT-instances-gaifman/sr15bench-gaifman/mrpp_8x8#22_10.gaifman_2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
106
- Reference result: published optimal value (treewidth) with a witness tree decomposition
107
- Reference source: td/exact/ex181.td.xz, unpacked line 1: "s td 74 19 109" (max bag size 19, so width 18); README.md line 9: "[td/exact/](td/exact/) contains optimal tree decompositions for the _exact_ instances."; the PACE 2017 report (LIPIcs.IPEC.2017.30, README line 3) is the reference for the exact track.
108
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 18. Since the published decomposition is optimal, a pass means width equals 18; a lower width would signal a checker or data error
109
  Companions (not included): td/exact/ex181.td.xz; README.md; LICENSE
110
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 10 of 10 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Width = (largest bag size) - 1; it passes when feasible and width <= 10. Since the published decomposition is optimal, a pass means width equals 10; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 9. Since the published decomposition is optimal, a pass means width equals 9; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 22. Since the published decomposition is optimal, a pass means width equals 22; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 6. Since the published decomposition is optimal, a pass means width equals 6; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 540. Since the published decomposition is optimal, a pass means width equals 540; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 34. Since the published decomposition is optimal, a pass means width equals 34; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 12. Since the published decomposition is optimal, a pass means width equals 12; a lower width would signal a checker or data error; width = (largest bag size) - 1; it passes when feasible and width <= 18. Since the published decomposition is optimal, a pass means width equals 18; a lower width would signal a checker or data error.
12
 
13
  **Licence (SPDX):** CC0-1.0
 
20
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex001.gr.xz
21
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex001.td.xz holds the published optimal tree decomposition in .td format; one file is one case
22
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 1: "ex001.gr task2.5_output_BFS2-7-8-9_clean/SAT-instances-gaifman/app16-gaifman/countbitsarray04_32.gaifman_8.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
23
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 10. Since the published decomposition is optimal, a pass means width equals 10; a lower width would signal a checker or data error
24
  Companions (not included): td/exact/ex001.td.xz; README.md; LICENSE
25
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
27
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex021.gr.xz
28
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex021.td.xz holds the published optimal tree decomposition in .td format; one file is one case
29
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 21: "ex021.gr task2.5_output_BFS-1-2-3-4-5-6_clean/task2_output/MMAP_UAI2014/Promedas_35_6.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
30
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 9. Since the published decomposition is optimal, a pass means width equals 9; a lower width would signal a checker or data error
31
  Companions (not included): td/exact/ex021.td.xz; README.md; LICENSE
32
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
34
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex041.gr.xz
35
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex041.td.xz holds the published optimal tree decomposition in .td format; one file is one case
36
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 41: "ex041.gr task2.5_output_BFS-7-8-9_clean/task2_output/MAR_prob/Promedus_27_9.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
37
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 9. Since the published decomposition is optimal, a pass means width equals 9; a lower width would signal a checker or data error
38
  Companions (not included): td/exact/ex041.td.xz; README.md; LICENSE
39
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
41
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex061.gr.xz
42
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex061.td.xz holds the published optimal tree decomposition in .td format; one file is one case
43
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 61: "ex061.gr task2.5_output_BFS-1-2-3-4-5-6_clean/task2_output/PR_prob/logistics.a.cnf_2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
44
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 22. Since the published decomposition is optimal, a pass means width equals 22; a lower width would signal a checker or data error
45
  Companions (not included): td/exact/ex061.td.xz; README.md; LICENSE
46
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
48
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex081.gr.xz
49
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex081.td.xz holds the published optimal tree decomposition in .td format; one file is one case
50
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 81: "ex081.gr task2.5_output_BFS-1-2-3-4-5-6_clean/SAT-instances-gaifman/app16-gaifman/div2.c.20.smt2-stp212.gaifman_2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
51
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 6. Since the published decomposition is optimal, a pass means width equals 6; a lower width would signal a checker or data error
52
  Companions (not included): td/exact/ex081.td.xz; README.md; LICENSE
53
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
55
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex101.gr.xz
56
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex101.td.xz holds the published optimal tree decomposition in .td format; one file is one case
57
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 101: "ex101.gr task2.5_output_BFS-1-2-3-4-5-6_clean/SAT-instances-gaifman/sc14-app-gaifman/grid-strips-grid-y-3.035-NOTKNOWN.gaifman_1.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
58
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 540. Since the published decomposition is optimal, a pass means width equals 540; a lower width would signal a checker or data error
59
  Companions (not included): td/exact/ex101.td.xz; README.md; LICENSE
60
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
62
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex121.gr.xz
63
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex121.td.xz holds the published optimal tree decomposition in .td format; one file is one case
64
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 121: "ex121.gr selected_C2_C3/jgiraldezlevy.2200.9086.08.40.83.gaifman_2_C2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
65
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 34. Since the published decomposition is optimal, a pass means width equals 34; a lower width would signal a checker or data error
66
  Companions (not included): td/exact/ex121.td.xz; README.md; LICENSE
67
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
69
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex141.gr.xz
70
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex141.td.xz holds the published optimal tree decomposition in .td format; one file is one case
71
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 141: "ex141.gr selected_C2_C3/modgen-n200-m90860q08c40-14808.gaifman_2_C2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
72
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 34. Since the published decomposition is optimal, a pass means width equals 34; a lower width would signal a checker or data error
73
  Companions (not included): td/exact/ex141.td.xz; README.md; LICENSE
74
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
76
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex161.gr.xz
77
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex161.td.xz holds the published optimal tree decomposition in .td format; one file is one case
78
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 161: "ex161.gr selected_C2_C3/aaai10-planning-ipc5-TPP-21-step11.gaifman_1_C29_C3.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
79
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 12. Since the published decomposition is optimal, a pass means width equals 12; a lower width would signal a checker or data error
80
  Companions (not included): td/exact/ex161.td.xz; README.md; LICENSE
81
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
 
83
  Decompressed from: https://raw.githubusercontent.com/PACE-challenge/Treewidth-PACE-2017-instances/44e40b0ce87d2360f6114ba2aa503ef8df5e9256/gr/exact/ex181.gr.xz
84
  Format: xz-compressed .gr text (PACE graph format: comment lines start with c, a p-line "p tw <n> <m>", then one edge "u v" per line, vertices 1..n); unpacks with Python lzma; companion td/exact/ex181.td.xz holds the published optimal tree decomposition in .td format; one file is one case
85
  Origin: PACE 2017 official exact-track instance; gr/exact/sources.txt line 181: "ex181.gr task2.5_output_BFS-1-2-3-4-5-6_clean/SAT-instances-gaifman/sr15bench-gaifman/mrpp_8x8#22_10.gaifman_2.gr.xz" names the graph it was derived from (Gaifman graphs of SAT instances, UAI probabilistic inference networks, road graphs and similar collections, preprocessed by the organisers)
 
 
86
  Check: width = (largest bag size) - 1; it passes when feasible and width <= 18. Since the published decomposition is optimal, a pass means width equals 18; a lower width would signal a checker or data error
87
  Companions (not included): td/exact/ex181.td.xz; README.md; LICENSE
88
  Dependencies: not packed; Python lzma module of the standard library to unpack .xz
campaign_files/2026-10-g-solver-instances-2/pace-2024-ocm/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Optimal ordering (minimum crossing number). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Minimisation: the agent returns a linear order of B. fewer crossings than published is also correct, so a published order that is not optimal never marks a better answer wrong.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -20,8 +18,6 @@ Files:
20
 
21
  - `g043-11/exact-public.zip`: PACE 2024 exact track public instances: one-sided crossing minimisation on 100 bipartite graphs, with optimal orderings.
22
  Format: zip of 100 .gr files (DIMACS-like: p ocr n0 n1 m, then edges); companion exact-public-sol.zip holds only 99 .sol files (order of the free layer, one vertex per line): 17.sol is missing, so instance 17.gr has no published answer and is dropped from the answers; a case is one .gr with its .sol, 99 cases
23
- Reference result: optimal ordering (minimum crossing number)
24
- Reference source: exact-public-sol.zip (99 .sol files; 17.sol missing, instance 17 dropped from the answers). The challenge page pacechallenge.org/2024 lists the download as "public instances + solutions" and states "The task is to compute an optimal solution for each given graph" and "Submissions that output an incorrect solution or a solution that is known to be non-optimal will be disqualified"; it does not say in so many words that the published solutions are optimal.
25
  Check: Minimisation: the agent returns a linear order of B. fewer crossings than published is also correct, so a published order that is not optimal never marks a better answer wrong
26
  Companions (not included): exact-public-sol.zip
27
  Dependencies: not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Minimisation: the agent returns a linear order of B. fewer crossings than published is also correct, so a published order that is not optimal never marks a better answer wrong.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
18
 
19
  - `g043-11/exact-public.zip`: PACE 2024 exact track public instances: one-sided crossing minimisation on 100 bipartite graphs, with optimal orderings.
20
  Format: zip of 100 .gr files (DIMACS-like: p ocr n0 n1 m, then edges); companion exact-public-sol.zip holds only 99 .sol files (order of the free layer, one vertex per line): 17.sol is missing, so instance 17.gr has no published answer and is dropped from the answers; a case is one .gr with its .sol, 99 cases
 
 
21
  Check: Minimisation: the agent returns a linear order of B. fewer crossings than published is also correct, so a published order that is not optimal never marks a better answer wrong
22
  Companions (not included): exact-public-sol.zip
23
  Dependencies: not packed
campaign_files/2026-10-gt-browser-tests-2/openscad-cmdline-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 4 of 10 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: golden_output_files. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** openscad -o out.stl --enable=predictible-output --render tests/data/scad/stl/stl-export.scad (and --export-format binstl for the binary file); openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-export.scad; openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-import-export_cube.scad (run from the unit layout so ../../obj/cube.obj resolves); openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-import-export_dodecahedron.scad; openscad -o out.obj --enable=predictible-output tests/data/scad/3D/features/polyhedron-cube.scad; openscad -o out.3mf --enable=predictible-output tests/data/scad/3mf/3mf-export.scad; openscad -o out.pov --enable=predictible-output --backend=manifold [--camera=tx,ty,tz,rx,ry,rz,dist] tests/data/scad/pov/pov-export.scad; openscad -o out.json --export-format param tests/data/scad/issues/issue6597.scad; openscad -o out.json --export-format param tests/data/scad/issues/issue7013.scad; openscad -o out.ast -p tests/data/scad/customizer/setofparameter.json -P firstSet tests/data/scad/customizer/setofparameter.scad (one run per set)
14
 
15
  **Licence (SPDX):** GPL-2.0-or-later
@@ -21,24 +19,18 @@ Files:
21
  - `tests/data/scad/stl/stl-export.scad`: OpenSCAD: render a small model and export it as ASCII STL and as binary STL with predictable output.
22
  Format: OpenSCAD .scad input; STL (ASCII and binary) goldens
23
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
24
- Reference result: golden_output_files
25
- Reference source: tests/regression/export-stl/stl-export-expected.stl and tests/regression/export-binstl/stl-export-expected.stl; the test is registered in tests/CMakeLists.txt lines 894-895, 972 and 978 with arguments `--enable=predictible-output --render (ASCII) and the same plus --export-format binstl (binary)`
26
  Check: Final file state: compare each written .stl with its golden by compare_default of tests/test_cmdline_tool.py (suffix stl has no compare_stl, so line 271-272 dispatch to compare_default): normalise, then exact line equality. No screenshot
27
  Tools: openscad -o out.stl --enable=predictible-output --render tests/data/scad/stl/stl-export.scad (and --export-format binstl for the binary file)
28
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
29
  - `tests/data/scad/obj/obj-export.scad`: OpenSCAD: export a centred unit cube as Wavefront OBJ.
30
  Format: OpenSCAD .scad input; Wavefront OBJ golden
31
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
32
- Reference result: golden_output_files
33
- Reference source: tests/regression/export-obj/obj-export-expected.obj; the test is registered in tests/CMakeLists.txt lines 904 and 981 with arguments `--enable=predictible-output`
34
  Check: Final file state: compare the written .obj with the golden by compare_default (no compare_obj, lines 271-272): normalise, then exact line equality. No screenshot
35
  Tools: openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-export.scad
36
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
37
  - `tests/data/scad/obj/obj-import-export_cube.scad`: OpenSCAD: import an OBJ mesh of a cube and export it again as OBJ.
38
  Format: OpenSCAD .scad input with an imported OBJ mesh; Wavefront OBJ golden
39
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
40
- Reference result: golden_output_files
41
- Reference source: tests/regression/export-obj/obj-import-export_cube-expected.obj; the test is registered in tests/CMakeLists.txt lines 906 and 981 with arguments `--enable=predictible-output`
42
  Check: Final file state: compare the written .obj with the golden by compare_default: normalise, then exact line equality. No screenshot
43
  Tools: openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-import-export_cube.scad (run from the unit layout so ../../obj/cube.obj resolves)
44
  Companions (not included): tests/data/obj/cube.obj (the mesh the .scad imports by relative path ../../obj/cube.obj)
@@ -46,8 +38,6 @@ Files:
46
  - `tests/data/scad/obj/obj-import-export_dodecahedron.scad`: OpenSCAD: import an OBJ mesh of a dodecahedron and export it again as OBJ.
47
  Format: OpenSCAD .scad input with an imported OBJ mesh; Wavefront OBJ golden
48
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
49
- Reference result: golden_output_files
50
- Reference source: tests/regression/export-obj/obj-import-export_dodecahedron-expected.obj; the test is registered in tests/CMakeLists.txt lines 905 and 981 with arguments `--enable=predictible-output`
51
  Check: Final file state: compare the written .obj with the golden by compare_default: normalise, then exact line equality. No screenshot
52
  Tools: openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-import-export_dodecahedron.scad
53
  Companions (not included): tests/data/obj/dodecahedron.obj (the mesh the .scad imports)
@@ -55,40 +45,30 @@ Files:
55
  - `tests/data/scad/3D/features/polyhedron-cube.scad`: OpenSCAD: build a cube from a polyhedron() face list and export it as OBJ.
56
  Format: OpenSCAD .scad input; Wavefront OBJ golden
57
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
58
- Reference result: golden_output_files
59
- Reference source: tests/regression/export-obj/polyhedron-cube-expected.obj; the test is registered in tests/CMakeLists.txt lines 907 and 981 with arguments `--enable=predictible-output`
60
  Check: Final file state: compare the written .obj with the golden by compare_default: normalise, then exact line equality. No screenshot
61
  Tools: openscad -o out.obj --enable=predictible-output tests/data/scad/3D/features/polyhedron-cube.scad
62
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
63
  - `tests/data/scad/3mf/3mf-export.scad`: OpenSCAD: export a model as a 3MF package.
64
  Format: OpenSCAD .scad input; 3MF (zip with 3D/3dmodel.model XML) golden
65
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
66
- Reference result: golden_output_files
67
- Reference source: tests/regression/export-3mf/3mf-export-expected.3mf; the test is registered in tests/CMakeLists.txt lines 909 and 983 with arguments `--enable=predictible-output`
68
  Check: Final file state: extract 3D/3dmodel.model from the written .3mf and normalise it by post_process_3mf (tests/test_cmdline_tool.py lines 280-298, called at line 488 when the suffix is 3mf), then compare with the golden by compare_default (exact line equality). No screenshot
69
  Tools: openscad -o out.3mf --enable=predictible-output tests/data/scad/3mf/3mf-export.scad
70
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
71
  - `tests/data/scad/pov/pov-export.scad`: OpenSCAD: export a model as a POV-Ray scene under ten camera settings.
72
  Format: OpenSCAD .scad input; POV-Ray .pov scene goldens
73
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
74
- Reference result: golden_output_files
75
- Reference source: tests/regression/export-pov-{as-is,translate-1..4,rotate-1..3,distance-1,all}/pov-export-expected.pov (10 goldens); the test is registered in tests/CMakeLists.txt lines 912 and 987-996 with arguments `--enable=predictible-output --backend=manifold, plus --camera=<translate x,y,z,rotate x,y,z,distance> (nine settings)`
76
  Check: Final file state: compare each written .pov with its golden by compare_default: normalise, then exact line equality. No screenshot (the scene file is judged, not a rendered image)
77
  Tools: openscad -o out.pov --enable=predictible-output --backend=manifold [--camera=tx,ty,tz,rx,ry,rz,dist] tests/data/scad/pov/pov-export.scad
78
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
79
  - `tests/data/scad/issues/issue6597.scad`: OpenSCAD: export the Customizer parameter definitions of a model as JSON.
80
  Format: OpenSCAD .scad input with Customizer annotations; parameter JSON golden
81
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
82
- Reference result: golden_output_files
83
- Reference source: tests/regression/export-param/issue6597-expected.json; the test is registered in tests/CMakeLists.txt lines 915 and 999 with arguments `--export-format param`
84
  Check: Final file state: parse the written JSON and the golden and require equality of the parsed values by compare_json (tests/test_cmdline_tool.py lines 195-208, picked for suffix json at line 271). No screenshot
85
  Tools: openscad -o out.json --export-format param tests/data/scad/issues/issue6597.scad
86
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
87
  - `tests/data/scad/issues/issue7013.scad`: OpenSCAD: export the Customizer parameter definitions of a model with a saved parameter set as JSON.
88
  Format: OpenSCAD .scad input with Customizer annotations and a parameter-set JSON; parameter JSON golden
89
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
90
- Reference result: golden_output_files
91
- Reference source: tests/regression/export-param/issue7013-expected.json; the test is registered in tests/CMakeLists.txt lines 916 and 999 with arguments `--export-format param`
92
  Check: Final file state: parse the written JSON and the golden and require equality by compare_json (tests/test_cmdline_tool.py lines 195-208). No screenshot
93
  Tools: openscad -o out.json --export-format param tests/data/scad/issues/issue7013.scad
94
  Companions (not included): tests/data/scad/issues/issue7013.json (the saved parameter-set file that sits beside the model)
@@ -96,8 +76,6 @@ Files:
96
  - `tests/data/scad/customizer/setofparameter.scad`: OpenSCAD: apply named Customizer parameter sets to a model and write the resulting program (AST dump).
97
  Format: OpenSCAD .scad input with a Customizer parameter-set JSON; .ast (program dump) goldens
98
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
99
- Reference result: golden_output_files
100
- Reference source: tests/regression/customizer-{first,wrong,incomplete,imgset,setNameWithDot}/setofparameter-expected.ast (5 goldens); the test is registered in tests/CMakeLists.txt lines 1102-1103 and 1112-1116 with arguments `-p setofparameter.json -P <firstSet | wrongSetValues | thirdSet | imagine | Name.dot>`
101
  Check: Final file state: compare each written .ast with its golden by compare_default: normalise, then exact line equality; the parameter values that land in the program are the answer. No screenshot
102
  Tools: openscad -o out.ast -p tests/data/scad/customizer/setofparameter.json -P firstSet tests/data/scad/customizer/setofparameter.scad (one run per set)
103
  Companions (not included): tests/data/scad/customizer/setofparameter.json (the Customizer parameter sets firstSet, wrongSetValues, thirdSet, imagine, Name.dot)
 
