Battam commited on
Commit
22d7b14
·
verified ·
1 Parent(s): a33e0c9

Release [run release-20261112T070614Z]

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. CHANGELOG.md +19 -0
  2. README.md +52 -356
  3. README.zh.md +52 -366
  4. campaign_files/2026-10-a-spreadsheets-with-answers-2/rmi-tools-noassertion/README.md +0 -2
  5. campaign_files/2026-10-b-disk-images-3/rust-fs-ntfs-windows-interrupted/README.md +0 -14
  6. campaign_files/2026-10-b-disk-images-3/saurav-astro-ctf-fat12-deleted/README.md +0 -3
  7. campaign_files/2026-10-c-repros-2/sbw-covariate-adjustment-repro/README.md +0 -2
  8. campaign_files/2026-10-g-solver-instances-2/sato7f-puzzle-generation-sat/LICENSE +3 -0
  9. campaign_files/2026-10-g-solver-instances-2/sato7f-puzzle-generation-sat/README.md +28 -0
  10. campaign_files/2026-10-g-solver-instances-2/schedulingbenchmarks-matsp/README.md +0 -4
  11. campaign_files/2026-10-g-solver-instances-2/schedulingbenchmarks-nrp-literature/README.md +0 -4
  12. campaign_files/2026-10-g-solver-instances-2/schedulingbenchmarks-nrp/README.md +0 -10
  13. campaign_files/2026-10-g-solver-instances-2/schedulinglab-fjsp-instances/README.md +0 -4
  14. campaign_files/2026-10-g-solver-instances/sdplib/README.md +0 -4
  15. campaign_files/2026-10-gt-browser-tests-2/ruby-icalendar-fixtures/README.md +0 -4
  16. campaign_files/2026-10-gt-browser-tests-2/ruby-mail-dsn-report-530/README.md +0 -4
  17. campaign_files/2026-10-gt-browser-tests-2/sabre-vobject-itip/README.md +0 -4
  18. campaign_files/2026-10-hw-eda-tests-2/rp-rs-rp2040-pac/README.md +0 -4
  19. campaign_files/2026-10-hw-eda-tests-2/rust-embedded-cortex-m-testsuite-github/README.md +0 -4
  20. campaign_files/2026-10-hw-eda-tests-2/rust-embedded-riscv-tests-qemu-examples-and-ci-expected-golden-files-github/README.md +0 -6
  21. campaign_files/2026-10-hw-eda-tests-2/rust-embedded-svdtools-test-fixture-res-cluster/README.md +0 -4
  22. campaign_files/2026-10-hw-eda-tests-2/secworks-chacha/README.md +0 -4
  23. campaign_files/2026-10-hw-eda-tests-2/secworks-cmac/README.md +0 -4
  24. campaign_files/2026-10-hw-eda-tests-2/secworks-gift/README.md +0 -4
  25. campaign_files/2026-10-hw-eda-tests-2/secworks-md5/README.md +0 -4
  26. campaign_files/2026-10-hw-eda-tests-2/secworks-modexp/README.md +0 -4
  27. campaign_files/2026-10-m-issue-trackers-2/rtcsnv-gtfs/README.md +0 -4
  28. campaign_files/2026-10-m-issue-trackers-2/rtems-users/README.md +0 -4
  29. campaign_files/2026-10-m-issue-trackers-2/ruby-redmine/README.md +0 -6
  30. campaign_files/2026-10-m-issue-trackers-2/ruby-talk/README.md +0 -4
  31. campaign_files/2026-10-m-issue-trackers-2/rust-github-issues/README.md +0 -4
  32. campaign_files/2026-10-m-issue-trackers-2/salisburyreds-passenger-gtfs/README.md +0 -4
  33. campaign_files/2026-10-m-issue-trackers-2/samba-samba/README.md +0 -4
  34. campaign_files/2026-10-m-issue-trackers-2/sanderscoaches-passenger-gtfs/README.md +0 -4
  35. campaign_files/2026-10-m-issue-trackers-2/saskatoon-transit-gtfs/README.md +0 -4
  36. campaign_files/2026-10-m-issue-trackers-2/sbmtd-gtfs/README.md +0 -4
  37. campaign_files/2026-10-md-media-and-documents/ronn-ng-definition-list-syntax/README.md +0 -4
  38. campaign_files/2026-10-md-media-and-documents/rst-to-myst-fixtures-render/README.md +0 -4
  39. campaign_files/2026-10-md-media-and-documents/rst2pdf-rst2pdf-tests-input-and-tests-reference/README.md +0 -4
  40. campaign_files/2026-10-md-media-and-documents/s3-symbolic-symphony-set/README.md +0 -10
  41. campaign_files/2026-10-ml-serving-configs-2/salesforce-merlion/README.md +0 -8
  42. campaign_files/2026-10-op-rule-calculations-2/saleor-warehouse-tests/README.md +0 -6
  43. campaign_files/2026-10-s-security-detection-2/securinets-insat-friendly-ctf-2026-public-challenges/README.md +0 -10
  44. campaign_files/2026-10-sc-science-reference-data-2/rutgers-gsl-monthly-snow-cover/README.md +0 -4
  45. campaign_files/2026-10-sc-science-reference-data-2/scilab-cacsd-module-unit-tests/README.md +0 -16
  46. campaign_files/2026-10-sc-science-reference-data-2/scine-sparrow-python-test/README.md +0 -4
  47. campaign_files/2026-10-sc-science-reference-data/samtools-test-suite/README.md +0 -4
  48. campaign_files/2026-10-sw-tests-and-exercises-2/rsbench/README.md +0 -4
  49. campaign_files/2026-10-sw-tests-and-exercises-2/rust-by-practice/README.md +0 -4
  50. campaign_files/2026-10-sw-tests-and-exercises-2/rust-latam-procedural-macros-workshop/README.md +0 -4
CHANGELOG.md CHANGED
@@ -3346,3 +3346,22 @@ Changes to the content of this index, oldest first. There is one entry for each
3346
  - hf-training-records: refused
3347
  - npm: refused
3348
  - stackexchange-api-posts: refused
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3346
  - hf-training-records: refused
3347
  - npm: refused
3348
  - stackexchange-api-posts: refused
3349
+
3350
+ ## 2026-11-12: sources held back
3351
+
3352
+ - Each source below keeps its newest version that passed every check; its newer batch waits until it passes them:
3353
+ - 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
3354
+ - 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
3355
+ - cfpb-complaints: refused
3356
+ - 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
3357
+ - 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
3358
+ - 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
3359
+ - 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
3360
+ - 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
3361
+ - 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
3362
+ - 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
3363
+ - 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
3364
+ - hf-training-records: refused
3365
+ - npm: refused
3366
+ - scotland-government-workbooks: changed pointer group file data/scotland-government-workbooks/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 2 key(s) of the old version are in no data file of source scotland-government-workbooks after the commit (row 3 from batch 2026-10-a-spreadsheets-with-answers-2, row 6 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
3367
+ - stackexchange-api-posts: refused
README.md CHANGED
@@ -8,6 +8,9 @@ tags:
8
  - manifest
9
  - urls
10
  - provenance
 