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 4 of 10 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** openscad -o out.stl --enable=predictible-output --render tests/data/scad/stl/stl-export.scad (and --export-format binstl for the binary file); openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-export.scad; openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-import-export_cube.scad (run from the unit layout so ../../obj/cube.obj resolves); openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-import-export_dodecahedron.scad; openscad -o out.obj --enable=predictible-output tests/data/scad/3D/features/polyhedron-cube.scad; openscad -o out.3mf --enable=predictible-output tests/data/scad/3mf/3mf-export.scad; openscad -o out.pov --enable=predictible-output --backend=manifold [--camera=tx,ty,tz,rx,ry,rz,dist] tests/data/scad/pov/pov-export.scad; openscad -o out.json --export-format param tests/data/scad/issues/issue6597.scad; openscad -o out.json --export-format param tests/data/scad/issues/issue7013.scad; openscad -o out.ast -p tests/data/scad/customizer/setofparameter.json -P firstSet tests/data/scad/customizer/setofparameter.scad (one run per set)
12
 
13
  **Licence (SPDX):** GPL-2.0-or-later
 
19
  - `tests/data/scad/stl/stl-export.scad`: OpenSCAD: render a small model and export it as ASCII STL and as binary STL with predictable output.
20
  Format: OpenSCAD .scad input; STL (ASCII and binary) goldens
21
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
22
  Check: Final file state: compare each written .stl with its golden by compare_default of tests/test_cmdline_tool.py (suffix stl has no compare_stl, so line 271-272 dispatch to compare_default): normalise, then exact line equality. No screenshot
23
  Tools: openscad -o out.stl --enable=predictible-output --render tests/data/scad/stl/stl-export.scad (and --export-format binstl for the binary file)
24
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
25
  - `tests/data/scad/obj/obj-export.scad`: OpenSCAD: export a centred unit cube as Wavefront OBJ.
26
  Format: OpenSCAD .scad input; Wavefront OBJ golden
27
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
28
  Check: Final file state: compare the written .obj with the golden by compare_default (no compare_obj, lines 271-272): normalise, then exact line equality. No screenshot
29
  Tools: openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-export.scad
30
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
31
  - `tests/data/scad/obj/obj-import-export_cube.scad`: OpenSCAD: import an OBJ mesh of a cube and export it again as OBJ.
32
  Format: OpenSCAD .scad input with an imported OBJ mesh; Wavefront OBJ golden
33
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
34
  Check: Final file state: compare the written .obj with the golden by compare_default: normalise, then exact line equality. No screenshot
35
  Tools: openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-import-export_cube.scad (run from the unit layout so ../../obj/cube.obj resolves)
36
  Companions (not included): tests/data/obj/cube.obj (the mesh the .scad imports by relative path ../../obj/cube.obj)
 
38
  - `tests/data/scad/obj/obj-import-export_dodecahedron.scad`: OpenSCAD: import an OBJ mesh of a dodecahedron and export it again as OBJ.
39
  Format: OpenSCAD .scad input with an imported OBJ mesh; Wavefront OBJ golden
40
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
41
  Check: Final file state: compare the written .obj with the golden by compare_default: normalise, then exact line equality. No screenshot
42
  Tools: openscad -o out.obj --enable=predictible-output tests/data/scad/obj/obj-import-export_dodecahedron.scad
43
  Companions (not included): tests/data/obj/dodecahedron.obj (the mesh the .scad imports)
 
45
  - `tests/data/scad/3D/features/polyhedron-cube.scad`: OpenSCAD: build a cube from a polyhedron() face list and export it as OBJ.
46
  Format: OpenSCAD .scad input; Wavefront OBJ golden
47
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
48
  Check: Final file state: compare the written .obj with the golden by compare_default: normalise, then exact line equality. No screenshot
49
  Tools: openscad -o out.obj --enable=predictible-output tests/data/scad/3D/features/polyhedron-cube.scad
50
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
51
  - `tests/data/scad/3mf/3mf-export.scad`: OpenSCAD: export a model as a 3MF package.
52
  Format: OpenSCAD .scad input; 3MF (zip with 3D/3dmodel.model XML) golden
53
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
54
  Check: Final file state: extract 3D/3dmodel.model from the written .3mf and normalise it by post_process_3mf (tests/test_cmdline_tool.py lines 280-298, called at line 488 when the suffix is 3mf), then compare with the golden by compare_default (exact line equality). No screenshot
55
  Tools: openscad -o out.3mf --enable=predictible-output tests/data/scad/3mf/3mf-export.scad
56
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
57
  - `tests/data/scad/pov/pov-export.scad`: OpenSCAD: export a model as a POV-Ray scene under ten camera settings.
58
  Format: OpenSCAD .scad input; POV-Ray .pov scene goldens
59
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
60
  Check: Final file state: compare each written .pov with its golden by compare_default: normalise, then exact line equality. No screenshot (the scene file is judged, not a rendered image)
61
  Tools: openscad -o out.pov --enable=predictible-output --backend=manifold [--camera=tx,ty,tz,rx,ry,rz,dist] tests/data/scad/pov/pov-export.scad
62
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
63
  - `tests/data/scad/issues/issue6597.scad`: OpenSCAD: export the Customizer parameter definitions of a model as JSON.
64
  Format: OpenSCAD .scad input with Customizer annotations; parameter JSON golden
65
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
66
  Check: Final file state: parse the written JSON and the golden and require equality of the parsed values by compare_json (tests/test_cmdline_tool.py lines 195-208, picked for suffix json at line 271). No screenshot
67
  Tools: openscad -o out.json --export-format param tests/data/scad/issues/issue6597.scad
68
  Dependencies: OpenSCAD built from commit c0ac8289f1c9db6f70891da5df3b4085b7eaf8b6 of openscad/openscad (GPL-2.0-or-later, open source, not packed), configured with experimental features on (the export tests are registered EXPERIMENTAL in tests/CMakeLists.txt); the openscad command-line binary does the work headless; judge: Python 3 with the comparison functions of tests/test_cmdline_tool.py at the same commit (standard library only: re, json, zipfile, difflib)
69
  - `tests/data/scad/issues/issue7013.scad`: OpenSCAD: export the Customizer parameter definitions of a model with a saved parameter set as JSON.
70
  Format: OpenSCAD .scad input with Customizer annotations and a parameter-set JSON; parameter JSON golden
71
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
72
  Check: Final file state: parse the written JSON and the golden and require equality by compare_json (tests/test_cmdline_tool.py lines 195-208). No screenshot
73
  Tools: openscad -o out.json --export-format param tests/data/scad/issues/issue7013.scad
74
  Companions (not included): tests/data/scad/issues/issue7013.json (the saved parameter-set file that sits beside the model)
 
76
  - `tests/data/scad/customizer/setofparameter.scad`: OpenSCAD: apply named Customizer parameter sets to a model and write the resulting program (AST dump).
77
  Format: OpenSCAD .scad input with a Customizer parameter-set JSON; .ast (program dump) goldens
78
  Origin: OpenSCAD command-line regression-test input and golden written by the OpenSCAD developers (tool_fixture)
 
 
79
  Check: Final file state: compare each written .ast with its golden by compare_default: normalise, then exact line equality; the parameter values that land in the program are the answer. No screenshot
80
  Tools: openscad -o out.ast -p tests/data/scad/customizer/setofparameter.json -P firstSet tests/data/scad/customizer/setofparameter.scad (one run per set)
81
  Companions (not included): tests/data/scad/customizer/setofparameter.json (the Customizer parameter sets firstSet, wrongSetValues, thirdSet, imagine, Name.dot)
campaign_files/2026-10-gt-browser-tests-2/pacparser-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_proxy_strings. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** write the PAC file with a text editor and point the browser at it in prefs.js (Firefox: network.proxy.type 2 and network.proxy.autoconfig_url) or in Chrome's ProxySettings policy JSON; check it with pacparser's command-line tool pactester (pactester -p proxy.pac -u URL [-c client IP]), the same tool the upstream tests run
14
 
15
  **Licence (SPDX):** LGPL-3.0-or-later
@@ -21,8 +19,6 @@ Files:
21
  - `tests/proxy.pac`: Browser proxy auto-config: a PAC file with host, URL and client-IP rules and 10 expected proxy results.
22
  Format: JavaScript proxy auto-config file (FindProxyForURL) plus a pipe-separated test table of pactester arguments and expected results
23
  Origin: test PAC script and expected results of pacparser's own test suite (tool_fixture)
24
- Reference result: expected_proxy_strings
25
- Reference source: the rules are in tests/proxy.pac lines 5-38
26
  Check: Final file state, no screenshot needed: run pactester on the agent's PAC file for each testdata line with the given -u and -c arguments and compare the returned string with the expected result by exact string equality; for the GUI half, also read the browser config (prefs.js network.proxy.type and network.proxy.autoconfig_url, or the policy JSON) and check it points at the PAC file. Line 8 is marked INTERNET_REQUIRED in testdata and needs DNS for www.google.com; line 9 needs the name to fail to resolve; run those only where DNS is available, as the upstream runner does
27
  Tools: write the PAC file with a text editor and point the browser at it in prefs.js (Firefox: network.proxy.type 2 and network.proxy.autoconfig_url) or in Chrome's ProxySettings policy JSON; check it with pacparser's command-line tool pactester (pactester -p proxy.pac -u URL [-c client IP]), the same tool the upstream tests run
28
  Companions (not included): tests/testdata; COPYING
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** write the PAC file with a text editor and point the browser at it in prefs.js (Firefox: network.proxy.type 2 and network.proxy.autoconfig_url) or in Chrome's ProxySettings policy JSON; check it with pacparser's command-line tool pactester (pactester -p proxy.pac -u URL [-c client IP]), the same tool the upstream tests run
12
 
13
  **Licence (SPDX):** LGPL-3.0-or-later
 
19
  - `tests/proxy.pac`: Browser proxy auto-config: a PAC file with host, URL and client-IP rules and 10 expected proxy results.
20
  Format: JavaScript proxy auto-config file (FindProxyForURL) plus a pipe-separated test table of pactester arguments and expected results
21
  Origin: test PAC script and expected results of pacparser's own test suite (tool_fixture)
 
 
22
  Check: Final file state, no screenshot needed: run pactester on the agent's PAC file for each testdata line with the given -u and -c arguments and compare the returned string with the expected result by exact string equality; for the GUI half, also read the browser config (prefs.js network.proxy.type and network.proxy.autoconfig_url, or the policy JSON) and check it points at the PAC file. Line 8 is marked INTERNET_REQUIRED in testdata and needs DNS for www.google.com; line 9 needs the name to fail to resolve; run those only where DNS is available, as the upstream runner does
23
  Tools: write the PAC file with a text editor and point the browser at it in prefs.js (Firefox: network.proxy.type 2 and network.proxy.autoconfig_url) or in Chrome's ProxySettings policy JSON; check it with pacparser's command-line tool pactester (pactester -p proxy.pac -u URL [-c client IP]), the same tool the upstream tests run
24
  Companions (not included): tests/testdata; COPYING
campaign_files/2026-10-hw-eda-tests-2/openroad-src-dpl-test/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: golden_def_and_log_exact_match. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/dpl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=simple01 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/dpl/test/. The unit passes when openroad exits 0, the log equals simple01.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against simple01.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of simple01.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
14
 
15
  **Licence (SPDX):** BSD-3-Clause AND NOASSERTION
@@ -21,8 +19,6 @@ Files:
21
  - `src/dpl/test/simple01.tcl`: OpenROAD detailed placer: legalize one cell placed off the row grid and match the golden DEF and the placement report exactly.
22
  Format: OpenROAD Tcl regression script (.tcl) with golden log (.ok) and golden DEF (.defok); inputs are LEF 5.x library files and a DEF design
23
  Origin: regression test written by the OpenROAD developers for their own placement tools (tool_fixture); the golden .ok and .defok are outputs of an earlier OpenROAD run committed as references.
24
- Reference result: golden_def_and_log_exact_match
25
- Reference source: src/dpl/test/simple01.defok (golden legalized DEF) and src/dpl/test/simple01.ok (golden log with `total displacement 4.7 u`, `legalized HPWL 9.1 u`, `No differences found.`)
26
  Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/dpl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=simple01 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/dpl/test/. The unit passes when openroad exits 0, the log equals simple01.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against simple01.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of simple01.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
27
  Companions (not included): src/dpl/test/simple01.ok; src/dpl/test/simple01.defok; src/dpl/test/simple01.def; test/Nangate45/Nangate45.lef; test/helpers.tcl; test/regression_test.sh; test/regression.bzl; LICENSE
28
  Dependencies: OpenROAD built from the same commit b1ad304cd6fc6252d5912e873ef3dd8569f1aea4 (2026-10-07, BSD-3-Clause), default CPU build: the test runs inside the openroad binary's Tcl shell (`openroad -no_splash -no_init -exit -threads 1`, test/regression_test.sh line 33); not packed; bash and diff from the build environment for test/regression_test.sh; not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/dpl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=simple01 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/dpl/test/. The unit passes when openroad exits 0, the log equals simple01.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against simple01.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of simple01.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
12
 
13
  **Licence (SPDX):** BSD-3-Clause AND NOASSERTION
 
19
  - `src/dpl/test/simple01.tcl`: OpenROAD detailed placer: legalize one cell placed off the row grid and match the golden DEF and the placement report exactly.
20
  Format: OpenROAD Tcl regression script (.tcl) with golden log (.ok) and golden DEF (.defok); inputs are LEF 5.x library files and a DEF design
21
  Origin: regression test written by the OpenROAD developers for their own placement tools (tool_fixture); the golden .ok and .defok are outputs of an earlier OpenROAD run committed as references.
 
 
22
  Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/dpl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=simple01 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/dpl/test/. The unit passes when openroad exits 0, the log equals simple01.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against simple01.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of simple01.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
23
  Companions (not included): src/dpl/test/simple01.ok; src/dpl/test/simple01.defok; src/dpl/test/simple01.def; test/Nangate45/Nangate45.lef; test/helpers.tcl; test/regression_test.sh; test/regression.bzl; LICENSE
24
  Dependencies: OpenROAD built from the same commit b1ad304cd6fc6252d5912e873ef3dd8569f1aea4 (2026-10-07, BSD-3-Clause), default CPU build: the test runs inside the openroad binary's Tcl shell (`openroad -no_splash -no_init -exit -threads 1`, test/regression_test.sh line 33); not packed; bash and diff from the build environment for test/regression_test.sh; not packed
campaign_files/2026-10-hw-eda-tests-2/openroad-src-gpl-test/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: golden_def_and_log_exact_match. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/gpl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=simple01-obs TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/gpl/test/. The unit passes when openroad exits 0, the log equals simple01-obs.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against simple01-obs.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of simple01-obs.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
14
 
15
  **Licence (SPDX):** BSD-3-Clause AND NOASSERTION
@@ -21,8 +19,6 @@ Files:
21
  - `src/gpl/test/simple01-obs.tcl`: OpenROAD global placer: place a small design around a placement obstruction at target density 0.8, matching the 237-iteration log and the golden DEF.
22
  Format: OpenROAD Tcl regression script (.tcl) with golden log (.ok) and golden DEF (.defok); inputs are LEF 5.x library files and a DEF design
23
  Origin: regression test written by the OpenROAD developers for their own placement tools (tool_fixture); the golden .ok and .defok are outputs of an earlier OpenROAD run committed as references.
24
- Reference result: golden_def_and_log_exact_match
25
- Reference source: src/gpl/test/simple01-obs.defok (golden placed DEF) and src/gpl/test/simple01-obs.ok (golden log: per-iteration overflow and HPWL, `Global placement finished at iteration 237`, `Final HPWL (um): 2714.72`, `No differences found.`)
26
  Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/gpl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=simple01-obs TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/gpl/test/. The unit passes when openroad exits 0, the log equals simple01-obs.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against simple01-obs.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of simple01-obs.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
27
  Companions (not included): src/gpl/test/simple01-obs.ok; src/gpl/test/simple01-obs.defok; src/gpl/test/simple01-obs.def; src/gpl/test/nangate45.lef; test/helpers.tcl; test/regression_test.sh; test/regression.bzl; LICENSE
28
  Dependencies: OpenROAD built from the same commit b1ad304cd6fc6252d5912e873ef3dd8569f1aea4 (2026-10-07, BSD-3-Clause), default CPU build: the test runs inside the openroad binary's Tcl shell (`openroad -no_splash -no_init -exit -threads 1`, test/regression_test.sh line 33); not packed; bash and diff from the build environment for test/regression_test.sh; not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/gpl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=simple01-obs TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/gpl/test/. The unit passes when openroad exits 0, the log equals simple01-obs.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against simple01-obs.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of simple01-obs.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
12
 
13
  **Licence (SPDX):** BSD-3-Clause AND NOASSERTION
 
19
  - `src/gpl/test/simple01-obs.tcl`: OpenROAD global placer: place a small design around a placement obstruction at target density 0.8, matching the 237-iteration log and the golden DEF.
20
  Format: OpenROAD Tcl regression script (.tcl) with golden log (.ok) and golden DEF (.defok); inputs are LEF 5.x library files and a DEF design
21
  Origin: regression test written by the OpenROAD developers for their own placement tools (tool_fixture); the golden .ok and .defok are outputs of an earlier OpenROAD run committed as references.
 
 
22
  Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/gpl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=simple01-obs TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/gpl/test/. The unit passes when openroad exits 0, the log equals simple01-obs.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against simple01-obs.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of simple01-obs.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
23
  Companions (not included): src/gpl/test/simple01-obs.ok; src/gpl/test/simple01-obs.defok; src/gpl/test/simple01-obs.def; src/gpl/test/nangate45.lef; test/helpers.tcl; test/regression_test.sh; test/regression.bzl; LICENSE
24
  Dependencies: OpenROAD built from the same commit b1ad304cd6fc6252d5912e873ef3dd8569f1aea4 (2026-10-07, BSD-3-Clause), default CPU build: the test runs inside the openroad binary's Tcl shell (`openroad -no_splash -no_init -exit -threads 1`, test/regression_test.sh line 33); not packed; bash and diff from the build environment for test/regression_test.sh; not packed
campaign_files/2026-10-hw-eda-tests-2/openroad-src-ppl-test/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: golden_def_and_log_exact_match. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/ppl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=add_constraint1 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/ppl/test/. The unit passes when openroad exits 0, the log equals add_constraint1.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against add_constraint1.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of add_constraint1.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
14
 
15
  **Licence (SPDX):** BSD-3-Clause AND NOASSERTION
@@ -21,8 +19,6 @@ Files:
21
  - `src/ppl/test/add_constraint1.tcl`: OpenROAD pin placer: place the 54 I/O pins of gcd with all inputs restricted to the top edge, matching the golden DEF exactly.
22
  Format: OpenROAD Tcl regression script (.tcl) with golden log (.ok) and golden DEF (.defok); inputs are LEF 5.x library files and a DEF design
23
  Origin: regression test written by the OpenROAD developers for their own placement tools (tool_fixture); the golden .ok and .defok are outputs of an earlier OpenROAD run committed as references.
24
- Reference result: golden_def_and_log_exact_match
25
- Reference source: src/ppl/test/add_constraint1.defok (golden DEF with every pin location) and src/ppl/test/add_constraint1.ok (golden log: 54 I/O, `Restrict pins ... to region 0.00u-100.13u at the TOP edge`, `I/O nets HPWL: 1859.88 um.`, `No differences found.`)
26
  Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/ppl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=add_constraint1 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/ppl/test/. The unit passes when openroad exits 0, the log equals add_constraint1.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against add_constraint1.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of add_constraint1.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
27
  Companions (not included): src/ppl/test/add_constraint1.ok; src/ppl/test/add_constraint1.defok; src/ppl/test/gcd.def; test/Nangate45/Nangate45.lef; test/helpers.tcl; test/regression_test.sh; test/regression.bzl; LICENSE
28
  Dependencies: OpenROAD built from the same commit b1ad304cd6fc6252d5912e873ef3dd8569f1aea4 (2026-10-07, BSD-3-Clause), default CPU build: the test runs inside the openroad binary's Tcl shell (`openroad -no_splash -no_init -exit -threads 1`, test/regression_test.sh line 33); not packed; bash and diff from the build environment for test/regression_test.sh; not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/ppl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=add_constraint1 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/ppl/test/. The unit passes when openroad exits 0, the log equals add_constraint1.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against add_constraint1.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of add_constraint1.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
12
 
13
  **Licence (SPDX):** BSD-3-Clause AND NOASSERTION
 
19
  - `src/ppl/test/add_constraint1.tcl`: OpenROAD pin placer: place the 54 I/O pins of gcd with all inputs restricted to the top edge, matching the golden DEF exactly.
20
  Format: OpenROAD Tcl regression script (.tcl) with golden log (.ok) and golden DEF (.defok); inputs are LEF 5.x library files and a DEF design
21
  Origin: regression test written by the OpenROAD developers for their own placement tools (tool_fixture); the golden .ok and .defok are outputs of an earlier OpenROAD run committed as references.
 
 
22
  Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/ppl/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=add_constraint1 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/ppl/test/. The unit passes when openroad exits 0, the log equals add_constraint1.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against add_constraint1.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of add_constraint1.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
23
  Companions (not included): src/ppl/test/add_constraint1.ok; src/ppl/test/add_constraint1.defok; src/ppl/test/gcd.def; test/Nangate45/Nangate45.lef; test/helpers.tcl; test/regression_test.sh; test/regression.bzl; LICENSE
24
  Dependencies: OpenROAD built from the same commit b1ad304cd6fc6252d5912e873ef3dd8569f1aea4 (2026-10-07, BSD-3-Clause), default CPU build: the test runs inside the openroad binary's Tcl shell (`openroad -no_splash -no_init -exit -threads 1`, test/regression_test.sh line 33); not packed; bash and diff from the build environment for test/regression_test.sh; not packed
campaign_files/2026-10-hw-eda-tests-2/openroad-src-tap-test/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: golden_def_and_log_exact_match. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/tap/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=gcd_nangate45 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/tap/test/. The unit passes when openroad exits 0, the log equals gcd_nangate45.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against gcd_nangate45.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of gcd_nangate45.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
14
 
15
  **Licence (SPDX):** BSD-3-Clause AND NOASSERTION
@@ -21,8 +19,6 @@ Files:
21
  - `src/tap/test/gcd_nangate45.tcl`: OpenROAD tapcell: insert endcaps and well taps every 20 um into gcd, expecting 114 endcaps and 118 tapcells and the golden DEF.
22
  Format: OpenROAD Tcl regression script (.tcl) with golden log (.ok) and golden DEF (.defok); inputs are LEF 5.x library files and a DEF design
23
  Origin: regression test written by the OpenROAD developers for their own placement tools (tool_fixture); the golden .ok and .defok are outputs of an earlier OpenROAD run committed as references.
24
- Reference result: golden_def_and_log_exact_match
25
- Reference source: src/tap/test/gcd_nangate45.defok (golden DEF with every inserted cell) and src/tap/test/gcd_nangate45.ok (golden log: `Inserted 114 endcaps.`, `Inserted 118 tapcells.`, `No differences found.`)
26
  Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/tap/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=gcd_nangate45 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/tap/test/. The unit passes when openroad exits 0, the log equals gcd_nangate45.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against gcd_nangate45.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of gcd_nangate45.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
27
  Companions (not included): src/tap/test/gcd_nangate45.ok; src/tap/test/gcd_nangate45.defok; src/tap/test/gcd_nangate45.def; test/Nangate45/Nangate45_tech.lef; test/Nangate45/Nangate45_stdcell.lef; test/helpers.tcl; test/regression_test.sh; test/regression.bzl; LICENSE
28
  Dependencies: OpenROAD built from the same commit b1ad304cd6fc6252d5912e873ef3dd8569f1aea4 (2026-10-07, BSD-3-Clause), default CPU build: the test runs inside the openroad binary's Tcl shell (`openroad -no_splash -no_init -exit -threads 1`, test/regression_test.sh line 33); not packed; bash and diff from the build environment for test/regression_test.sh; not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/tap/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=gcd_nangate45 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/tap/test/. The unit passes when openroad exits 0, the log equals gcd_nangate45.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against gcd_nangate45.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of gcd_nangate45.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
12
 
13
  **Licence (SPDX):** BSD-3-Clause AND NOASSERTION
 
19
  - `src/tap/test/gcd_nangate45.tcl`: OpenROAD tapcell: insert endcaps and well taps every 20 um into gcd, expecting 114 endcaps and 118 tapcells and the golden DEF.
20
  Format: OpenROAD Tcl regression script (.tcl) with golden log (.ok) and golden DEF (.defok); inputs are LEF 5.x library files and a DEF design
21
  Origin: regression test written by the OpenROAD developers for their own placement tools (tool_fixture); the golden .ok and .defok are outputs of an earlier OpenROAD run committed as references.
 
 
22
  Check: Build OpenROAD at the pinned commit (CPU build) in a sandbox without network, restore the upstream layout of src/tap/test/ (helpers.tcl and Nangate45 as links or copies of the test/ targets), and run test/regression_test.sh with TEST_NAME=gcd_nangate45 TEST_EXT=tcl TEST_CHECK_LOG=True from inside src/tap/test/. The unit passes when openroad exits 0, the log equals gcd_nangate45.ok line for line (plain diff, test/regression_test.sh line 69), and the script's own diff_files call prints `No differences found.` for the written DEF against gcd_nangate45.defok (exact line compare, no ignore pattern, test/helpers.tcl lines 171 to 202); that line is also the last line of gcd_nangate45.ok, so the log diff covers it. For a task, give the input DEF and LEF and the target (the .defok or the asserted figures in the .ok) and ask the agent to write the Tcl that produces it; judge with the upstream script's checks.
23
  Companions (not included): src/tap/test/gcd_nangate45.ok; src/tap/test/gcd_nangate45.defok; src/tap/test/gcd_nangate45.def; test/Nangate45/Nangate45_tech.lef; test/Nangate45/Nangate45_stdcell.lef; test/helpers.tcl; test/regression_test.sh; test/regression.bzl; LICENSE
24
  Dependencies: OpenROAD built from the same commit b1ad304cd6fc6252d5912e873ef3dd8569f1aea4 (2026-10-07, BSD-3-Clause), default CPU build: the test runs inside the openroad binary's Tcl shell (`openroad -no_splash -no_init -exit -threads 1`, test/regression_test.sh line 33); not packed; bash and diff from the build environment for test/regression_test.sh; not packed
campaign_files/2026-10-m-issue-trackers-2/openldap-technical/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the JSON with jq or Python's json module; rebuild the tree from the parent field (or fetch the per-email raw mbox from HyperKitty and use mailbox/email modules).
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m329-09/emails.json`: back_mdb: does rtxnsize affect slapcat live backups? (7 messages, 2019-03-29 to 2019-04-01).
22
  Format: HyperKitty REST API thread email list (JSON array, one object per email: url, message_id, message_id_hash, thread, sender (obfuscated address), sender_name, subject, date (ISO 8601 with UTC offset), parent (URL of the parent email or null), children (URLs)). The thread is the whole array; the tree is given by parent.
23
  Origin: Real list traffic on the OpenLDAP openldap-technical list. The unit is one thread as HyperKitty's REST API serves it: a JSON list of the thread's emails with message_id, sender_name, date, parent and children; the thread had its last message on 2019-04-01 and has been closed since. Live API without licence, so a pointer row (general rule 4): two spaced fetches were byte-identical. A late reply to this 2019 thread would change the bytes and break the pin; the pin check at use time catches that.
24
- Reference result: facts computed from the thread headers
25
- Reference source: message_id, parent, sender_name and date fields of the emails in the pinned thread JSON (HyperKitty derives them from the Message-ID, In-Reply-To/References, From and Date headers).
26
  Check: Exact match against values computed from the pinned JSON: the root is the one email with parent null, the thread is the whole array, the tree is given by parent; dates are converted from their UTC offset to UTC; people are counted by sender_name. The first-reply interval is the exact difference in whole seconds between the root date and the earliest other date. Deterministic. The program was checked on 2 threads of this source: the chosen thread, also counted independently from the HyperKitty HTML thread page (7 message anchors, the same 5 sender names) and read message by message, and the thread 'Problem with "force user to password reset at first login' (emails JSON of thread 2T3OFPN4ZNMNNXZ6EHWIMRS4ICZEJPCC, 10 messages, 5 people, fetched once), whose parent chain and counts were checked by hand; message count, last message and first-reply seconds agreed.
27
  Tools: Read the JSON with jq or Python's json module; rebuild the tree from the parent field (or fetch the per-email raw mbox from HyperKitty and use mailbox/email modules).
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the JSON with jq or Python's json module; rebuild the tree from the parent field (or fetch the per-email raw mbox from HyperKitty and use mailbox/email modules).
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m329-09/emails.json`: back_mdb: does rtxnsize affect slapcat live backups? (7 messages, 2019-03-29 to 2019-04-01).
20
  Format: HyperKitty REST API thread email list (JSON array, one object per email: url, message_id, message_id_hash, thread, sender (obfuscated address), sender_name, subject, date (ISO 8601 with UTC offset), parent (URL of the parent email or null), children (URLs)). The thread is the whole array; the tree is given by parent.
21
  Origin: Real list traffic on the OpenLDAP openldap-technical list. The unit is one thread as HyperKitty's REST API serves it: a JSON list of the thread's emails with message_id, sender_name, date, parent and children; the thread had its last message on 2019-04-01 and has been closed since. Live API without licence, so a pointer row (general rule 4): two spaced fetches were byte-identical. A late reply to this 2019 thread would change the bytes and break the pin; the pin check at use time catches that.
 
 
22
  Check: Exact match against values computed from the pinned JSON: the root is the one email with parent null, the thread is the whole array, the tree is given by parent; dates are converted from their UTC offset to UTC; people are counted by sender_name. The first-reply interval is the exact difference in whole seconds between the root date and the earliest other date. Deterministic. The program was checked on 2 threads of this source: the chosen thread, also counted independently from the HyperKitty HTML thread page (7 message anchors, the same 5 sender names) and read message by message, and the thread 'Problem with "force user to password reset at first login' (emails JSON of thread 2T3OFPN4ZNMNNXZ6EHWIMRS4ICZEJPCC, 10 messages, 5 people, fetched once), whose parent chain and counts were checked by hand; message count, last message and first-reply seconds agreed.
23
  Tools: Read the JSON with jq or Python's json module; rebuild the tree from the parent field (or fetch the per-email raw mbox from HyperKitty and use mailbox/email modules).
campaign_files/2026-10-m-issue-trackers-2/opensips-users/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m141-09/2019-March.txt.gz`: [OpenSIPS-Users] RFC2833 to INFO (9 messages, 2019-03-18 to 2019-03-18).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned month file.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and '[OpenSIPS-Users] listen to IPv6' (7 messages, root <CABoceAKMXuQatOfDzUYTfvgkf6zJD8Z0Z7c9wsAG7W5NroLTsw@mail.gmail.com>); message count, people, last sender and time, first-reply seconds and direct replies agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m141-09/2019-March.txt.gz`: [OpenSIPS-Users] RFC2833 to INFO (9 messages, 2019-03-18 to 2019-03-18).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and '[OpenSIPS-Users] listen to IPv6' (7 messages, root <CABoceAKMXuQatOfDzUYTfvgkf6zJD8Z0Z7c9wsAG7W5NroLTsw@mail.gmail.com>); message count, people, last sender and time, first-reply seconds and direct replies agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/openssl-users/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m089-10/2019-March.txt.gz`: OpenVPNGui 2.4.7 fails: format error in certificate's notAfter field (22 messages, 2019-03-04 to 2019-03-08).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned month file.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing double quotes (one archive writes a quoted local part that replies cite unquoted; found in the hand check). People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). Deterministic. The program was checked by hand on 2 threads of this source before giving answers: the chosen thread (read message by message) and 'i2d_X509_REQ() -> d2i_X509_REQ() = asn1 encoding routines' (25 messages by headers, one subject throughout); it agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m089-10/2019-March.txt.gz`: OpenVPNGui 2.4.7 fails: format error in certificate's notAfter field (22 messages, 2019-03-04 to 2019-03-08).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing double quotes (one archive writes a quoted local part that replies cite unquoted; found in the hand check). People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). Deterministic. The program was checked by hand on 2 threads of this source before giving answers: the chosen thread (read message by message) and 'i2d_X509_REQ() -> d2i_X509_REQ() = asn1 encoding routines' (25 messages by headers, one subject throughout); it agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/openvswitch-ovs-discuss/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m089-12/2019-March.txt.gz`: [ovs-discuss] Amazon like networking vpc with ovn (9 messages, 2019-03-07 to 2019-03-29).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned month file.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing double quotes (one archive writes a quoted local part that replies cite unquoted; found in the hand check). People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). Deterministic. The program was checked by hand on 2 threads of this source before giving answers: the chosen thread (read message by message) and 'OVS' (29 messages by headers, one subject throughout); it agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m089-12/2019-March.txt.gz`: [ovs-discuss] Amazon like networking vpc with ovn (9 messages, 2019-03-07 to 2019-03-29).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing double quotes (one archive writes a quoted local part that replies cite unquoted; found in the hand check). People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). Deterministic. The program was checked by hand on 2 threads of this source before giving answers: the chosen thread (read message by message) and 'OVS' (29 messages by headers, one subject throughout); it agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/oran-sc-gerrit-changes/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** curl the Gerrit REST API (GET https://gerrit.o-ran-sc.org/r/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `changes/14257/detail.json`: Aiml-fw/awmf/modelmgmtservice change 14257: adding ProblemDetail model for error message.
22
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
23
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
24
- Reference result: facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments)
25
- Reference source: The pinned response: changes/14257/detail.json.
26
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: oran-jobbuilder, oran.required.gha, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
27
  Tools: curl the Gerrit REST API (GET https://gerrit.o-ran-sc.org/r/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** curl the Gerrit REST API (GET https://gerrit.o-ran-sc.org/r/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `changes/14257/detail.json`: Aiml-fw/awmf/modelmgmtservice change 14257: adding ProblemDetail model for error message.
20
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
21
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
 