 
 
11
  configs:
12
  - config_name: catalog
13
  default: true
@@ -24,25 +27,26 @@ 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-2026-11-12.1/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-2026-11-12.1/docs/figures/numbers-narrow-light.svg">
35
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.1/docs/figures/numbers-wide-light.svg" alt="1,809 material sets with 4,197 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 32,867 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-2026-11-12.1/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-2026-11-12.1/docs/figures/numbers-wide-dark.svg" alt="1,809 material sets with 4,197 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 32,867 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-2026-11-12.1/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 +56,7 @@ 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-2026-11-12.1`.
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 +84,7 @@ 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-2026-11-12.1/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 +95,7 @@ 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-2026-11-12.1/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-12.1/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
@@ -106,377 +110,69 @@ A data file may hold the rows of several sources, so the query keeps this one's
106
 
107
  ## What's inside
108
 
109
- | Task material | Corpora |
110
- |---|---|
111
- | Material sets, each the files admitted from one site, repository or dataset for building tasks, grouped into units that you fetch whole. Counted in material sets and units, by collection. | Whole upstreams read in bulk, such as package registries, bug trackers, mailing lists and web archives, to search when no material set fits. Counted in corpora and rows. |
112
-
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-2026-11-12.1/docs/figures/material-narrow-light.svg">
117
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.1/docs/figures/material-wide-light.svg" alt="Units by collection, longest bar first: Spreadsheets 537; 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 107; 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-2026-11-12.1/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-2026-11-12.1/docs/figures/material-wide-dark.svg" alt="Units by collection, longest bar first: Spreadsheets 537; 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 107; In no collection 9; SEC filing reviews 0.">
122
  </picture>
123
 
124
- 1,809 material sets with 4,197 units, in 14 collections (1 of them with no material set yet); 1 of the material sets are in no collection.
125
-
126
  <details>
127
- <summary>Task material by collection, as a table</summary>
128
-
129
- | Collection | Material sets | Units | Files | Upstreams | Recorded size | Deduplicated size |
130
- |---|---:|---:|---:|---:|---:|---:|
131
- | Spreadsheets | 68 | 537 | 673 | 55 | 250 MB | 249 MB |
132
- | Disk images | 52 | 179 | 324 | 40 | 272 MB | 272 MB |
133
- | Reproduction packages and runs | 88 | 107 | 871 | 88 | 120 MB | 120 MB |
134
- | SEC filing reviews | 0 | 0 | 0 | 0 | 0 B | 0 B |
135
- | Solver instances | 71 | 197 | 381 | 67 | 127 MB | 127 MB |
136
- | Browser tests | 123 | 439 | 852 | 88 | 31.9 MB | 31.9 MB |
137
- | Hardware design (EDA) tests | 81 | 258 | 651 | 75 | 80.8 MB | 80.7 MB |
138
- | Issue trackers | 212 | 258 | 682 | 212 | 196 MB | 196 MB |
139
- | Media and documents | 276 | 391 | 1,088 | 269 | 569 MB | 569 MB |
140
- | Model serving configurations | 103 | 398 | 1,168 | 83 | 250 MB | 229 MB |
141
- | Rule calculations | 177 | 329 | 727 | 154 | 196 MB | 186 MB |
142
- | Security detection | 110 | 328 | 983 | 77 | 684 MB | 652 MB |
143
- | Science reference data | 161 | 342 | 671 | 157 | 63.8 MB | 63.8 MB |
144
- | Tests and exercises | 286 | 425 | 2,544 | 265 | 42.6 MB | 42.2 MB |
145
- | In no collection | 1 | 9 | 19 | 1 | 78.5 KB | 78.5 KB |
146
-
147
- Upstreams are the upstream works the material sets of a collection come from.
148
-
149
- </details>
150
-
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-2026-11-12.1/docs/figures/corpora-narrow-light.svg">
155
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.1/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-2026-11-12.1/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-2026-11-12.1/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.
 
 
 
 
163
 
164
- <details>
165
- <summary>Corpora by main domain, as a table</summary>
166
-
167
- | Domain | Corpora | Rows | Recorded size | Deduplicated size |
168
- |---|---:|---:|---:|---:|
169
- | Data science | 23 | 750,815,212 | 634 TB | 600 TB |
170
- | System administration | 29 | 185,826,217 | 280 TB | 187 TB |
171
- | Many domains | 16 | 109,262,742 | 234 TB | 224 TB |
172
- | Software engineering | 12 | 45,267,701 | 116 TB | 87.3 TB |
173
- | Science | 15 | 44,633,675 | 251 TB | 251 TB |
174
- | Operations | 6 | 39,404,642 | 1.78 TB | 1.78 TB |
175
- | Media | 9 | 12,751,612 | 874 TB | 874 TB |
176
- | Machine learning | 7 | 9,636,889 | 2.04 PB | 1.86 PB |
177
- | Hardware | 5 | 8,716,520 | 1.67 TB | 1.67 TB |
178
- | Unclassified | 4 | 5,284,162 | 1.02 TB | 997 GB |
179
- | Security | 10 | 4,521,631 | 11.9 TB | 11.9 TB |
180
- | Misc | 3 | 241,125 | 2.87 TB | 2.87 TB |
181
-
182
- Each corpus is counted under its main domain.
183
 