 
22
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: oran-jobbuilder, oran.required.gha, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
23
  Tools: curl the Gerrit REST API (GET https://gerrit.o-ran-sc.org/r/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
campaign_files/2026-10-m-issue-trackers-2/osgeo-gdal-dev/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers. For HyperKitty, curl the archive REST API (?format=json) and walk the parent links.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m027-09/2019-March.txt.gz`: [gdal-dev] Segfault in gdalrasterize (5 messages, 2019-03-26 to 2019-03-27).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure. Addresses written as 'x at y'.
23
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. People are counted by person, not address (the row's scale names the people). Deterministic. The thread-cutting program was checked against a hand reading on 2 threads of this list (spec general rule 5): on the chosen thread the header closure and the Subject lines agree message for message; on a second thread every difference was traced by hand to the In-Reply-To headers (a renamed subject inside the reply chain, or a separate root with the same subject), and the program followed the headers correctly. Questions that need the body (what was asked, what fixed it) need a model or human judge and are listed separately. Hand-check second sample: 'Closing remaining open Trac tickets ?' (14 messages by headers; 9 under the renamed subject 'Trac wiki (was: ...)' linked by In-Reply-To).
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers. For HyperKitty, curl the archive REST API (?format=json) and walk the parent links.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers. For HyperKitty, curl the archive REST API (?format=json) and walk the parent links.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m027-09/2019-March.txt.gz`: [gdal-dev] Segfault in gdalrasterize (5 messages, 2019-03-26 to 2019-03-27).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure. Addresses written as 'x at y'.
21
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. People are counted by person, not address (the row's scale names the people). Deterministic. The thread-cutting program was checked against a hand reading on 2 threads of this list (spec general rule 5): on the chosen thread the header closure and the Subject lines agree message for message; on a second thread every difference was traced by hand to the In-Reply-To headers (a renamed subject inside the reply chain, or a separate root with the same subject), and the program followed the headers correctly. Questions that need the body (what was asked, what fixed it) need a model or human judge and are listed separately. Hand-check second sample: 'Closing remaining open Trac tickets ?' (14 messages by headers; 9 under the renamed subject 'Trac wiki (was: ...)' linked by In-Reply-To).
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers. For HyperKitty, curl the archive REST API (?format=json) and walk the parent links.
campaign_files/2026-10-m-issue-trackers-2/osmocom-gerrit-changes/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** curl the Gerrit REST API (GET https://gerrit.osmocom.org/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `changes/40327/detail.json`: Libosmo-sigtran change 40327: asp: Avoid double-free of received msg if conn is teared down.
22
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
23
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
24
- Reference result: facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments)
25
- Reference source: The pinned response: changes/40327/detail.json.
26
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: jenkins, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
27
  Tools: curl the Gerrit REST API (GET https://gerrit.osmocom.org/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** curl the Gerrit REST API (GET https://gerrit.osmocom.org/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `changes/40327/detail.json`: Libosmo-sigtran change 40327: asp: Avoid double-free of received msg if conn is teared down.
20
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
21
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
 
 
22
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: jenkins, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
23
  Tools: curl the Gerrit REST API (GET https://gerrit.osmocom.org/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
campaign_files/2026-10-m-issue-trackers-2/osmocom-openbsc/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m209-08/2019-March.txt.gz`: Libosmocore wishlist (10 messages, 2019-03-20 to 2019-03-26).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real list traffic in a closed past month of the list's archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned month file.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC (a Date header without zone is read as UTC) and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and 'easier way to avoid msgb leaks in libosmo-sccp callers?' (3 messages, root <20190310033105.GA3276@my.box>); message count, last message time, first-reply seconds and the file's distinct message count agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m209-08/2019-March.txt.gz`: Libosmocore wishlist (10 messages, 2019-03-20 to 2019-03-26).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real list traffic in a closed past month of the list's archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC (a Date header without zone is read as UTC) and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and 'easier way to avoid msgb leaks in libosmo-sccp callers?' (3 messages, root <20190310033105.GA3276@my.box>); message count, last message time, first-reply seconds and the file's distinct message count agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/oxfordbus-passenger-gtfs/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the service calendar itself (active services on a date, removed or added dates, day counts). Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
14
 
15
  **Licence (SPDX):** OGL-UK-3.0
@@ -22,8 +20,6 @@ Units:
22
  Read out of the archive: https://s3-eu-west-1.amazonaws.com/passenger-sources/oxfordbus/gtfs/oxfordbus_1790925682.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
24
  Origin: Live static GTFS feed published by Oxford Bus (published through Passenger, data.discoverpassenger.com) at https://s3-eu-west-1.amazonaws.com/passenger-sources/oxfordbus/gtfs/oxfordbus_1790925682.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
25
- Reference result: facts computed from the service calendar itself (active services on a date, removed or added dates, day counts)
26
- Reference source: calendar.txt weekday flags with start_date and end_date, calendar_dates.txt exception rows, agency_timezone of agency.txt, all from the pinned members.
27
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
28
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
29
  Companions (not included): calendar_dates.txt; agency.txt
 
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
12
 
13
  **Licence (SPDX):** OGL-UK-3.0
 
20
  Read out of the archive: https://s3-eu-west-1.amazonaws.com/passenger-sources/oxfordbus/gtfs/oxfordbus_1790925682.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
22
  Origin: Live static GTFS feed published by Oxford Bus (published through Passenger, data.discoverpassenger.com) at https://s3-eu-west-1.amazonaws.com/passenger-sources/oxfordbus/gtfs/oxfordbus_1790925682.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
 
 
23
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
24
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
25
  Companions (not included): calendar_dates.txt; agency.txt
campaign_files/2026-10-m-issue-trackers-2/ozlabs-openbmc/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m089-13/2019-March.txt.gz`: Sensor value calculation formula (12 messages, 2019-03-06 to 2019-03-15).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned month file.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing double quotes (one archive writes a quoted local part that replies cite unquoted; found in the hand check). People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). Deterministic. The program was checked by hand on 2 threads of this source before giving answers: the chosen thread (read message by message) and '[PATCH] ARM: actions: fix a leaked reference by adding missing of_node_put' (a patch series cross-posted to the list: 32 messages by headers, 25 with their own patch titles linked by In-Reply-To); it agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m089-13/2019-March.txt.gz`: Sensor value calculation formula (12 messages, 2019-03-06 to 2019-03-15).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing double quotes (one archive writes a quoted local part that replies cite unquoted; found in the hand check). People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). Deterministic. The program was checked by hand on 2 threads of this source before giving answers: the chosen thread (read message by message) and '[PATCH] ARM: actions: fix a leaked reference by adding missing of_node_put' (a patch series cross-posted to the list: 32 messages by headers, 25 with their own patch titles linked by In-Reply-To); it agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/pace-gtfs/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the service calendar itself (active services on a date, removed or added dates, day counts). Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
14
 
15
  **Licence (SPDX):** LicenseRef-Pace-Limited-License-Agreement
@@ -22,8 +20,6 @@ Units:
22
  Read out of the archive: https://www.pacebus.com/sites/default/files/2026-08/GTFS.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt)
24
  Origin: Live static GTFS feed published by Pace, the Suburban Bus Division of the Regional Transportation Authority at https://www.pacebus.com/sites/default/files/2026-08/GTFS.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
25
- Reference result: facts computed from the service calendar itself (active services on a date, removed or added dates, day counts)
26
- Reference source: calendar.txt weekday flags with start_date and end_date, calendar_dates.txt exception rows, agency_timezone of agency.txt, all from the pinned members.
27
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
28
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
29
  Companions (not included): calendar_dates.txt; agency.txt; feed_info.txt
 
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
12
 
13
  **Licence (SPDX):** LicenseRef-Pace-Limited-License-Agreement
 
20
  Read out of the archive: https://www.pacebus.com/sites/default/files/2026-08/GTFS.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt)
22
  Origin: Live static GTFS feed published by Pace, the Suburban Bus Division of the Regional Transportation Authority at https://www.pacebus.com/sites/default/files/2026-08/GTFS.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
 
 
23
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
24
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
25
  Companions (not included): calendar_dates.txt; agency.txt; feed_info.txt
campaign_files/2026-10-m-issue-trackers-2/path-gtfs/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the service calendar itself (active services on a date, removed or added dates, day counts). Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -22,8 +20,6 @@ Units:
22
  Read out of the archive: http://data.trilliumtransit.com/gtfs/path-nj-us/path-nj-us.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt)
24
  Origin: Live static GTFS feed published by Port Authority of New York and New Jersey (PATH); feed produced by Trillium Solutions at http://data.trilliumtransit.com/gtfs/path-nj-us/path-nj-us.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
25
- Reference result: facts computed from the service calendar itself (active services on a date, removed or added dates, day counts)
26
- Reference source: calendar.txt weekday flags with start_date and end_date, calendar_dates.txt exception rows, agency_timezone of agency.txt, all from the pinned members.
27
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
28
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
29
  Companions (not included): calendar_dates.txt; agency.txt; feed_info.txt
 
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
12
 
13
  **Licence (SPDX):** NOASSERTION
 
20
  Read out of the archive: http://data.trilliumtransit.com/gtfs/path-nj-us/path-nj-us.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt)
22
  Origin: Live static GTFS feed published by Port Authority of New York and New Jersey (PATH); feed produced by Trillium Solutions at http://data.trilliumtransit.com/gtfs/path-nj-us/path-nj-us.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
 
 
23
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
24
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
25
  Companions (not included): calendar_dates.txt; agency.txt; feed_info.txt
campaign_files/2026-10-md-media-and-documents/openslr-70/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Orthographic transcript of one utterance. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Check: Give the agent the wav only (hide the line index); it writes the transcript.
14
 
15
  **Licence (SPDX):** CC-BY-SA-4.0
@@ -22,7 +20,5 @@ Units:
22
  Read out of the archive: https://openslr.trmal.net/resources/70/en_ng_female.zip
23
  Format: WAV (48 kHz, 16-bit, mono) single utterance, a member of the upstream zip
24
  Origin: crowd-sourced read speech recorded by the publisher with volunteers, released as a research corpus
25
- Reference result: orthographic transcript of one utterance
26
- Reference source: line_index_female.tsv line 1160: "Large landscape image is required"; the OpenSLR resource page (about section) reads "This data set contains transcribed high-quality audio" and "The data set has been manually quality checked, but there might still be errors."
27
  Check: give the agent the wav only (hide the line index); it writes the transcript.
28
  Companions (not included): LICENSE; line_index_female.tsv
 
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Give the agent the wav only (hide the line index); it writes the transcript.
12
 
13
  **Licence (SPDX):** CC-BY-SA-4.0
 
20
  Read out of the archive: https://openslr.trmal.net/resources/70/en_ng_female.zip
21
  Format: WAV (48 kHz, 16-bit, mono) single utterance, a member of the upstream zip
22
  Origin: crowd-sourced read speech recorded by the publisher with volunteers, released as a research corpus
 
 
23
  Check: give the agent the wav only (hide the line index); it writes the transcript.
24
  Companions (not included): LICENSE; line_index_female.tsv
campaign_files/2026-10-md-media-and-documents/openslr-83/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 2 of 2 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Orthographic transcript of one utterance. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Check: Give the agent the wav only (hide the line index); it writes the transcript.
14
 
15
  **Licence (SPDX):** CC-BY-SA-4.0
@@ -22,15 +20,11 @@ Units:
22
  Read out of the archive: https://openslr.trmal.net/resources/83/irish_english_male.zip
23
  Format: WAV (48 kHz, 16-bit, mono) single utterance, a member of the upstream zip
24
  Origin: crowd-sourced read speech recorded by the publisher with volunteers, released as a research corpus
25
- Reference result: orthographic transcript of one utterance
26
- Reference source: line_index_all.csv line 2965: "Maw refers to a fish's air bladder"; the OpenSLR resource page (about section) reads "This data set contains transcribed high-quality audio" and "The data set has been manually quality checked, but there might still be errors."
27
  Check: give the agent the wav only (hide the line index); it writes the transcript.
28
  Companions (not included): LICENSE; line_index_all.csv
29
  - `md053-04/wef_02484_00003859772.wav`: OpenSLR 83 Welsh English female utterance wef_02484_00003859772 with transcript.
30
  Read out of the archive: https://openslr.trmal.net/resources/83/welsh_english_female.zip
31
  Format: WAV (48 kHz, 16-bit, mono) single utterance, a member of the upstream zip
32
  Origin: crowd-sourced read speech recorded by the publisher with volunteers, released as a research corpus
33
- Reference result: orthographic transcript of one utterance
34
- Reference source: line_index_all.csv line 1156: "It is twenty degrees with a chance of rain in Ripon"; the OpenSLR resource page (about section) reads "This data set contains transcribed high-quality audio" and "The data set has been manually quality checked, but there might still be errors."
35
  Check: give the agent the wav only (hide the line index); it writes the transcript.
36
  Companions (not included): LICENSE; line_index_all.csv
 
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 2 of 2 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Give the agent the wav only (hide the line index); it writes the transcript.
12
 
13
  **Licence (SPDX):** CC-BY-SA-4.0
 
20
  Read out of the archive: https://openslr.trmal.net/resources/83/irish_english_male.zip
21
  Format: WAV (48 kHz, 16-bit, mono) single utterance, a member of the upstream zip
22
  Origin: crowd-sourced read speech recorded by the publisher with volunteers, released as a research corpus
 
 
23
  Check: give the agent the wav only (hide the line index); it writes the transcript.
24
  Companions (not included): LICENSE; line_index_all.csv
25
  - `md053-04/wef_02484_00003859772.wav`: OpenSLR 83 Welsh English female utterance wef_02484_00003859772 with transcript.
26
  Read out of the archive: https://openslr.trmal.net/resources/83/welsh_english_female.zip
27
  Format: WAV (48 kHz, 16-bit, mono) single utterance, a member of the upstream zip
28
  Origin: crowd-sourced read speech recorded by the publisher with volunteers, released as a research corpus
 
 
29
  Check: give the agent the wav only (hide the line index); it writes the transcript.
30
  Companions (not included): LICENSE; line_index_all.csv
campaign_files/2026-10-md-media-and-documents/org-ruby-html-examples-tables/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected HTML for one Org mode file. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** org-ruby Ruby library (Orgmode::Parser.new(text).to_html) or its org-ruby command line program; emacs --batch with org-html-export-to-html is the native route
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `spec/html_examples/tables.org`: Org-ruby HTML export test: Org mode tables converted to HTML.
22
  Format: Org mode input file and an HTML file with the same base name
23
  Origin: HTML export examples written by the org-ruby maintainers (Org input and the HTML org-ruby must produce); not copied from another suite
24
- Reference result: expected HTML for one Org mode file
25
- Reference source: spec/html_examples/tables.html
26
  Check: show the agent tables.org only; the agent writes the HTML as Orgmode::Parser.new(text, allow_include_files: true).to_html does; pass on exact string equality with tables.html
27
  Tools: org-ruby Ruby library (Orgmode::Parser.new(text).to_html) or its org-ruby command line program; emacs --batch with org-html-export-to-html is the native route
28
  Companions (not included): spec/html_examples/tables.html; spec/parser_spec.rb; README.org
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** org-ruby Ruby library (Orgmode::Parser.new(text).to_html) or its org-ruby command line program; emacs --batch with org-html-export-to-html is the native route
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `spec/html_examples/tables.org`: Org-ruby HTML export test: Org mode tables converted to HTML.
20
  Format: Org mode input file and an HTML file with the same base name
21
  Origin: HTML export examples written by the org-ruby maintainers (Org input and the HTML org-ruby must produce); not copied from another suite
 
 
22
  Check: show the agent tables.org only; the agent writes the HTML as Orgmode::Parser.new(text, allow_include_files: true).to_html does; pass on exact string equality with tables.html
23
  Tools: org-ruby Ruby library (Orgmode::Parser.new(text).to_html) or its org-ruby command line program; emacs --batch with org-html-export-to-html is the native route
24
  Companions (not included): spec/html_examples/tables.html; spec/parser_spec.rb; README.org
campaign_files/2026-10-md-media-and-documents/ox-hugo/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected Hugo Markdown file of an Org post. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** emacs --batch with test/setup-ox-hugo.el loaded, calling org-hugo-export-all-wim-to-md on post-draft.org (make md1), output compared with the committed golden file by diff -r
14
 
15
  **Licence (SPDX):** GPL-3.0-or-later
@@ -21,8 +19,6 @@ Files:
21
  - `test/site/content-org/single-posts/post-draft.org`: Ox-hugo test: Org post marked draft exported to Hugo Markdown with TOML front matter.
22
  Format: Org input and one golden Markdown file
23
  Origin: Org test post and golden Markdown written by the ox-hugo author; the Makefile test target (do_test, diffgolden) is the judge
24
- Reference result: expected Hugo Markdown file of an Org post
25
- Reference source: test/site/content/singles/post-draft.md
26
  Check: give the agent post-draft.org and ask for the Markdown file ox-hugo writes (TOML front matter, headings with ids); compare with the golden file by diff -r; any difference fails
27
  Tools: emacs --batch with test/setup-ox-hugo.el loaded, calling org-hugo-export-all-wim-to-md on post-draft.org (make md1), output compared with the committed golden file by diff -r
28
  Companions (not included): test/site/content/singles/post-draft.md; test/setup-ox-hugo.el; Makefile; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** emacs --batch with test/setup-ox-hugo.el loaded, calling org-hugo-export-all-wim-to-md on post-draft.org (make md1), output compared with the committed golden file by diff -r
12
 
13
  **Licence (SPDX):** GPL-3.0-or-later
 
19
  - `test/site/content-org/single-posts/post-draft.org`: Ox-hugo test: Org post marked draft exported to Hugo Markdown with TOML front matter.
20
  Format: Org input and one golden Markdown file
21
  Origin: Org test post and golden Markdown written by the ox-hugo author; the Makefile test target (do_test, diffgolden) is the judge
 
 
22
  Check: give the agent post-draft.org and ask for the Markdown file ox-hugo writes (TOML front matter, headings with ids); compare with the golden file by diff -r; any difference fails
23
  Tools: emacs --batch with test/setup-ox-hugo.el loaded, calling org-hugo-export-all-wim-to-md on post-draft.org (make md1), output compared with the committed golden file by diff -r
24
  Companions (not included): test/site/content/singles/post-draft.md; test/setup-ox-hugo.el; Makefile; LICENSE
campaign_files/2026-10-md-media-and-documents/pandoc-crossref-m2m-divtable/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected pandoc Markdown after cross-reference numbering. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** pandoc --filter pandoc-crossref -t markdown input.md (or -t latex); the filter is a pandoc JSON filter run from the command line
14
 
15
  **Licence (SPDX):** GPL-2.0-or-later
@@ -21,8 +19,6 @@ Files:
21
  - `test/m2m/divTable/input.md`: Pandoc-crossref test "divTable": tables in crossref divs numbered and captioned in pandoc Markdown.
22
  Format: test directory: input.md (pandoc Markdown), expect.md (expected Markdown output), expect.tex (expected LaTeX, compared only in the flaky build)
23
  Origin: integration test written by the pandoc-crossref maintainers for tables inside crossref divs
24
- Reference result: expected pandoc Markdown after cross-reference numbering
25
- Reference source: expect.md
26
  Check: show the agent input.md (hide expect.md and expect.tex); the agent produces pandoc Markdown with the cross-references resolved; split both texts into lines and pass when every paired line is equal and the counts of non-empty lines match, as the driver does; the LaTeX output is not judged (the upstream compares it only under the FLAKY flag)
27
  Tools: pandoc --filter pandoc-crossref -t markdown input.md (or -t latex); the filter is a pandoc JSON filter run from the command line
28
  Companions (not included): test/m2m/divTable/expect.md; test/m2m/divTable/expect.tex; test/test-integrative.hs
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** pandoc --filter pandoc-crossref -t markdown input.md (or -t latex); the filter is a pandoc JSON filter run from the command line
12
 
13
  **Licence (SPDX):** GPL-2.0-or-later
 
19
  - `test/m2m/divTable/input.md`: Pandoc-crossref test "divTable": tables in crossref divs numbered and captioned in pandoc Markdown.
20
  Format: test directory: input.md (pandoc Markdown), expect.md (expected Markdown output), expect.tex (expected LaTeX, compared only in the flaky build)
21
  Origin: integration test written by the pandoc-crossref maintainers for tables inside crossref divs
 
 
22
  Check: show the agent input.md (hide expect.md and expect.tex); the agent produces pandoc Markdown with the cross-references resolved; split both texts into lines and pass when every paired line is equal and the counts of non-empty lines match, as the driver does; the LaTeX output is not judged (the upstream compares it only under the FLAKY flag)
23
  Tools: pandoc --filter pandoc-crossref -t markdown input.md (or -t latex); the filter is a pandoc JSON filter run from the command line
24
  Companions (not included): test/m2m/divTable/expect.md; test/m2m/divTable/expect.tex; test/test-integrative.hs
campaign_files/2026-10-md-media-and-documents/pandoc-ext-abstract-section/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected Markdown with the abstract and the Sammanfattning section moved into YAML metadata. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** pandoc lua test/runtests.lua test/perevirky (perevir runs the input block through pandoc with filter abstract-section.lua and compares document structure with the output block)
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/perevirky/two-sections.md`: Pandoc abstract-section filter test: two abstract-like sections moved into document metadata.
22
  Format: perevir Markdown test file: a ``` markdown {#input} block and a ``` markdown {#output} block, plus a perevir.yaml with filters and options
23
  Origin: perevir test file written by the filter author in pandoc-ext/abstract-section: prose explanation, a Markdown input block {#input} and an expected Markdown block {#output} in one file
24
- Reference result: expected Markdown with the abstract and the Sammanfattning section moved into YAML metadata
25
- Reference source: the ``` markdown {#output} block of test/perevirky/two-sections.md (lines 41 to 54)
26
  Check: show the agent the {#input} block and ask for the document pandoc makes with abstract-section.lua (abstract and sammanfattning sections become metadata fields); parse its answer and the {#output} block with pandoc and compare the two documents as structures, with soft line breaks ignored (perevir)
27
  Tools: pandoc lua test/runtests.lua test/perevirky (perevir runs the input block through pandoc with filter abstract-section.lua and compares document structure with the output block)
28
  Companions (not included): test/perevirky/perevir.yaml; test/runtests.lua; Makefile; _extensions/abstract-section/abstract-section.lua; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** pandoc lua test/runtests.lua test/perevirky (perevir runs the input block through pandoc with filter abstract-section.lua and compares document structure with the output block)
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/perevirky/two-sections.md`: Pandoc abstract-section filter test: two abstract-like sections moved into document metadata.
20
  Format: perevir Markdown test file: a ``` markdown {#input} block and a ``` markdown {#output} block, plus a perevir.yaml with filters and options
21
  Origin: perevir test file written by the filter author in pandoc-ext/abstract-section: prose explanation, a Markdown input block {#input} and an expected Markdown block {#output} in one file
 
 
22
  Check: show the agent the {#input} block and ask for the document pandoc makes with abstract-section.lua (abstract and sammanfattning sections become metadata fields); parse its answer and the {#output} block with pandoc and compare the two documents as structures, with soft line breaks ignored (perevir)
23
  Tools: pandoc lua test/runtests.lua test/perevirky (perevir runs the input block through pandoc with filter abstract-section.lua and compares document structure with the output block)
24
  Companions (not included): test/perevirky/perevir.yaml; test/runtests.lua; Makefile; _extensions/abstract-section/abstract-section.lua; LICENSE
campaign_files/2026-10-md-media-and-documents/pandoc-ext-diagram/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected HTML in which a dot code block becomes an image reference. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** pandoc --defaults test/test-no-alt-or-caption.yaml (filter diagram.lua runs Graphviz dot), output compared with diff
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/input-no-alt-or-caption.md`: Pandoc diagram filter test: Graphviz code block without alt text or caption converted to an HTML image.
22
  Format: Markdown input with a dot code block, pandoc defaults YAML, one expected HTML file
23
  Origin: test input and expected HTML written by the filter author in pandoc-ext/diagram; the input-dot case was not taken because its graph is copied from the Graphviz gallery
24
- Reference result: expected HTML in which a dot code block becomes an image reference
25
- Reference source: test/expected-no-alt-or-caption.html (the image file name is the SHA1 of the diagram content, 6724c5a4982daa748d11f126ff9a00256ebaf5e1.svg)
26
  Check: give the agent the input and the settings of the YAML (to html, filter diagram); compare its HTML with the expected file using plain diff; the expected img name depends on the dot code, so the agent must keep the filter's naming; any difference fails
27
  Tools: pandoc --defaults test/test-no-alt-or-caption.yaml (filter diagram.lua runs Graphviz dot), output compared with diff
28
  Companions (not included): test/expected-no-alt-or-caption.html; test/test-no-alt-or-caption.yaml; Makefile; _extensions/diagram/diagram.lua; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** pandoc --defaults test/test-no-alt-or-caption.yaml (filter diagram.lua runs Graphviz dot), output compared with diff
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/input-no-alt-or-caption.md`: Pandoc diagram filter test: Graphviz code block without alt text or caption converted to an HTML image.
20
  Format: Markdown input with a dot code block, pandoc defaults YAML, one expected HTML file
21
  Origin: test input and expected HTML written by the filter author in pandoc-ext/diagram; the input-dot case was not taken because its graph is copied from the Graphviz gallery
 
 
22
  Check: give the agent the input and the settings of the YAML (to html, filter diagram); compare its HTML with the expected file using plain diff; the expected img name depends on the dot code, so the agent must keep the filter's naming; any difference fails
23
  Tools: pandoc --defaults test/test-no-alt-or-caption.yaml (filter diagram.lua runs Graphviz dot), output compared with diff
24
  Companions (not included): test/expected-no-alt-or-caption.html; test/test-no-alt-or-caption.yaml; Makefile; _extensions/diagram/diagram.lua; LICENSE
campaign_files/2026-10-md-media-and-documents/pandoc-ext-list-table/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected HTML of a Markdown document whose list-table divs become tables. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** pandoc --lua-filter list-table.lua --to=html test/sample.md, output compared with diff
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/sample.md`: Pandoc list-table filter test: nested bullet lists in a list-table div converted to HTML tables.
22
  Format: Markdown input (fenced divs of class list-table) and one expected HTML fragment
23
  Origin: test input and expected HTML written by the filter authors in pandoc-ext/list-table; the Makefile test target is the judge
24
- Reference result: expected HTML of a Markdown document whose list-table divs become tables
25
- Reference source: test/expected.html
26
  Check: give the agent test/sample.md and ask for the HTML pandoc writes when list-table divs become tables (as pandoc --lua-filter list-table.lua --to=html); compare with test/expected.html using diff --strip-trailing-cr -u; any line difference fails
27
  Tools: pandoc --lua-filter list-table.lua --to=html test/sample.md, output compared with diff
28
  Companions (not included): test/expected.html; Makefile; _extensions/list-table/list-table.lua; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** pandoc --lua-filter list-table.lua --to=html test/sample.md, output compared with diff
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/sample.md`: Pandoc list-table filter test: nested bullet lists in a list-table div converted to HTML tables.
20
  Format: Markdown input (fenced divs of class list-table) and one expected HTML fragment
21
  Origin: test input and expected HTML written by the filter authors in pandoc-ext/list-table; the Makefile test target is the judge
 
 
22
  Check: give the agent test/sample.md and ask for the HTML pandoc writes when list-table divs become tables (as pandoc --lua-filter list-table.lua --to=html); compare with test/expected.html using diff --strip-trailing-cr -u; any line difference fails
23
  Tools: pandoc --lua-filter list-table.lua --to=html test/sample.md, output compared with diff
24
  Companions (not included): test/expected.html; Makefile; _extensions/list-table/list-table.lua; LICENSE
campaign_files/2026-10-md-media-and-documents/pandoc-ext-multibib/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected plain-text document with two separately generated bibliographies. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** pandoc -d test/test.yaml (to plain, standalone, filter multibib.lua, citeproc), output compared with diff
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/input.md`: Pandoc multibib filter test: one Markdown document citing two BibTeX files, rendered to plain text with two reference lists.
22
  Format: Markdown input, two BibTeX files, pandoc defaults YAML and one expected plain-text file
23
  Origin: test input, two BibTeX files and the expected plain-text output written by the filter authors in pandoc-ext/multibib
24
- Reference result: expected plain-text document with two separately generated bibliographies
25
- Reference source: test/expected.txt (test/expected.<PANDOC_VERSION>.txt when present; none at the pinned commit)
26
  Check: give the agent the input, both .bib files and the settings of test/test.yaml (to plain, standalone, multibib); compare its plain-text output with test/expected.txt using plain diff; any difference fails
27
  Tools: pandoc -d test/test.yaml (to plain, standalone, filter multibib.lua, citeproc), output compared with diff
28
  Companions (not included): test/primary.bib; test/secondary.bib; test/test.yaml; test/expected.txt; Makefile; _extensions/multibib/multibib.lua; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** pandoc -d test/test.yaml (to plain, standalone, filter multibib.lua, citeproc), output compared with diff
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/input.md`: Pandoc multibib filter test: one Markdown document citing two BibTeX files, rendered to plain text with two reference lists.
20
  Format: Markdown input, two BibTeX files, pandoc defaults YAML and one expected plain-text file
21
  Origin: test input, two BibTeX files and the expected plain-text output written by the filter authors in pandoc-ext/multibib
 
 
22
  Check: give the agent the input, both .bib files and the settings of test/test.yaml (to plain, standalone, multibib); compare its plain-text output with test/expected.txt using plain diff; any difference fails
23
  Tools: pandoc -d test/test.yaml (to plain, standalone, filter multibib.lua, citeproc), output compared with diff
24
  Companions (not included): test/primary.bib; test/secondary.bib; test/test.yaml; test/expected.txt; Makefile; _extensions/multibib/multibib.lua; LICENSE
campaign_files/2026-10-md-media-and-documents/pandoc-ext-pagebreak/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected troff ms output with page breaks for a Markdown document. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** pandoc --defaults test/test-ms.yaml (input test/input.md, to ms, filter pagebreak.lua, form feed breaks on), output compared with diff
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/input.md`: Pandoc pagebreak filter test: Markdown with \newpage and form-feed page breaks converted to troff ms.
22
  Format: Markdown input, pandoc defaults YAML and one expected ms file
23
  Origin: test input and expected outputs written by the filter authors in pandoc-ext/pagebreak; one expected file per output format (adoc, html, ms, typst), this row takes ms
24
- Reference result: expected troff ms output with page breaks for a Markdown document
25
- Reference source: test/expected.ms
26
  Check: give the agent test/input.md and the settings of test/test-ms.yaml (to ms, form feed counts as a page break); compare its ms output with test/expected.ms using plain diff; any difference fails
27
  Tools: pandoc --defaults test/test-ms.yaml (input test/input.md, to ms, filter pagebreak.lua, form feed breaks on), output compared with diff
28
  Companions (not included): test/expected.ms; test/test-ms.yaml; Makefile; pagebreak.lua; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** pandoc --defaults test/test-ms.yaml (input test/input.md, to ms, filter pagebreak.lua, form feed breaks on), output compared with diff
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/input.md`: Pandoc pagebreak filter test: Markdown with \newpage and form-feed page breaks converted to troff ms.
20
  Format: Markdown input, pandoc defaults YAML and one expected ms file
21
  Origin: test input and expected outputs written by the filter authors in pandoc-ext/pagebreak; one expected file per output format (adoc, html, ms, typst), this row takes ms
 