184
  </details>
185
 
186
- Browse task material by collection, or corpora, on the [data page](https://huggingface.co/spaces/Battam/training-data-foundation), and build a fetch command there.
187
  Every material set and every corpus, with its counts, size, license tier and login, is in [`docs/sources.md`](docs/sources.md).
188
 
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-2026-11-12.1/docs/figures/steps-narrow-light.svg">
193
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.1/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-2026-11-12.1.">
194
  </picture>
195
  <picture class="not-prose">
196
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.1/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-2026-11-12.1/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-2026-11-12.1.">
198
  </picture>
199
 
200
- The terms, each step and how every number is made are folded below.
201
-
202
- <details>
203
- <summary>Glossary</summary>
204
-
205
- - **Row:** one pointer, a row of the data files under `data/`: where to download one file or record, its size, its hashes and its license.
206
- Corpora are counted in rows.
207
- - **Unit:** a group of files that its upstream publishes as one thing (a test case, one workbook, one package), delivered whole.
208
- Its key `unit_id` is the `unit_group` its rows share in the row column `extra` (`extra.source_meta.unit_group`), unique within its material set.
209
- Material sets are counted in units.
210
- A unit is material, not a task: one unit may give several tasks, and one task may use several units.
211
- - **Material set:** the units admitted from one source (a site, a repository or a dataset) for building tasks, with the README.md and LICENSE the index writes for it, which belong to no unit.
212
- It is counted in units;
213
- its files are the files of its units and those two.
214
- - **Corpus** (plural corpora): one whole upstream that a collector keeps reading in bulk, such as a package registry, a bug tracker, a mailing list or a web archive.
215
- It is counted in rows.
216
- - **Collection:** the material sets of one task family, such as spreadsheets or disk images;
217
- a corpus that holds a collection's kind of material is listed under it and stays a corpus.
218
- - **Form:** the shape of a source's material, one id such as `spreadsheets` or `test-suites` (the catalog column `form`, the option `--form` of `tools/fetch.py`).
219
- - **Default profile:** the rules `tools/fetch.py` applies unless an option relaxes them: it leaves out duplicate copies, rows that matched a benchmark or were not compared with the benchmark list yet, rows in a language other than English and files above 20 MB; rows of any license and rows marked as carrying personal data come unless `--training-only` or `--no-personal-data` leaves them out ([`docs/for-agents.md`](docs/for-agents.md)).
220
-
221
- Every row carries a `unit_id`: in a material set, the key of its unit;
222
- in a corpus, the row's own key (its `uri`, then `@<range_offset>` for a byte range, then `!/<member_path>` for an archive member), and `tools/fetch.py` puts the file of a corpus row in a folder named after it.
223
-
224
- </details>
225
-
226
- <details>
227
- <summary>From corpora to material sets</summary>
228
-
229
- Corpora are whole upstreams read in bulk and counted in rows; a material set is admitted for building tasks from works the search found upstream or from a group of corpus rows, its files checked and grouped into units, and it is counted in units, apart from any corpus its rows came from.
230
-
231
- - **Corpora** come from the collectors: each reads one whole upstream in bulk and records a row for every file or record it lists, with its address, size, hashes and license as the upstream gives them.
232
- - **Admission** makes the material sets.
233
- Its candidates are the works the search found upstream for the task family of a collection, and the units of a group of corpus rows chosen with `tools/fetch.py` (`--where` and `--sample`), which the fetch tool fetched and checked against their recorded size and sha256.
234
- It takes a candidate when its pointer facts hold: a download address, a recorded size and sha256, a recorded license, no legal block, no fact found wrong, and content not already in the index;
235
- every candidate it does not take is listed with its reason in `meta/not_included.parquet` ([`docs/access.md`](docs/access.md)).
236
- - **What admission adds:** the candidate's files as one unit with its `unit_id`, the material set's README.md and LICENSE, its collection (the one of its task family), its form, and labels such as language, robots and terms.
237
- The collect step then fetches every file of every unit again and checks its sha256 before the material set is published.
238
- - A material set made from corpus rows is a source of its own, apart from the corpus: the same file is a row of the corpus and a file of a unit, so the two layers' counts are never added;
239
- the deduplicated size counts its content once.
240
-
241
- </details>
242
-
243
- <details>
244
- <summary>How sizes are measured</summary>
245
-
246
- The two layers are counted apart and never added together;
247
- 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 | 33,745 | 32,867 | 32,740 (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 | 2,151 | 2,144 | 2,130 (99.3%) | 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.
257
- The estimate is for the rows without a recorded size, and says it is one;
258
- above one source its range is the sum of its sources' ranges, wider than a joint interval.
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
-
266
- <details>
267
- <summary>How the training-stage tables are counted</summary>
268
-
269
- The stage table `catalog/stages.yaml` of the code repository gives each training stage a condition on fields the index already has: the kind, form and domain of a row's source, the file types of a unit and, for some stages, rules of their own such as a size limit.
270
- The release records those fields for every unit of a material set and every row of a corpus, with how many of the rules every stage shares it passes, in this order: the benchmark rule (a row that matches a public benchmark task is left out), the rule on rows not compared with the benchmark list yet and deduplication (`meta/stage_fields.parquet`);
271
- each stage counts what meets its condition before those rules and after each of them, and its count after the last one is what it can take.
272
- The training permit and the personal-data mark are labels of every row, neither a condition nor a rule of any stage.
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
-
280
- Material set units for building tasks: every file of the unit in English or no natural language, and at most 20 MB.
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
-
293
- Corpus rows of code, documentation and the terminal side: conditions and counts only.
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
-
306
- Corpus rows of plain text, the mid-training types not counted here: conditions and counts only.
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
-
319
- <details>
320
- <summary>Collection, checks and release</summary>
321
-
322
- - **Data collection:** rows are collected from the public listings, APIs and archives of each source and checked when added:
323
- the address must answer, and size and hash are fetched again and compared.
324
- Material sets come from admission, which takes the works the search found upstream and groups of corpus rows (From corpora to material sets, above).
325
- Every material set and every corpus with its counts: [`docs/sources.md`](docs/sources.md).
326
- - **Checks before release:** every source has its license recorded and grouped into a tier, personal information is read from sample files, and rows are compared with public benchmarks;
327
- rows that copy a public benchmark are removed, and units that do not meet a requirement are listed with the reason: [`docs/access.md`](docs/access.md).
328
- - **Release:** each release is one commit with its own tag;
329
- this card, its figures under `docs/figures/`, `cards/*.md`, `meta/catalog.parquet`, `meta/croissant.json` and `meta/size_figures.json` are generated from the same commit, so their numbers agree.
330
- Every count and size of this card is read from [`meta/size_figures.json`](meta/size_figures.json), where each is computed once for this release, and the stage tables from [`meta/stage_figures.json`](meta/stage_figures.json).
331
- What each file of a release holds: [`docs/fields.md`](docs/fields.md).
332
- - **Chinese version:** every name and description the Chinese card and its pages show is translated by a language model once per text and kept, by the sha256 of its English text, in `meta/translations.parquet`.
333
-
334
- </details>
335
 
336
  ## Reference
337
 
338
- | To | Read |
339
- |---|---|
340
- | Read every field of a row and every column of the catalog | [`docs/fields.md`](docs/fields.md) |
341
- | See every material set and every corpus, with its counts, size, license tier and login | [`docs/sources.md`](docs/sources.md) |
342
- | See what was not included and why, and which sources need a login | [`docs/access.md`](docs/access.md) |
343
- | Build tasks from the files, and keep what an agent must not see out of its environment | [`docs/for-agents.md`](docs/for-agents.md) |
344
- | See every release and what it changed | [CHANGELOG.md](CHANGELOG.md) |
345
- | Read the dataset as Croissant metadata | [Croissant generated by the Hub](https://huggingface.co/api/datasets/Battam/training-data-foundation/croissant); [`meta/croissant.json`](meta/croissant.json) describes the catalog and every source |
346
- | Look through the catalog, one row per material set or corpus with its `kind` and `collection` | [The `catalog` subset in the Dataset Viewer](https://huggingface.co/datasets/Battam/training-data-foundation/viewer/catalog) |
347
- | Query the catalog in the browser; it lists every source as published, without the default profile of `tools/fetch.py`, which the command applies | [SQL Console](https://huggingface.co/datasets/Battam/training-data-foundation?sql_console=true&sql=--+every+source+of+the+catalog+as+published%3A+the+default+profile+of+tools%2Ffetch.py+is+not+applied+here%0ASELECT+source%2C+kind%2C+collection%2C+title%2C+units%2C+files%0AFROM+train%0AORDER+BY+kind%2C+collection%2C+files+DESC%0ALIMIT+50) |
348
- | Watch the index in use, in a video in Chinese and in English, each with its subtitles; the data page plays the video and the subtitles of its own language, in its Get the data section | [Usage video on the data page](https://huggingface.co/spaces/Battam/training-data-foundation#usage-video) |
349
- | Check every number of this card | [`meta/size_figures.json`](meta/size_figures.json) and [`meta/stage_figures.json`](meta/stage_figures.json) |
350
-
351
- <details>
352
- <summary>Dataset structure: the fields used most</summary>
353
-
354
- Each row of the data files under `data/` points to one file or record (a row of the glossary);
355
- the catalog's column `data_files` names the files of each source, and one material set or corpus is read from exactly those files, as in the queries above and with `tools/fetch.py`.
356
- The Dataset Viewer has two subsets: `catalog`, one row per material set or corpus with its `kind` and `collection`, and `preview`, the first 10 rows of each.
357
- The fields used most:
358
-
359
- | Field | Type | Meaning |
360
- |---|---|---|
361
- | `source` | text | Name of the source; the same as its folder under data/. |
362
- | `uri` | text | Where to download the item (https, s3, git, rsync or hf). |
363
- | `size_bytes` | integer | Size in bytes as delivered; empty when the source does not publish it. |
364
- | `sha256` | 32 bytes | SHA-256 of the item, 32 bytes (hex-encode to compare); empty when not known. |
365
- | `ext` | text | File extension from the upstream file name, lower case. |
366
- | `media_type` | text | Media type of the item when known. |
367
- | `license_spdx` | text | The row's own license (SPDX); empty when it takes its batch license. |
368
- | `batch` | text | Batch the row was published in; its license is in meta/sources.json. |
369
- | `sensitive` | text | health or minors when the item is about health or about minors. |
370
- | `extra` | text | JSON with the row's other fields; in a material set, extra.source_meta.unit_group is the key of the row's unit. |
371
-
372
- Every row field and every column of the catalog: [`docs/fields.md`](docs/fields.md).
373
-
374
- </details>
375
 