 
22
  Check: give the agent test/input.md and the settings of test/test-ms.yaml (to ms, form feed counts as a page break); compare its ms output with test/expected.ms using plain diff; any difference fails
23
  Tools: pandoc --defaults test/test-ms.yaml (input test/input.md, to ms, filter pagebreak.lua, form feed breaks on), output compared with diff
24
  Companions (not included): test/expected.ms; test/test-ms.yaml; Makefile; pagebreak.lua; LICENSE
campaign_files/2026-10-md-media-and-documents/pandoc-ext-section-bibliographies/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected pandoc native AST with one bibliography per section. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** pandoc --defaults test/test.yaml --defaults test/test-unnumbered-section.yaml (citeproc, section-bibliographies.lua, unify-refs-divs.lua, to native), output compared with diff
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/input-unnumbered-section.md`: Pandoc section-bibliographies filter test: per-section reference lists in a document with an unnumbered section, as pandoc native AST.
22
  Format: Markdown input, BibTeX file, two pandoc defaults YAML files, one expected native (Haskell-style AST) file
23
  Origin: test input, bibliography and expected pandoc native AST written by the filter author in pandoc-ext/section-bibliographies
24
- Reference result: expected pandoc native AST with one bibliography per section
25
- Reference source: test/expected-unnumbered-section.native
26
  Check: give the agent the input, the bibliography and both defaults files (citeproc, then section-bibliographies with cleanup-first, then unify-refs-divs, to native); compare its native output with the expected file using plain diff; any difference fails
27
  Tools: pandoc --defaults test/test.yaml --defaults test/test-unnumbered-section.yaml (citeproc, section-bibliographies.lua, unify-refs-divs.lua, to native), output compared with diff
28
  Companions (not included): test/test.yaml; test/test-unnumbered-section.yaml; test/expected-unnumbered-section.native; test/bibliography.bib; test/unify-refs-divs.lua; Makefile; _extensions/section-bibliographies/section-bibliographies.lua; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** pandoc --defaults test/test.yaml --defaults test/test-unnumbered-section.yaml (citeproc, section-bibliographies.lua, unify-refs-divs.lua, to native), output compared with diff
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/input-unnumbered-section.md`: Pandoc section-bibliographies filter test: per-section reference lists in a document with an unnumbered section, as pandoc native AST.
20
  Format: Markdown input, BibTeX file, two pandoc defaults YAML files, one expected native (Haskell-style AST) file
21
  Origin: test input, bibliography and expected pandoc native AST written by the filter author in pandoc-ext/section-bibliographies
 
 
22
  Check: give the agent the input, the bibliography and both defaults files (citeproc, then section-bibliographies with cleanup-first, then unify-refs-divs, to native); compare its native output with the expected file using plain diff; any difference fails
23
  Tools: pandoc --defaults test/test.yaml --defaults test/test-unnumbered-section.yaml (citeproc, section-bibliographies.lua, unify-refs-divs.lua, to native), output compared with diff
24
  Companions (not included): test/test.yaml; test/test-unnumbered-section.yaml; test/expected-unnumbered-section.native; test/bibliography.bib; test/unify-refs-divs.lua; Makefile; _extensions/section-bibliographies/section-bibliographies.lua; LICENSE
campaign_files/2026-10-md-media-and-documents/parsoid-parsertests-definition-lists/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Expected HTML for each wikitext case. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** php bin/parse.php --wt2html < input.wikitext (Parsoid standalone command line) or bin/parserTests.php; pandoc -f mediawiki -t html is comparable
14
 
15
  **Licence (SPDX):** GPL-2.0-or-later
@@ -21,8 +19,6 @@ Files:
21
  - `tests/parser/definitionLists.txt`: Parsoid parser tests "definitionLists": MediaWiki wikitext definition lists converted to HTML.
22
  Format: MediaWiki parserTests file (19301 bytes): 52 "!! test" blocks with "!! wikitext" input and "!! html" expected output, plus article definitions (Template:1x and others) used by the tests; companion knownFailures JSON (1049 bytes) lists the 2 cases Parsoid currently fails
23
  Origin: parser test file of the Parsoid wikitext converter, maintained by Wikimedia; covers ; and : list markup, nesting and mixes with other block markup
24
- Reference result: expected HTML for each wikitext case
25
- Reference source: the "!! html" (or "!! html/parsoid") section of each test block
26
  Check: parse the test blocks with Parsoid TestFileReader rules; show only the wikitext; convert wt2html; normalise both sides with Test::normalizeHTML and compare as TestRunner::checkHTML does; cases named in definitionLists-knownFailures.json are expected failures and are left out of scoring
27
  Tools: php bin/parse.php --wt2html < input.wikitext (Parsoid standalone command line) or bin/parserTests.php; pandoc -f mediawiki -t html is comparable
28
  Companions (not included): tests/parser/definitionLists-knownFailures.json
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** php bin/parse.php --wt2html < input.wikitext (Parsoid standalone command line) or bin/parserTests.php; pandoc -f mediawiki -t html is comparable
12
 
13
  **Licence (SPDX):** GPL-2.0-or-later
 
19
  - `tests/parser/definitionLists.txt`: Parsoid parser tests "definitionLists": MediaWiki wikitext definition lists converted to HTML.
20
  Format: MediaWiki parserTests file (19301 bytes): 52 "!! test" blocks with "!! wikitext" input and "!! html" expected output, plus article definitions (Template:1x and others) used by the tests; companion knownFailures JSON (1049 bytes) lists the 2 cases Parsoid currently fails
21
  Origin: parser test file of the Parsoid wikitext converter, maintained by Wikimedia; covers ; and : list markup, nesting and mixes with other block markup
 
 
22
  Check: parse the test blocks with Parsoid TestFileReader rules; show only the wikitext; convert wt2html; normalise both sides with Test::normalizeHTML and compare as TestRunner::checkHTML does; cases named in definitionLists-knownFailures.json are expected failures and are left out of scoring
23
  Tools: php bin/parse.php --wt2html < input.wikitext (Parsoid standalone command line) or bin/parserTests.php; pandoc -f mediawiki -t html is comparable
24
  Companions (not included): tests/parser/definitionLists-knownFailures.json
campaign_files/2026-10-ml-serving-configs-2/openvino-op-reference-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_output_tensors. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Run each TEST_DATA case through the ScaledDotProductAttention operator of the plugin under test (the reference test in scaled_dot_product_attention.cpp builds the model from the header data) and compare the output with ov::test::utils::compare as base_reference_test.cpp line 123 does: every element must satisfy |expected - actual| <= abs_threshold + rel_threshold * |expected|, shapes equal; Build the MulticlassNms model of each case and run it on the plugin under test, then compare each output with the expected tensors as base_reference_test.cpp does (line 123, ov::test::utils::compare for float outputs; integer index and count outputs are compared through the same element rule, so within the default epsilon they must be equal).
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,16 +19,12 @@ Files:
21
  - `src/tests/test_utils/unit_test_utils/tests_data/scaled_dot_product_attention_data.h`: OpenVINO reference test data for ScaledDotProductAttention: 5 TEST_DATA cases (query, key, value, attention mask, scale, causal flag) with expected outputs generated by torch.nn.functional.scaled_dot_product_attention.
22
  Format: C++ header with TEST_DATA macro entries holding shapes and inline float arrays
23
  Origin: reference fixture of the OpenVINO operator set, expected outputs computed with PyTorch by the OpenVINO team
24
- Reference result: expected_output_tensors
25
- Reference source: the expected output arrays inside each TEST_DATA entry of the unit; header line 19: expected output were generated using torch.nn.functional.scaled_dot_product_attention implementation
26
  Check: Run each TEST_DATA case through the ScaledDotProductAttention operator of the plugin under test (the reference test in scaled_dot_product_attention.cpp builds the model from the header data) and compare the output with ov::test::utils::compare as base_reference_test.cpp line 123 does: every element must satisfy |expected - actual| <= abs_threshold + rel_threshold * |expected|, shapes equal
27
  Companions (not included): src/plugins/template/tests/functional/op_reference/scaled_dot_product_attention.cpp; src/plugins/template/tests/functional/op_reference/base_reference_test.hpp; src/plugins/template/tests/functional/op_reference/base_reference_test.cpp; src/tests/test_utils/common_test_utils/src/ov_tensor_utils.cpp; src/tests/test_utils/common_test_utils/include/common_test_utils/ov_tensor_utils.hpp; LICENSE
28
  Dependencies: OpenVINO runtime and its template plugin test build (ov_template_func_tests) at commit 43a09aaf. GoogleTest.
29
  - `src/plugins/template/tests/functional/op_reference/multiclass_nms.cpp`: OpenVINO operator reference test multiclass_nms.cpp: 18 MulticlassNmsParams cases (score and IoU thresholds, nms_top_k, keep_top_k, background class, normalized boxes, sort by class or score, eta, rois_num) with inline boxes, scores and expected selected outputs, indices and counts.
30
  Format: C++ GoogleTest source with inline parameter structs
31
  Origin: reference test of the OpenVINO operator set written by the OpenVINO team
32
- Reference result: expected_output_tensors
33
- Reference source: the expected selected outputs, selected indices and valid output counts written inline in each MulticlassNmsParams entry of the unit (18 entries), written by the OpenVINO team from the MulticlassNms operator specification
34
  Check: Build the MulticlassNms model of each case and run it on the plugin under test, then compare each output with the expected tensors as base_reference_test.cpp does (line 123, ov::test::utils::compare for float outputs; integer index and count outputs are compared through the same element rule, so within the default epsilon they must be equal)
35
  Companions (not included): src/plugins/template/tests/functional/op_reference/base_reference_test.hpp; src/plugins/template/tests/functional/op_reference/base_reference_test.cpp; src/tests/test_utils/common_test_utils/src/ov_tensor_utils.cpp; src/tests/test_utils/common_test_utils/include/common_test_utils/ov_tensor_utils.hpp; LICENSE
36
  Dependencies: OpenVINO runtime and its template plugin test build (ov_template_func_tests) at commit 43a09aaf. GoogleTest.
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Run each TEST_DATA case through the ScaledDotProductAttention operator of the plugin under test (the reference test in scaled_dot_product_attention.cpp builds the model from the header data) and compare the output with ov::test::utils::compare as base_reference_test.cpp line 123 does: every element must satisfy |expected - actual| <= abs_threshold + rel_threshold * |expected|, shapes equal; Build the MulticlassNms model of each case and run it on the plugin under test, then compare each output with the expected tensors as base_reference_test.cpp does (line 123, ov::test::utils::compare for float outputs; integer index and count outputs are compared through the same element rule, so within the default epsilon they must be equal).
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `src/tests/test_utils/unit_test_utils/tests_data/scaled_dot_product_attention_data.h`: OpenVINO reference test data for ScaledDotProductAttention: 5 TEST_DATA cases (query, key, value, attention mask, scale, causal flag) with expected outputs generated by torch.nn.functional.scaled_dot_product_attention.
20
  Format: C++ header with TEST_DATA macro entries holding shapes and inline float arrays
21
  Origin: reference fixture of the OpenVINO operator set, expected outputs computed with PyTorch by the OpenVINO team
 
 
22
  Check: Run each TEST_DATA case through the ScaledDotProductAttention operator of the plugin under test (the reference test in scaled_dot_product_attention.cpp builds the model from the header data) and compare the output with ov::test::utils::compare as base_reference_test.cpp line 123 does: every element must satisfy |expected - actual| <= abs_threshold + rel_threshold * |expected|, shapes equal
23
  Companions (not included): src/plugins/template/tests/functional/op_reference/scaled_dot_product_attention.cpp; src/plugins/template/tests/functional/op_reference/base_reference_test.hpp; src/plugins/template/tests/functional/op_reference/base_reference_test.cpp; src/tests/test_utils/common_test_utils/src/ov_tensor_utils.cpp; src/tests/test_utils/common_test_utils/include/common_test_utils/ov_tensor_utils.hpp; LICENSE
24
  Dependencies: OpenVINO runtime and its template plugin test build (ov_template_func_tests) at commit 43a09aaf. GoogleTest.
25
  - `src/plugins/template/tests/functional/op_reference/multiclass_nms.cpp`: OpenVINO operator reference test multiclass_nms.cpp: 18 MulticlassNmsParams cases (score and IoU thresholds, nms_top_k, keep_top_k, background class, normalized boxes, sort by class or score, eta, rois_num) with inline boxes, scores and expected selected outputs, indices and counts.
26
  Format: C++ GoogleTest source with inline parameter structs
27
  Origin: reference test of the OpenVINO operator set written by the OpenVINO team
 