376
  ## License, privacy and limits
377
 
378
- | Topic | What it means for you |
379
- |---|---|
380
- | License | The catalog is released under CC0 1.0. Each file keeps the license of its source, given in its row. |
381
- | Personal data | The catalog holds no personal records of its own. Rows about health or minors carry `health` or `minors` in the row column `sensitive`, so you can filter them out. |
382
- | Not a copy | The index holds one row per public file or record: its address, size, hashes and license. The files stay at their sources; the index is not a copy of them, not a list of ready tasks or environments, and not a training corpus by itself. |
383
- | Removal | To have a source removed or an entry corrected, open a discussion in the Community tab of this repository. |
384
-
385
- <details>
386
- <summary>License details</summary>
387
-
388
- - This catalog itself (addresses, sizes, hashes and descriptive text) is released under CC0 1.0; that dedication does not extend to the files it points to.
389
- - The files under `media/` (the usage example video and its subtitles) and under `examples/` (the worked examples) are released under CC0 1.0 as well.
390
- - Each file keeps the license of its source: see the license columns of each row (a row without its own license takes the license of its batch, `batches[].license` in `meta/sources.json`).
391
- - License tiers group each source's license by what it allows;
392
- the tier of every source and what each tier means: [`docs/sources.md`](docs/sources.md).
393
- - License tiers and reason codes are computed from the license categories of ScanCode LicenseDB, licensed CC-BY-4.0;
394
- its attribution: [`NOTICE`](NOTICE).
395
-
396
- </details>
397
-
398
- <details>
399
- <summary>Personal and sensitive information</summary>
400
-
401
- - The catalog itself holds addresses, sizes, hashes and license information;
402
- it contains no personal records of its own.
403
- - Some sources point to material that names its authors, such as mailing lists, bug trackers and forums.
404
- - Rows that point to data about health or about minors carry `health` or `minors` in the row column `sensitive`, so they can be filtered out.
405
- - Personal information read from sample files, by level, as each source's card gives it (`risk.pii` and `risk.pii_evidence` in `cards/<source>.yaml`, with the sample files it rests on):
406
-
407
- | Level | Corpora | Task material |
408
- |---|---:|---:|
409
- | `none_known` | 3 | 0 |
410
- | `present` | 14 | 0 |
411
- | `suspected` | 1 | 0 |
412
- | Not judged yet | 121 | 1,809 |
413
- | Sample files read per judged source | 9 to 30 | none |
414
- | Days the sample files were read | 2026-10-04 and 2026-10-05 | none |
415
-
416
- </details>
417
-
418
- <details>
419
- <summary>Known limits</summary>
420
-
421
- | Sources | Corpora | Task material |
422
- |---|---:|---:|
423
- | With a recorded size for every row | 92 of 139 | 1,809 of 1,809 |
424
- | With a summary read from sample files | 0 of 139 | 0 of 1,809 |
425
- | With personal information read from sample files | 18 of 139 | 0 of 1,809 |
426
- | Still being collected | 25 of 139 | 0 of 1,809 |
427
-
428
- - Rows whose size the source does not publish are shown as unknown and left out of the size totals, never counted as 0.
429
- - Fields not judged yet say so;
430
- the counts, types, sizes and licenses of every source come from the index itself.
431
- - A source still being collected grows with each release.
432
- - The Chinese names and descriptions are machine translations.
433
-
434
- </details>
435
-
436
- <details>
437
- <summary>Uses</summary>
438
-
439
- - **Who it is for:** people and agents that choose material for building agent tasks and environments, and then fetch the files themselves.
440
- - **Direct use:** choose sources and files for building agent tasks and environments, then fetch each file from its source and check it against the size and hashes in its row.
441
- - **Out-of-scope use:** this dataset is not a copy of the files, and not a training corpus by itself:
442
- the files stay at their sources under their own licenses.
443
- What a task checks is decided when the task is built from the material;
444
- the index holds material only.
445
- - How to build tasks from the files, what the default profile of `tools/fetch.py` removes and what to keep out of an agent's environment: [`docs/for-agents.md`](docs/for-agents.md).
446
-
447
- </details>
448
-
449
- <details>
450
- <summary>Releases</summary>
451
-
452
- - Latest release: `release-2026-11-12.1` (its newest CHANGELOG entry is dated 2026-11-12).
453
- - Releases are listed in [CHANGELOG.md](CHANGELOG.md);
454
- each one is a commit with its own release tag (this one is [`release-2026-11-12.1`](https://huggingface.co/datasets/Battam/training-data-foundation/tree/release-2026-11-12.1)),
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);
458
- `meta/croissant.json` describes the catalog and every source.
459
-
460
- Days with a release, from the commit history read for this card:
461
-
462
- | Window | Days | Days with a release |
463
- |---|---:|---:|
464
- | 2026-10-14 to 2026-11-12 | 30 | 17 |
465
-
466
- </details>
467
-
468
- <details>
469
- <summary>Removal requests</summary>
470
-
471
- This index lists pointers to files hosted elsewhere (addresses, sizes, checksums and license information); it does not host the files themselves.
472
-
473
- If you maintain a source listed here and would like it removed, or you find an entry that is wrong, please open a discussion in the Community tab of this repository. Name the source, and the files if the request covers only some of them.
474
-
475
- We review new discussions regularly. Removal requests are honoured: the source's entries are taken out in the next update of the index, and the removal is recorded in CHANGELOG.md. Reports of wrong entries are checked and corrected the same way.
476
-
477
- A removal covers the current version of this repository (the main branch) and the Parquet copy that Hugging Face makes automatically for the dataset viewer (the refs/convert/parquet branch). After a removal, we check that the removed entries are no longer in the Parquet copy; when a whole source is removed, we check that its folder is gone from it. Earlier versions of the files stay in the git history of this repository, which pinned commits rely on; if you also need them removed from the history, say so in your request.
478
-
479
- </details>
480
 
481
  ## Citation
482
 
@@ -484,7 +180,7 @@ 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-2026-11-12.1},
488
  year = {2026}
489
  }
490
  ```
 
8
  - manifest
9
  - urls
10
  - provenance
11
+ - training-data
12
+ - pre-training
13
+ - post-training
14
  configs:
15
  - config_name: catalog
16
  default: true
 
27
  <!-- artifact-pointer-layout: v2 -->
28
  <!-- 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. -->
29
 
30
+ # Training Data Foundation
31
 
32
+ Real public data for pre-training, mid-training and post-training, each file with its address, size, hashes and license, fetched with one command;
33
+ this dataset is the foundation's public index.
34
 
35
  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)
36
 
37
  <picture class="not-prose">
38
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/stages-narrow-light.svg">
39
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/stages-wide-light.svg" alt="What each training stage can use in this release. Pre-training (provisional): 30.6 TB deduplicated size, 149,118,607 rows, 29.9 TB allowed for training. Mid-training (provisional): 3.86 TB deduplicated size, 34,079,767 rows, 1.42 TB allowed for training. Post-training: 1,758 material sets, from 1,817 upstream sources, 1,405 allowed for training.">
40
  </picture>
41
  <picture class="not-prose">
42
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/stages-narrow-dark.svg">
43
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/stages-wide-dark.svg" alt="What each training stage can use in this release. Pre-training (provisional): 30.6 TB deduplicated size, 149,118,607 rows, 29.9 TB allowed for training. Mid-training (provisional): 3.86 TB deduplicated size, 34,079,767 rows, 1.42 TB allowed for training. Post-training: 1,758 material sets, from 1,817 upstream sources, 1,405 allowed for training.">
44
  </picture>
45
 
46
  ## Get data in one command
47
 
48
  ```sh
49
+ uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/tools/fetch.py --form spreadsheets --sample 3 --out ./data
50
  ```
51
 
52
  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.
 
56
  <details>
57
  <summary>What the command takes, and in which order</summary>
58
 
59
+ The command and the queries below are pinned to the release tag `release-2026-11-12.2`.
60
  [uv](https://docs.astral.sh/uv/) reads the script's dependencies, at the versions published before the day of this release.
61
 
62
  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`.
 
84
 
85
  ```sql
86
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
87
+ FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-12.2/meta/catalog.parquet'
88
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
89
  ORDER BY collection, units DESC;
90
  ```
 
95
  Choose files in one corpus:
96
 
97
  ```sql
98
+ SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-12.2/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-12.2/meta/catalog.parquet' WHERE source = 'if-archive'));
99
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
100
  FROM read_parquet(getvariable('files'))
101
  WHERE source = 'if-archive' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
 
110
 
111
  ## What's inside
112
 
 
 
 
 
 
 
113
  <picture class="not-prose">
114
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/inside-narrow-light.svg">
115
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/inside-wide-light.svg" alt="The training stages by top-level domain, largest first. Pre-training (provisional), deduplicated size, by domain: Data science 29.8 TB; Science 531 GB; Machine learning 82.8 GB; Media 48.9 GB; System administration 35.7 GB; Hardware 29.6 GB; Many domains 29.6 GB; Security 4.18 GB; Unclassified 1.43 GB; Misc 376 MB; Software engineering 175 MB; Operations 4.08 KB. Mid-training (provisional), deduplicated size, by domain: System administration 1.27 TB; Many domains 1.00 TB; Machine learning 503 GB; Science 461 GB; Data science 429 GB; Hardware 177 GB; Security 13.9 GB; Misc 3.45 GB; Software engineering 2.48 GB; Unclassified 2.21 GB; Media 701 MB; Operations 13.8 MB. Post-training, material sets, by domain: Data science 347; Software engineering 286; Media 273; Operations 206; Machine learning 182; Science 174; Security 103; Hardware 87; System administration 71; Misc 29.">
116
  </picture>
117
  <picture class="not-prose">
118
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/inside-narrow-dark.svg">
119
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/inside-wide-dark.svg" alt="The training stages by top-level domain, largest first. Pre-training (provisional), deduplicated size, by domain: Data science 29.8 TB; Science 531 GB; Machine learning 82.8 GB; Media 48.9 GB; System administration 35.7 GB; Hardware 29.6 GB; Many domains 29.6 GB; Security 4.18 GB; Unclassified 1.43 GB; Misc 376 MB; Software engineering 175 MB; Operations 4.08 KB. Mid-training (provisional), deduplicated size, by domain: System administration 1.27 TB; Many domains 1.00 TB; Machine learning 503 GB; Science 461 GB; Data science 429 GB; Hardware 177 GB; Security 13.9 GB; Misc 3.45 GB; Software engineering 2.48 GB; Unclassified 2.21 GB; Media 701 MB; Operations 13.8 MB. Post-training, material sets, by domain: Data science 347; Software engineering 286; Media 273; Operations 206; Machine learning 182; Science 174; Security 103; Hardware 87; System administration 71; Misc 29.">
120
  </picture>
121
 
 
 
122
  <details>
123
+ <summary>The training stages by top-level domain, as a table</summary>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
124
 
125
+ | Domain | Pre-training: deduplicated size | Pre-training: rows | Mid-training: deduplicated size | Mid-training: rows | Post-training: material sets |
126
+ |---|---:|---:|---:|---:|---:|
127
+ | Software engineering | 175 MB | 58,928 | 2.48 GB | 241,069 | 286 |
128
+ | System administration | 35.7 GB | 1,997,853 | 1.27 TB | 11,372,393 | 71 |
129
+ | Data science | 29.8 TB | 105,523,069 | 429 GB | 264,581 | 347 |
130
+ | Machine learning | 82.8 GB | 237,230 | 503 GB | 219,905 | 182 |
131
+ | Science | 531 GB | 13,352,880 | 461 GB | 197,348 | 174 |
132
+ | Operations | 4.08 KB | 18,862,045 | 13.8 MB | 423 | 206 |
133
+ | Hardware | 29.6 GB | 388,557 | 177 GB | 4,945,550 | 87 |
134
+ | Security | 4.18 GB | 80,662 | 13.9 GB | 117,484 | 103 |
135
+ | Media | 48.9 GB | 16,365 | 701 MB | 34,670 | 273 |
136
+ | Misc | 376 MB | 2,403 | 3.45 GB | 102,989 | 29 |
137
+ | Many domains | 29.6 GB | 8,262,018 | 1.00 TB | 16,181,781 | 0 |
138
+ | Unclassified | 1.43 GB | 336,597 | 2.21 GB | 401,574 | 0 |
139
 
140
+ Each source is counted under its main domain, the top-level domain of its first domain code.
141
+ Every size is deduplicated, and every count is what a stage keeps after the rules every stage shares ([How it is made in the data page's manual](https://battam-training-data-foundation.static.hf.space/index.html?lng=en#manual)).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
142
 
143
  </details>
144
 
145
+ Browse by domain and training stage on the [data page](https://huggingface.co/spaces/Battam/training-data-foundation), and build a fetch command there.
146
  Every material set and every corpus, with its counts, size, license tier and login, is in [`docs/sources.md`](docs/sources.md).
147
 