 
28
  Check: Build the MulticlassNms model of each case and run it on the plugin under test, then compare each output with the expected tensors as base_reference_test.cpp does (line 123, ov::test::utils::compare for float outputs; integer index and count outputs are compared through the same element rule, so within the default epsilon they must be equal)
29
  Companions (not included): src/plugins/template/tests/functional/op_reference/base_reference_test.hpp; src/plugins/template/tests/functional/op_reference/base_reference_test.cpp; src/tests/test_utils/common_test_utils/src/ov_tensor_utils.cpp; src/tests/test_utils/common_test_utils/include/common_test_utils/ov_tensor_utils.hpp; LICENSE
30
  Dependencies: OpenVINO runtime and its template plugin test build (ov_template_func_tests) at commit 43a09aaf. GoogleTest.
campaign_files/2026-10-ml-serving-configs-2/openvinotoolkit-nncf/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each unit needs other upstream files is not stated. 11 of 11 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: aggregate_metric_reference. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Tools are named in the check written per unit in the notes.
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,84 +19,62 @@ Units:
21
  - `examples/post_training_quantization/onnx/mobilenet_v2/` (3 files): NNCF post-training quantization of MobileNetV2 (ONNX backend) on Imagenette, with CI accuracy expectations (fp32 and int8 top-1, accuracy drop).
22
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
23
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
24
- Reference result: aggregate_metric_reference
25
- Reference source: tests/cross_fw/examples/example_scope.json entry post_training_quantization_onnx_mobilenet_v2 (line 2): fp32_top1 0.9864968152866243, int8_top1 0.9844585987261146, accuracy_drop 0.0020382165605096203 (example_scope.json lines 6-10); written by the NNCF maintainers from CI runs on the reference CPU named in the entry
26
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_onnx_mobilenet_v2 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_onnx_mobilenet_v2 with an absolute tolerance of 0.002. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
27
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
28
  Dependencies: onnx==1.22.0, onnxruntime==1.24.3, openvino==2026.4.1, torch==2.14.0, torchvision==0.29.0, fastdownload==0.0.7 per the unit's requirements.txt. data and weights fetched by main.py at run time: Imagenette imagenette2-320.tgz (fastai, s3.amazonaws.com/fast-ai-imageclas, about 340 MB; a 10-class subset of ImageNet; validation split 3,925 images) and the ONNX model from Hugging Face alexsu52/mobilenet_v2_imagenette (licence apache-2.0 per the HF model API), main.py lines 30-31.
29
  - `examples/post_training_quantization/openvino/mobilenet_v2/` (3 files): NNCF post-training quantization of MobileNetV2 (OpenVINO backend) on Imagenette, with CI accuracy expectations (fp32 and int8 top-1, accuracy drop).
30
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
31
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
32
- Reference result: aggregate_metric_reference
33
- Reference source: tests/cross_fw/examples/example_scope.json entry post_training_quantization_openvino_mobilenet_v2_quantize (line 22): fp32_top1 0.9864968152866243, int8_top1 0.9844585987261146, accuracy_drop 0.0020382165605096203 (example_scope.json lines 26-30); written by the NNCF maintainers from CI runs on the reference CPU named in the entry
34
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_openvino_mobilenet_v2_quantize does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_openvino_mobilenet_v2_quantize with an absolute tolerance of 0.002. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
35
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
36
  Dependencies: openvino==2026.4.1, torch==2.14.0, torchvision==0.29.0, fastdownload==0.0.7 per the unit's requirements.txt. data and weights fetched by main.py at run time: Imagenette imagenette2-320.tgz (fastai, about 340 MB) and the OpenVINO IR openvino_model.tgz from Hugging Face alexsu52/mobilenet_v2_imagenette (apache-2.0 per the HF model API), main.py lines 30-31.
37
  - `examples/post_training_quantization/torch/mobilenet_v2/` (3 files): NNCF post-training quantization of MobileNetV2 (PyTorch backend, exported to OpenVINO) on Imagenette, with CI accuracy expectations.
38
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
39
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
40
- Reference result: aggregate_metric_reference
41
- Reference source: tests/cross_fw/examples/example_scope.json entry post_training_quantization_torch_mobilenet_v2 (line 98): fp32_top1 0.9864968152866243, int8_top1 0.9849681528662421, accuracy_drop 0.0015286624203821875 (example_scope.json lines 102-106); written by the NNCF maintainers from CI runs on the reference CPU named in the entry
42
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_torch_mobilenet_v2 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_torch_mobilenet_v2 with an absolute tolerance of 0.002. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
43
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
44
  Dependencies: torch==2.14.0, torchvision==0.29.0, openvino==2026.4.1, fastdownload==0.0.7 per the unit's requirements.txt. data and weights fetched by main.py at run time: Imagenette imagenette2-320.tgz (fastai, about 340 MB) and the checkpoint pytorch_model.bin from Hugging Face alexsu52/mobilenet_v2_imagenette (apache-2.0 per the HF model API), main.py lines 30-31.
45
  - `examples/post_training_quantization/torch/ssd300_vgg16/` (3 files): NNCF post-training quantization of torchvision SSD300-VGG16 on COCO128, with CI mAP expectations.
46
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
47
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
48
- Reference result: aggregate_metric_reference
49
- Reference source: tests/cross_fw/examples/example_scope.json entry post_training_quantization_torch_ssd300_vgg16 (line 118): fp32_mAP 0.5228869318962097, int8_mAP 0.5145288109779358, accuracy_drop 0.0080191493 (example_scope.json lines 122-126; the listed drop differs from fp32 minus int8 by less than the tolerance); written by the NNCF maintainers from CI runs on the reference CPU named in the entry
50
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_torch_ssd300_vgg16 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_torch_ssd300_vgg16 with an absolute tolerance of 0.002. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
51
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
52
  Dependencies: torch==2.14.0, torchvision==0.29.0, torchmetrics==1.0.1, pycocotools==2.0.7, openvino==2026.4.1, onnx==1.22.0 per the unit's requirements.txt. data and weights fetched by main.py at run time: COCO128 coco128.zip from ultralytics.com/assets (128 COCO train2017 images with labels, about 7 MB; COCO images carry per-image Flickr licences, some non-commercial) and torchvision SSD300_VGG16_Weights.COCO_V1 weights, main.py line 34.
53
  - `examples/post_training_quantization/torch_fx/resnet18/` (3 files): NNCF post-training quantization of ResNet18 through torch.fx (OpenVINO backend of torch.compile) on Tiny ImageNet-200, with CI accuracy expectations.
54
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
55
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
56
- Reference result: aggregate_metric_reference
57
- Reference source: tests/cross_fw/examples/example_scope.json entry post_training_quantization_torch_fx_resnet18 (line 138): fp32_top1 55.52, int8_top1 55.27 (percent; example_scope.json entry starting line 138); written by the NNCF maintainers from CI runs on the reference CPU named in the entry
58
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_torch_fx_resnet18 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_torch_fx_resnet18 with an absolute tolerance of 0.2. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
59
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
60
  Dependencies: torch==2.14.0, torchvision==0.29.0, openvino==2026.4.1 per the unit's requirements.txt. data and weights fetched by main.py at run time: Tiny ImageNet-200 tiny-imagenet-200.zip from cs231n.stanford.edu (about 237 MB; 10,000 validation images of 64 px, derived from ImageNet) and the checkpoint 302_resnet18_fp32_v1.pth from storage.openvinotoolkit.org, main.py lines 43-45.
61
  - `examples/quantization_aware_training/torch/resnet18/` (3 files): NNCF quantization-aware training of ResNet18 on Tiny ImageNet-200 (PTQ init, then 2 epochs of QAT), with CI accuracy expectations before and after training.
62
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
63
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
64
- Reference result: aggregate_metric_reference
65
- Reference source: tests/cross_fw/examples/example_scope.json entry quantization_aware_training_torch_resnet18 (line 153): fp32_top1 55.52000045776367, int8_init_top1 55.279998779296875 (tolerance 0.2); after training int8_top1 57.08069620253165, accuracy_drop -1.5606957447679761 (tolerance 1.0); example_scope.json lines 153-167; written by the NNCF maintainers from CI runs on the reference CPU named in the entry
66
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function quantization_aware_training_torch_resnet18 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry quantization_aware_training_torch_resnet18 with an absolute tolerance of 0.2; the metrics after training are compared with tolerance 1.0 (test_examples.py lines 179-182). Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
67
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
68
  Dependencies: torch==2.14.0, torchvision==0.29.0, openvino==2026.4.1 per the unit's requirements.txt. data and weights fetched by main.py at run time: Tiny ImageNet-200 (cs231n.stanford.edu, about 237 MB, derived from ImageNet) and the checkpoint 302_resnet18_fp32_v1.pth from storage.openvinotoolkit.org, main.py lines 48-50; TRAINING_EPOCHS = 2 at main.py line 42.
69
  - `examples/pruning/torch/resnet18/` (3 files): NNCF magnitude pruning of ResNet18 to 70 percent sparsity with pruning-aware fine-tuning on Tiny ImageNet-200, with a CI accuracy expectation.
70
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
71
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
72
- Reference result: aggregate_metric_reference
73
- Reference source: tests/cross_fw/examples/example_scope.json entry magnitude_pruning_torch_resnet18 (line 331): acc1 54.96 (percent; example_scope.json entry starting line 331); written by the NNCF maintainers from CI runs on the reference CPU named in the entry
74
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function magnitude_pruning_torch_resnet18 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry magnitude_pruning_torch_resnet18 with an absolute tolerance of 0.2. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
75
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
76
  Dependencies: torch==2.14.0, torchvision==0.29.0, openvino==2026.4.1 per the unit's requirements.txt. data and weights fetched by main.py at run time: Tiny ImageNet-200 (cs231n.stanford.edu, about 237 MB, derived from ImageNet) and the checkpoint 302_resnet18_fp32_v1.pth from storage.openvinotoolkit.org, main.py lines 45-47; default mode mag at main.py line 57, num_epochs = 2 at line 224.
77
  - `tests/onnx/data/reference_model_accuracy/classification/accuracy_checker-reference.csv`: NNCF ONNX end-to-end reference accuracies: FP32 top-1 of 14 ONNX model-zoo classifiers on ImageNet-1k validation, measured with OpenVINO accuracy_checker configs.
78
  Format: CSV (model, metric, value, abs_threshold, rel_threshold) plus accuracy_checker YAML configs
79
  Origin: reference accuracies kept by a compression library's end-to-end tests
80
- Reference result: aggregate_metric_reference
81
- Reference source: accuracy_checker-reference.csv: FP32 accuracy@top1 per model (fraction, 5 decimals), produced by running accuracy_checker with the YAML config of each model on the ImageNet-1k validation set; reference.json holds the same FP32 target and an INT8 target per model
82
  Check: Pick one model line, run accuracy_checker (open_model_zoo) with that model's YAML config from onnx_models_configs on the 50,000-image ImageNet-1k validation set with onnxruntime, read accuracy@top1, and compare with the CSV value to half a unit of its last decimal. The CSV's own abs_threshold and rel_threshold columns are 0. The INT8 check of reference.json (window -1.0 to 1.0 percentage points around target_int8, test_e2e_ptq.py lines 505-512) needs NNCF quantization first and is optional.
83
  Companions (not included): tests/onnx/data/reference_model_accuracy/reference.json; tests/onnx/benchmarking/classification/onnx_models_configs/bvlcalexnet-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/caffenet-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/densenet-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/efficientnet-lite4-11.yml; tests/onnx/benchmarking/classification/onnx_models_configs/googlenet-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/inception-v1-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/mobilenetv2-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/resnet50-v1-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/resnet50-v2-7.yml; tests/onnx/benchmarking/classification/onnx_models_configs/shufflenet-9.yml; tests/onnx/benchmarking/classification/onnx_models_configs/shufflenet-v2-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/squeezenet1.0-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/vgg16-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/zfnet512-12.yml; tests/onnx/test_e2e_ptq.py; tests/onnx/benchmarking/ac_wrapper.py; tests/onnx/benchmarking/README.md; LICENSE
84
  - `tests/post_training/data/ptq_reference_data.yaml`: NNCF post-training quantization conformance references: ImageNet-1k top-1 of torchvision and timm models per backend (FP32, OpenVINO, ONNX, PyTorch, torch.fx).
85
  Format: YAML (entry name to metric_value)
86
  Origin: conformance references kept by a compression library's nightly tests
87
- Reference result: aggregate_metric_reference
88
- Reference source: ptq_reference_data.yaml: top-1 accuracy per model and backend on the ImageNet-1k validation set, written by the NNCF maintainers from conformance runs; FP32 entries are the unquantized baselines
89
  Check: Pick one non-null entry of a CPU backend (for example timm/resnet18_backend_OV or torchvision/resnet18_backend_FP32), run tests/post_training/test_quantize_conformance.py for that model and backend at the pinned commit on the pinned imagenet-1k revision, and compare the measured top-1 with metric_value by np.isclose with the default atol (base.py lines 360-361; the YAML sets no per-entry atol). Entries with metric_value null and CUDA backends are excluded.
90
  Companions (not included): tests/post_training/model_scope.py; tests/post_training/pipelines/base.py; tests/post_training/pipelines/image_classification_base.py; tests/post_training/pipelines/image_classification_timm.py; tests/post_training/pipelines/image_classification_torchvision.py; tests/post_training/test_quantize_conformance.py; tests/post_training/conftest.py; tests/post_training/README.md; tests/post_training/requirements.txt; LICENSE
91
  - `tests/post_training/data/wc_reference_data.yaml`: NNCF weight-compression conformance references: whowhatbench similarity of int4 and int8 compressed TinyLlama-1.1B to its FP32 answers, with exact counts of compressed layers.
92
  Format: YAML (entry name to metric_value, num_int4, num_int8, atol)
93
  Origin: conformance references kept by a compression library's nightly tests
94
- Reference result: aggregate_metric_reference
95
- Reference source: wc_reference_data.yaml: whowhatbench similarity of the compressed model's answers to the FP32 reference answers in ref_qa.csv (lm_weight_compression.py lines 393-440, rounded to 5 decimals), and the exact numbers of int4 and int8 layers after compression (lines 469-484)
96
  Check: Pick one OV-backend entry (for example tinyllama_data_free_backend_OV), run test_quantize_conformance.py for it at the pinned commit; compare the similarity with metric_value by np.isclose with the entry's atol or the default 1e-4, and compare num_int4 and num_int8 for exact equality. The entry with atol 0.2 (yaml line 60) is loose on a 0-1 scale and is flagged.
97
  Companions (not included): tests/post_training/data/wwb_ref_answers/tinyllama__tinyllama-1.1b-step-50k-105b/ref_qa.csv; tests/post_training/pipelines/lm_weight_compression.py; tests/post_training/model_scope.py; tests/post_training/pipelines/base.py; tests/post_training/test_quantize_conformance.py; tests/post_training/conftest.py; tests/post_training/README.md; tests/post_training/requirements.txt; LICENSE
98
  - `tests/post_training/experimental/sparsify_activations/reference_data.yaml`: NNCF activation-sparsification conformance references: TinyLlama similarity and DeiT3-small ImageNet-1k top-1 after sparsifying activations, with exact counts of sparsified and int8 layers.
99
  Format: YAML (entry name to metric_value, atol, num_int4, num_int8, num_sparse_activations)
100
  Origin: conformance references kept by a compression library's nightly tests
101
- Reference result: aggregate_metric_reference
102
- Reference source: reference_data.yaml: for TinyLlama the whowhatbench similarity to the FP32 answers (as in the weight-compression pipeline it extends, pipelines.py line 160), for DeiT3-small the ImageNet-1k top-1 (pipelines.py lines 280-285), plus the exact numbers of sparsified activations and int8 layers, written by the NNCF maintainers from conformance runs
103
  Check: Pick one TORCH (CPU) entry, run test_sparsify_activations_conformance.py for it at the pinned commit; compare the metric with metric_value by np.isclose with the entry's atol (base.py lines 360-361), and compare num_int4, num_int8 and num_sparse_activations for exact equality (pipelines.py lines 141-158). CUDA_TORCH entries are excluded.
104
  Companions (not included): tests/post_training/experimental/sparsify_activations/pipelines.py; tests/post_training/experimental/sparsify_activations/model_scope.py; tests/post_training/experimental/sparsify_activations/test_sparsify_activations_conformance.py; tests/post_training/pipelines/lm_weight_compression.py; tests/post_training/pipelines/image_classification_base.py; tests/post_training/pipelines/image_classification_timm.py; tests/post_training/pipelines/base.py; tests/post_training/data/wwb_ref_answers/tinyllama__tinyllama-1.1b-step-50k-105b/ref_qa.csv; tests/post_training/requirements.txt; LICENSE
 
8
 
9
  **Self-contained:** Whether each unit needs other upstream files is not stated. 11 of 11 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Tools are named in the check written per unit in the notes.
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `examples/post_training_quantization/onnx/mobilenet_v2/` (3 files): NNCF post-training quantization of MobileNetV2 (ONNX backend) on Imagenette, with CI accuracy expectations (fp32 and int8 top-1, accuracy drop).
20
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
21
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
 
 
22
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_onnx_mobilenet_v2 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_onnx_mobilenet_v2 with an absolute tolerance of 0.002. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
23
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
24
  Dependencies: onnx==1.22.0, onnxruntime==1.24.3, openvino==2026.4.1, torch==2.14.0, torchvision==0.29.0, fastdownload==0.0.7 per the unit's requirements.txt. data and weights fetched by main.py at run time: Imagenette imagenette2-320.tgz (fastai, s3.amazonaws.com/fast-ai-imageclas, about 340 MB; a 10-class subset of ImageNet; validation split 3,925 images) and the ONNX model from Hugging Face alexsu52/mobilenet_v2_imagenette (licence apache-2.0 per the HF model API), main.py lines 30-31.
25
  - `examples/post_training_quantization/openvino/mobilenet_v2/` (3 files): NNCF post-training quantization of MobileNetV2 (OpenVINO backend) on Imagenette, with CI accuracy expectations (fp32 and int8 top-1, accuracy drop).
26
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
27
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
 
 
28
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_openvino_mobilenet_v2_quantize does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_openvino_mobilenet_v2_quantize with an absolute tolerance of 0.002. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
29
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
30
  Dependencies: openvino==2026.4.1, torch==2.14.0, torchvision==0.29.0, fastdownload==0.0.7 per the unit's requirements.txt. data and weights fetched by main.py at run time: Imagenette imagenette2-320.tgz (fastai, about 340 MB) and the OpenVINO IR openvino_model.tgz from Hugging Face alexsu52/mobilenet_v2_imagenette (apache-2.0 per the HF model API), main.py lines 30-31.
31
  - `examples/post_training_quantization/torch/mobilenet_v2/` (3 files): NNCF post-training quantization of MobileNetV2 (PyTorch backend, exported to OpenVINO) on Imagenette, with CI accuracy expectations.
32
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
33
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
 
 
34
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_torch_mobilenet_v2 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_torch_mobilenet_v2 with an absolute tolerance of 0.002. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
35
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
36
  Dependencies: torch==2.14.0, torchvision==0.29.0, openvino==2026.4.1, fastdownload==0.0.7 per the unit's requirements.txt. data and weights fetched by main.py at run time: Imagenette imagenette2-320.tgz (fastai, about 340 MB) and the checkpoint pytorch_model.bin from Hugging Face alexsu52/mobilenet_v2_imagenette (apache-2.0 per the HF model API), main.py lines 30-31.
37
  - `examples/post_training_quantization/torch/ssd300_vgg16/` (3 files): NNCF post-training quantization of torchvision SSD300-VGG16 on COCO128, with CI mAP expectations.
38
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
39
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
 
 
40
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_torch_ssd300_vgg16 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_torch_ssd300_vgg16 with an absolute tolerance of 0.002. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
41
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
42
  Dependencies: torch==2.14.0, torchvision==0.29.0, torchmetrics==1.0.1, pycocotools==2.0.7, openvino==2026.4.1, onnx==1.22.0 per the unit's requirements.txt. data and weights fetched by main.py at run time: COCO128 coco128.zip from ultralytics.com/assets (128 COCO train2017 images with labels, about 7 MB; COCO images carry per-image Flickr licences, some non-commercial) and torchvision SSD300_VGG16_Weights.COCO_V1 weights, main.py line 34.
43
  - `examples/post_training_quantization/torch_fx/resnet18/` (3 files): NNCF post-training quantization of ResNet18 through torch.fx (OpenVINO backend of torch.compile) on Tiny ImageNet-200, with CI accuracy expectations.
44
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
45
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
 
 
46
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function post_training_quantization_torch_fx_resnet18 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry post_training_quantization_torch_fx_resnet18 with an absolute tolerance of 0.2. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
47
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
48
  Dependencies: torch==2.14.0, torchvision==0.29.0, openvino==2026.4.1 per the unit's requirements.txt. data and weights fetched by main.py at run time: Tiny ImageNet-200 tiny-imagenet-200.zip from cs231n.stanford.edu (about 237 MB; 10,000 validation images of 64 px, derived from ImageNet) and the checkpoint 302_resnet18_fp32_v1.pth from storage.openvinotoolkit.org, main.py lines 43-45.
49
  - `examples/quantization_aware_training/torch/resnet18/` (3 files): NNCF quantization-aware training of ResNet18 on Tiny ImageNet-200 (PTQ init, then 2 epochs of QAT), with CI accuracy expectations before and after training.
50
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
51
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
 
 
52
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function quantization_aware_training_torch_resnet18 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry quantization_aware_training_torch_resnet18 with an absolute tolerance of 0.2; the metrics after training are compared with tolerance 1.0 (test_examples.py lines 179-182). Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
53
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
54
  Dependencies: torch==2.14.0, torchvision==0.29.0, openvino==2026.4.1 per the unit's requirements.txt. data and weights fetched by main.py at run time: Tiny ImageNet-200 (cs231n.stanford.edu, about 237 MB, derived from ImageNet) and the checkpoint 302_resnet18_fp32_v1.pth from storage.openvinotoolkit.org, main.py lines 48-50; TRAINING_EPOCHS = 2 at main.py line 42.
55
  - `examples/pruning/torch/resnet18/` (3 files): NNCF magnitude pruning of ResNet18 to 70 percent sparsity with pruning-aware fine-tuning on Tiny ImageNet-200, with a CI accuracy expectation.
56
  Format: Python example directory (main.py, README.md, requirements.txt) plus a JSON expectation entry
57
  Origin: CI expectations of a model-compression library, recorded from its maintainers' runs of the published examples
 