148
  ## How it is made
149
 
150
  <picture class="not-prose">
151
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/steps-narrow-light.svg">
152
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/steps-wide-light.svg" alt="How the index is made, in four steps. 1, collect: collectors read whole public upstreams into corpora, one row per file or record, the material of pre-training and mid-training. 2, admit: works the search found upstream and groups of corpus rows become material sets, the material of post-training. 3, check: licenses, personal data, overlap with benchmarks and sizes are checked before release. 4, release: one commit with its own tag, release-2026-11-12.2.">
153
  </picture>
154
  <picture class="not-prose">
155
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/steps-narrow-dark.svg">
156
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/steps-wide-dark.svg" alt="How the index is made, in four steps. 1, collect: collectors read whole public upstreams into corpora, one row per file or record, the material of pre-training and mid-training. 2, admit: works the search found upstream and groups of corpus rows become material sets, the material of post-training. 3, check: licenses, personal data, overlap with benchmarks and sizes are checked before release. 4, release: one commit with its own tag, release-2026-11-12.2.">
157
  </picture>
158
 
159
+ Collectors read whole public upstreams into corpora, the material of pre-training and mid-training, and admission turns works found upstream and chosen corpus rows into material sets, the material of post-training.
160
+ Before each release every source is checked for its license, personal data, overlap with benchmarks and sizes, and each release is one tagged commit whose card and figures are computed from its own files.
161
+ The glossary, each step, how sizes and the stage counts are measured, and the checks of this release: [How it is made in the data page's manual](https://battam-training-data-foundation.static.hf.space/index.html?lng=en#manual).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
162
 
163
  ## Reference
164
 
165
+ Every page and file of this dataset, the fields used most, Croissant metadata, the Dataset Viewer and the SQL Console: [Pages and files of the dataset in the data page's manual](https://battam-training-data-foundation.static.hf.space/index.html?lng=en#manual).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
166
 
167
  ## License, privacy and limits
168
 
169
+ Every row carries two labels: its file's training permit, `license_training_ok` (`yes`, `conditional`, `no` or `unknown`), and a personal-data mark when it has one: `health` or `minors` in its column `sensitive`, `personal_data` in its pool fields, or its source's `pii.level` in the catalog.
170
+ They are labels, not rules: `tools/fetch.py` takes every row whatever they say, and filters on them only when asked, `--training-only` leaving out files whose license does not let them be shipped and `--no-personal-data` rows marked as carrying personal data.
171
+ On the data page, the Training filter chooses rows by their permit, and each source's card says whether it holds personal data.
172
+ Rows that copy a public benchmark task, and rows not compared with the benchmark list yet, are left out of every training stage and of the fetch tool's default;
173
+ only `--allow-benchmark` and `--allow-not-compared` take them.
174
+ The catalog itself is released under CC0 1.0, and each file keeps the license of its source.
175
+ The full text, with the license details, personal information read from sample files, known limits, uses, releases and removal requests: [License, privacy and limits in the data page's manual](https://battam-training-data-foundation.static.hf.space/index.html?lng=en#manual).
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
176
 