 
58
  Check: Run the unit's main.py at the pinned commit (as tests/cross_fw/examples/run_example.py function magnitude_pruning_torch_resnet18 does, seed 42 where the example trains) on the named data; it writes the metrics; compare each accuracy metric listed in example_scope.json entry magnitude_pruning_torch_resnet18 with an absolute tolerance of 0.2. Performance and model-size metrics (rel 0.05, test_examples.py lines 34-35, 187, 191) are hardware-bound or descriptive and are not used.
59
  Companions (not included): tests/cross_fw/examples/example_scope.json; tests/cross_fw/examples/test_examples.py; tests/cross_fw/examples/run_example.py; tests/cross_fw/examples/README.md; LICENSE
60
  Dependencies: torch==2.14.0, torchvision==0.29.0, openvino==2026.4.1 per the unit's requirements.txt. data and weights fetched by main.py at run time: Tiny ImageNet-200 (cs231n.stanford.edu, about 237 MB, derived from ImageNet) and the checkpoint 302_resnet18_fp32_v1.pth from storage.openvinotoolkit.org, main.py lines 45-47; default mode mag at main.py line 57, num_epochs = 2 at line 224.
61
  - `tests/onnx/data/reference_model_accuracy/classification/accuracy_checker-reference.csv`: NNCF ONNX end-to-end reference accuracies: FP32 top-1 of 14 ONNX model-zoo classifiers on ImageNet-1k validation, measured with OpenVINO accuracy_checker configs.
62
  Format: CSV (model, metric, value, abs_threshold, rel_threshold) plus accuracy_checker YAML configs
63
  Origin: reference accuracies kept by a compression library's end-to-end tests
 
 
64
  Check: Pick one model line, run accuracy_checker (open_model_zoo) with that model's YAML config from onnx_models_configs on the 50,000-image ImageNet-1k validation set with onnxruntime, read accuracy@top1, and compare with the CSV value to half a unit of its last decimal. The CSV's own abs_threshold and rel_threshold columns are 0. The INT8 check of reference.json (window -1.0 to 1.0 percentage points around target_int8, test_e2e_ptq.py lines 505-512) needs NNCF quantization first and is optional.
65
  Companions (not included): tests/onnx/data/reference_model_accuracy/reference.json; tests/onnx/benchmarking/classification/onnx_models_configs/bvlcalexnet-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/caffenet-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/densenet-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/efficientnet-lite4-11.yml; tests/onnx/benchmarking/classification/onnx_models_configs/googlenet-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/inception-v1-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/mobilenetv2-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/resnet50-v1-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/resnet50-v2-7.yml; tests/onnx/benchmarking/classification/onnx_models_configs/shufflenet-9.yml; tests/onnx/benchmarking/classification/onnx_models_configs/shufflenet-v2-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/squeezenet1.0-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/vgg16-12.yml; tests/onnx/benchmarking/classification/onnx_models_configs/zfnet512-12.yml; tests/onnx/test_e2e_ptq.py; tests/onnx/benchmarking/ac_wrapper.py; tests/onnx/benchmarking/README.md; LICENSE
66
  - `tests/post_training/data/ptq_reference_data.yaml`: NNCF post-training quantization conformance references: ImageNet-1k top-1 of torchvision and timm models per backend (FP32, OpenVINO, ONNX, PyTorch, torch.fx).
67
  Format: YAML (entry name to metric_value)
68
  Origin: conformance references kept by a compression library's nightly tests
 
 
69
  Check: Pick one non-null entry of a CPU backend (for example timm/resnet18_backend_OV or torchvision/resnet18_backend_FP32), run tests/post_training/test_quantize_conformance.py for that model and backend at the pinned commit on the pinned imagenet-1k revision, and compare the measured top-1 with metric_value by np.isclose with the default atol (base.py lines 360-361; the YAML sets no per-entry atol). Entries with metric_value null and CUDA backends are excluded.
70
  Companions (not included): tests/post_training/model_scope.py; tests/post_training/pipelines/base.py; tests/post_training/pipelines/image_classification_base.py; tests/post_training/pipelines/image_classification_timm.py; tests/post_training/pipelines/image_classification_torchvision.py; tests/post_training/test_quantize_conformance.py; tests/post_training/conftest.py; tests/post_training/README.md; tests/post_training/requirements.txt; LICENSE
71
  - `tests/post_training/data/wc_reference_data.yaml`: NNCF weight-compression conformance references: whowhatbench similarity of int4 and int8 compressed TinyLlama-1.1B to its FP32 answers, with exact counts of compressed layers.
72
  Format: YAML (entry name to metric_value, num_int4, num_int8, atol)
73
  Origin: conformance references kept by a compression library's nightly tests
 
 
74
  Check: Pick one OV-backend entry (for example tinyllama_data_free_backend_OV), run test_quantize_conformance.py for it at the pinned commit; compare the similarity with metric_value by np.isclose with the entry's atol or the default 1e-4, and compare num_int4 and num_int8 for exact equality. The entry with atol 0.2 (yaml line 60) is loose on a 0-1 scale and is flagged.
75
  Companions (not included): tests/post_training/data/wwb_ref_answers/tinyllama__tinyllama-1.1b-step-50k-105b/ref_qa.csv; tests/post_training/pipelines/lm_weight_compression.py; tests/post_training/model_scope.py; tests/post_training/pipelines/base.py; tests/post_training/test_quantize_conformance.py; tests/post_training/conftest.py; tests/post_training/README.md; tests/post_training/requirements.txt; LICENSE
76
  - `tests/post_training/experimental/sparsify_activations/reference_data.yaml`: NNCF activation-sparsification conformance references: TinyLlama similarity and DeiT3-small ImageNet-1k top-1 after sparsifying activations, with exact counts of sparsified and int8 layers.
77
  Format: YAML (entry name to metric_value, atol, num_int4, num_int8, num_sparse_activations)
78
  Origin: conformance references kept by a compression library's nightly tests
 
 
79
  Check: Pick one TORCH (CPU) entry, run test_sparsify_activations_conformance.py for it at the pinned commit; compare the metric with metric_value by np.isclose with the entry's atol (base.py lines 360-361), and compare num_int4, num_int8 and num_sparse_activations for exact equality (pipelines.py lines 141-158). CUDA_TORCH entries are excluded.
80
  Companions (not included): tests/post_training/experimental/sparsify_activations/pipelines.py; tests/post_training/experimental/sparsify_activations/model_scope.py; tests/post_training/experimental/sparsify_activations/test_sparsify_activations_conformance.py; tests/post_training/pipelines/lm_weight_compression.py; tests/post_training/pipelines/image_classification_base.py; tests/post_training/pipelines/image_classification_timm.py; tests/post_training/pipelines/base.py; tests/post_training/data/wwb_ref_answers/tinyllama__tinyllama-1.1b-step-50k-105b/ref_qa.csv; tests/post_training/requirements.txt; LICENSE
campaign_files/2026-10-ml-serving-configs-2/outlines-core-json-schema-regex-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether the file needs other upstream files is not stated. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_test_outputs. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: In the outlines-core crate at tag 0.2.14 run cargo test --lib json_schema on CPU. Each tuple asserts the built regex equals the expected string, and should_match and should_not_match (lines 207 to 233) check full matches with the regex crate. Hide the expected regex strings and the match lists; comparisons are exact string equality and full-match booleans; there is no numeric tolerance.
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `src/json_schema/mod.rs`: Outlines-core JSON Schema to regex: 69 schemas with the exact regular expression the converter must build and JSON strings it must and must not fully match.
22
  Format: Rust source module whose #[cfg(test)] block (lines 200 to the end) holds schema, expected regex, matching and non-matching strings as tuples inline
23
  Origin: expectations written by the outlines-core maintainers into the test sources
24
- Reference result: expected_test_outputs
25
- Reference source: the expected regex string and the a_match and not_a_match lists of each tuple of test_schema_matches_regex (from line 236), written by the maintainers
26
  Check: In the outlines-core crate at tag 0.2.14 run cargo test --lib json_schema on CPU. Each tuple asserts the built regex equals the expected string, and should_match and should_not_match (lines 207 to 233) check full matches with the regex crate. Hide the expected regex strings and the match lists; comparisons are exact string equality and full-match booleans; there is no numeric tolerance.
27
  Companions (not included): LICENSE
28
  Dependencies: outlines-core source at tag 0.2.14 (Rust crate with serde_json and regex); Rust stable with cargo; the library part needs no tokenizer; not bundled
 
8
 
9
  **Self-contained:** Whether the file needs other upstream files is not stated. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: In the outlines-core crate at tag 0.2.14 run cargo test --lib json_schema on CPU. Each tuple asserts the built regex equals the expected string, and should_match and should_not_match (lines 207 to 233) check full matches with the regex crate. Hide the expected regex strings and the match lists; comparisons are exact string equality and full-match booleans; there is no numeric tolerance.
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `src/json_schema/mod.rs`: Outlines-core JSON Schema to regex: 69 schemas with the exact regular expression the converter must build and JSON strings it must and must not fully match.
20
  Format: Rust source module whose #[cfg(test)] block (lines 200 to the end) holds schema, expected regex, matching and non-matching strings as tuples inline
21
  Origin: expectations written by the outlines-core maintainers into the test sources
 
 
22
  Check: In the outlines-core crate at tag 0.2.14 run cargo test --lib json_schema on CPU. Each tuple asserts the built regex equals the expected string, and should_match and should_not_match (lines 207 to 233) check full matches with the regex crate. Hide the expected regex strings and the match lists; comparisons are exact string equality and full-match booleans; there is no numeric tolerance.
23
  Companions (not included): LICENSE
24
  Dependencies: outlines-core source at tag 0.2.14 (Rust crate with serde_json and regex); Rust stable with cargo; the library part needs no tokenizer; not bundled
campaign_files/2026-10-op-rule-calculations-2/openpeppol-peppol-bis-invoice-3/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained. 10 of 10 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_validation_outcome_per_test. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R040 fires. Other files: see the notes.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,80 +19,60 @@ Files:
21
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R040.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R040: allowance or charge Amount equals BaseAmount * MultiplierFactorNumeric / 100 when both are given, within 0.02.
22
  Format: XML (VEFA validator testSet), 10 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
23
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
24
- Reference result: expected_validation_outcome_per_test
25
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R040</success> or <error>PEPPOL-EN16931-R040</error>; the file has 7 success and 3 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 325
26
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R040 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
27
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R040 fires
28
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
29
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R044.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R044: a price-level AllowanceCharge must have ChargeIndicator false (charges on price level are not allowed).
30
  Format: XML (VEFA validator testSet), 5 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
31
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
32
- Reference result: expected_validation_outcome_per_test
33
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R044</success> or <error>PEPPOL-EN16931-R044</error>; the file has 3 success and 2 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 416
34
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R044 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
35
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R044 fires
36
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
37
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R046.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R046: item net price equals gross price (BaseAmount) minus price allowance Amount when gross price is given, exact decimal equality.
38
  Format: XML (VEFA validator testSet), 4 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
39
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
40
- Reference result: expected_validation_outcome_per_test
41
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R046</success> or <error>PEPPOL-EN16931-R046</error>; the file has 3 success and 1 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 418
42
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R046 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
43
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R046 fires
44
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
45
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R051.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R051: every currencyID attribute on amounts equals the document currency code (BT-5).
46
  Format: XML (VEFA validator testSet), 6 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
47
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
48
- Reference result: expected_validation_outcome_per_test
49
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R051</success> or <error>PEPPOL-EN16931-R051</error>; the file has 1 success and 5 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 346
50
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R051 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
51
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R051 fires
52
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
53
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R053.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R053: exactly one TaxTotal with TaxSubtotal elements.
54
  Format: XML (VEFA validator testSet), 3 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
55
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
56
- Reference result: expected_validation_outcome_per_test
57
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R053</success> or <error>PEPPOL-EN16931-R053</error>; the file has 1 success and 2 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 293
58
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R053 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
59
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R053 fires
60
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
61
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R110.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R110: start date of an invoice line period is on or after the invoice period start date.
62
  Format: XML (VEFA validator testSet), 6 tests, each an inline UBL CreditNote / Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
63
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
64
- Reference result: expected_validation_outcome_per_test
65
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R110</success> or <error>PEPPOL-EN16931-R110</error>; the file has 4 success and 2 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 351
66
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R110 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
67
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R110 fires
68
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
69
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R111.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R111: end date of an invoice line period is on or before the invoice period end date.
70
  Format: XML (VEFA validator testSet), 6 tests, each an inline UBL CreditNote / Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
71
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
72
- Reference result: expected_validation_outcome_per_test
73
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R111</success> or <error>PEPPOL-EN16931-R111</error>; the file has 4 success and 2 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 356
74
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R111 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
75
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R111 fires
76
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
77
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R120.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R120: invoice line net amount equals quantity * (price / base quantity) + line charges - line allowances, within 0.02 (charge and allowance sums rounded to 2 decimals).
78
  Format: XML (VEFA validator testSet), 11 tests, each an inline UBL CreditNote / Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
79
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
80
- Reference result: expected_validation_outcome_per_test
81
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R120</success> or <error>PEPPOL-EN16931-R120</error>; the file has 7 success and 4 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 403
82
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R120 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
83
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R120 fires
84
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
85
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R121.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R121: price BaseQuantity, when given, is greater than zero.
86
  Format: XML (VEFA validator testSet), 4 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
87
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
88
- Reference result: expected_validation_outcome_per_test
89
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R121</success> or <error>PEPPOL-EN16931-R121</error>; the file has 2 success and 2 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 406
90
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R121 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
91
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R121 fires
92
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
93
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R130.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R130: unit code of the price BaseQuantity equals the unit code of the invoiced (or credited) quantity.
94
  Format: XML (VEFA validator testSet), 3 tests, each an inline UBL CreditNote / Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
95
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
96
- Reference result: expected_validation_outcome_per_test
97
- Reference source: each <test> carries <assert><success>PEPPOL-EN16931-R130</success> or <error>PEPPOL-EN16931-R130</error>; the file has 1 success and 2 error expectations; the rule text is PEPPOL-EN16931-UBL.sch line 431
98
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R130 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
99
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R130 fires
100
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
 
8
 
9
  **Self-contained:** Each file is self-contained. 10 of 10 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R040 fires. Other files: see the notes.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R040.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R040: allowance or charge Amount equals BaseAmount * MultiplierFactorNumeric / 100 when both are given, within 0.02.
20
  Format: XML (VEFA validator testSet), 10 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
21
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
22
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R040 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
23
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R040 fires
24
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
25
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R044.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R044: a price-level AllowanceCharge must have ChargeIndicator false (charges on price level are not allowed).
26
  Format: XML (VEFA validator testSet), 5 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
27
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
28
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R044 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
29
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R044 fires
30
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
31
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R046.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R046: item net price equals gross price (BaseAmount) minus price allowance Amount when gross price is given, exact decimal equality.
32
  Format: XML (VEFA validator testSet), 4 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
33
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
34
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R046 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
35
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R046 fires
36
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
37
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R051.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R051: every currencyID attribute on amounts equals the document currency code (BT-5).
38
  Format: XML (VEFA validator testSet), 6 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
39
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
40
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R051 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
41
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R051 fires
42
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
43
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R053.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R053: exactly one TaxTotal with TaxSubtotal elements.
44
  Format: XML (VEFA validator testSet), 3 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
45
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
46
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R053 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
47
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R053 fires
48
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
49
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R110.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R110: start date of an invoice line period is on or after the invoice period start date.
50
  Format: XML (VEFA validator testSet), 6 tests, each an inline UBL CreditNote / Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
51
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
52
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R110 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
53
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R110 fires
54
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
55
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R111.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R111: end date of an invoice line period is on or before the invoice period end date.
56
  Format: XML (VEFA validator testSet), 6 tests, each an inline UBL CreditNote / Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
57
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
58
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R111 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
59
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R111 fires
60
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
61
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R120.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R120: invoice line net amount equals quantity * (price / base quantity) + line charges - line allowances, within 0.02 (charge and allowance sums rounded to 2 decimals).
62
  Format: XML (VEFA validator testSet), 11 tests, each an inline UBL CreditNote / Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
63
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
64
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R120 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
65
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R120 fires
66
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
67
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R121.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R121: price BaseQuantity, when given, is greater than zero.
68
  Format: XML (VEFA validator testSet), 4 tests, each an inline UBL Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
69
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
70
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R121 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
71
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R121 fires
72
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch
73
  - `rules/unit-UBL-PEPPOL/PEPPOL-EN16931-R130.xml`: Peppol BIS Billing 3.0 unit tests for rule PEPPOL-EN16931-R130: unit code of the price BaseQuantity equals the unit code of the invoiced (or credited) quantity.
74
  Format: XML (VEFA validator testSet), 3 tests, each an inline UBL CreditNote / Invoice fragment with the expected outcome; companion: the Peppol Schematron rule file (release 3.0.20)
75
  Origin: unit tests written by the Peppol BIS maintainers to test their Schematron rule; fragments are hand-built, not real invoices
 
 
76
  Check: hide the <assert> blocks; give the agent the fragments and the rule text (or the whole Schematron) and ask for each test whether rule PEPPOL-EN16931-R130 passes or fails; compare with the expected outcome exactly, score = share of tests right; reference run: the Schematron at the same commit through an XSLT 2.0 Schematron engine
77
  Tools: parse each UBL Invoice or CreditNote fragment with Python lxml (or run the Schematron through SchXslt and Saxon-HE on the command line) and print, per test, whether rule PEPPOL-EN16931-R130 fires
78
  Companions (not included): rules/sch/PEPPOL-EN16931-UBL.sch