177
  ## Citation
178
 
 
180
  @misc{training_data_foundation,
181
  title = {training-data-foundation: an index of public materials for building agent environments},
182
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
183
+ note = {Release release-2026-11-12.2},
184
  year = {2026}
185
  }
186
  ```
README.zh.md CHANGED
@@ -8,28 +8,32 @@ tags:
8
  - manifest
9
  - urls
10
  - provenance
 
 
 
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-2026-11-12.1/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-2026-11-12.1/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-2026-11-12.1/docs/figures/numbers-wide-light.zh.svg" alt="1,809 个材料集,共 4,197 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 32,867 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-2026-11-12.1/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-2026-11-12.1/docs/figures/numbers-wide-dark.zh.svg" alt="1,809 个材料集,共 4,197 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 32,867 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-2026-11-12.1/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 +43,7 @@ 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-2026-11-12.1` 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 +71,7 @@ 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-2026-11-12.1/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 +82,7 @@ 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-2026-11-12.1/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-12.1/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
@@ -93,387 +97,69 @@ ORDER BY size_bytes LIMIT 20;
93
 
94
  ## 内容概览
95
 
96
- | 任务材料 | 语料库 |
97
- |---|---|
98
- | 材料集:每个材料集是为构建任务从一个网站、一个仓库或一个数据集收录的文件,分组为单元,每个单元整体取用。按材料集和单元计数,按合集分组。 | 整批读取的完整上游,例如软件包仓库、问题跟踪系统、邮件列表和网页存档;没有合适的材料集时在这里查找。按语料库和行计数。 |
99
-
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-2026-11-12.1/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-2026-11-12.1/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 537;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 107;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-2026-11-12.1/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-2026-11-12.1/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 537;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 107;不属于任何合集 9;SEC 申报审查 0。">
109
  </picture>
110
 
111
- 1,809 个材料集,共 4,197 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。
112
-
113
  <details>
114
- <summary>按合集的任务材料(表格)</summary>
115
-
116
- | 合集 | 材料集数 | 单元数 | 文件数 | 上游数 | 记录的大小 | 去重后大小 |
117
- |---|---:|---:|---:|---:|---:|---:|
118
- | 电子表格 | 68 | 537 | 673 | 55 | 250 MB | 249 MB |
119
- | 磁盘镜像 | 52 | 179 | 324 | 40 | 272 MB | 272 MB |
120
- | 复现包和运行 | 88 | 107 | 871 | 88 | 120 MB | 120 MB |
121
- | SEC 申报审查 | 0 | 0 | 0 | 0 | 0 B | 0 B |
122
- | 求解器实例 | 71 | 197 | 381 | 67 | 127 MB | 127 MB |
123
- | 浏览器测试 | 123 | 439 | 852 | 88 | 31.9 MB | 31.9 MB |
124
- | 硬件设计(EDA)测试 | 81 | 258 | 651 | 75 | 80.8 MB | 80.7 MB |
125
- | 问题跟踪系统 | 212 | 258 | 682 | 212 | 196 MB | 196 MB |
126
- | 媒体和文档 | 276 | 391 | 1,088 | 269 | 569 MB | 569 MB |
127
- | 模型服务配置 | 103 | 398 | 1,168 | 83 | 250 MB | 229 MB |
128
- | 规则计算 | 177 | 329 | 727 | 154 | 196 MB | 186 MB |
129
- | 安全检测 | 110 | 328 | 983 | 77 | 684 MB | 652 MB |
130
- | 科学参考数据 | 161 | 342 | 671 | 157 | 63.8 MB | 63.8 MB |
131
- | 测试和练习 | 286 | 425 | 2,544 | 265 | 42.6 MB | 42.2 MB |
132
- | 不属于任何合集 | 1 | 9 | 19 | 1 | 78.5 KB | 78.5 KB |
133
-
134
- 上游是一个合集的材料集所来自的上游作品。
135
-
136
- </details>
137
-
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-2026-11-12.1/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-2026-11-12.1/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-2026-11-12.1/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-2026-11-12.1/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。
 
 
 
 
150
 
151
- <details>
152
- <summary>按主要领域的语料库(表格)</summary>
153
-
154
- | 领域 | 语料库数 | 行数 | 记录的大小 | 去重后大小 |
155
- |---|---:|---:|---:|---:|
156
- | 数据科学 | 23 | 750,815,212 | 634 TB | 600 TB |
157
- | 系统管理 | 29 | 185,826,217 | 280 TB | 187 TB |
158
- | 多领域 | 16 | 109,262,742 | 234 TB | 224 TB |
159
- | 软件工程 | 12 | 45,267,701 | 116 TB | 87.3 TB |
160
- | 科学 | 15 | 44,633,675 | 251 TB | 251 TB |
161
- | 运营 | 6 | 39,404,642 | 1.78 TB | 1.78 TB |
162
- | 媒体 | 9 | 12,751,612 | 874 TB | 874 TB |
163
- | 机器学习 | 7 | 9,636,889 | 2.04 PB | 1.86 PB |
164
- | 硬件 | 5 | 8,716,520 | 1.67 TB | 1.67 TB |
165
- | 未分类 | 4 | 5,284,162 | 1.02 TB | 997 GB |
166
- | 安全 | 10 | 4,521,631 | 11.9 TB | 11.9 TB |
167
- | 杂项 | 3 | 241,125 | 2.87 TB | 2.87 TB |
168
-
169
- 每个语料库计入它的主要领域。
170
 
171
  </details>
172
 
173
- 在[数据页](https://huggingface.co/spaces/Battam/training-data-foundation)上按合集浏览任务材料,或浏览语料库,并在那里生成下载命令。
174
  每个材料集和每个语料库,以及它的计数、大小、许可证等级和登录要求,都在 [`docs/sources.zh.md`](docs/sources.zh.md) 中。
175
 
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-2026-11-12.1/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-2026-11-12.1/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-2026-11-12.1。">
181
  </picture>
182
  <picture class="not-prose">
183
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.1/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-2026-11-12.1/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-2026-11-12.1。">
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。
188
-
189
- <details>
190
- <summary>术语</summary>
191
-
192
- - **行:** 一个指针,即 `data/` 下数据文件中的一行:在哪里下载一个文件或一条记录,以及它的大小、哈希值和许可证。
193
- 语料库按行计数。
194
- - **单元:** 上游作为一件东西发布的一组文件(一个测试用例、一个工作簿、一个软件包),整体交付。
195
- 它的键 `unit_id` 是它的各行在行列 `extra` 中共用的 `unit_group`(`extra.source_meta.unit_group`),在所属材料集内唯一。
196
- 材料集按单元计数。
197
- 单元是材料,不是题:一个单元可以出几道题,一道题也可以用几个单元。
198
- - **材料集:** 为构建任务从一个来源(一个网站、一个仓库或一个数据集)收录的单元,加上索引为它写的 README.md 和 LICENSE,这两个文件不属于任何单元。
199
- 材料集按单元计数;
200
- 它的文件是它的单元的文件和这两个文件。
201
- - **语料库:** 采集器持续整批读取的一个完整上游,例如软件包仓库、问题跟踪系统、邮件列表或网页存档。
202
- 语料库按行计数。
203
- - **合集:** 同一个任务类别的材料集,例如电子表格或磁盘镜像;
204
- 装着某个合集那一类材料的语料库列在该合集下,仍然是语料库。
205
- - **形式:** 一个来源的材料的形态,一个标识,例如 `spreadsheets` 或 `test-suites`(目录列 `form`,`tools/fetch.py` 的选项 `--form`)。
206
- - **默认配置:** `tools/fetch.py` 在没有选项放宽时应用的规则:去掉重复的副本、与基准测试相符或还没有与基准测试清单比对的行、记录为英语以外语言的行,以及大于 20 MB 的文件;任何许可证的行和标为带有个人数据的行都会进来,除非 `--training-only` 或 `--no-personal-data` 把它们去掉([`docs/for-agents.zh.md`](docs/for-agents.zh.md))。
207
-
208
- 每一行都带有 `unit_id`:在材料集中,是它所属单元的键;
209
- 在语料库中,是该行自己的键(它的 `uri`,字节范围再接 `@<range_offset>`,存档成员再接 `!/<member_path>`),`tools/fetch.py` 把语料库一行的文件放进以它命名的文件夹。
210
-
211
- </details>
212
-
213
- <details>
214
- <summary>从语料库到材料集</summary>
215
-
216
- 语料库是整批读取的完整上游,按行计数;材料集为构建任务而收录,来自检索在上游找到的作品或一组语料库行,它的文件经过核对并分组为单元,按单元计数,与它的行所来自的语料库分开。
217
-
218
- - **语料库**来自采集器:每个采集器整批读取一个完整上游,为它列出的每个文件或每条记录记一行,带有上游给出的地址、大小、哈希值和许可证。
219
- - **收录**产生材料集。
220
- 它的候选是检索为某个合集的任务类别在上游找到的作品,以及用 `tools/fetch.py`(`--where` 和 `--sample`)选出的一组语料库行中的单元,这些单元由下载工具取回,并按记录的大小和 sha256 校验过。
221
- 候选的指针事实成立时才收录:有下载地址,记录了大小和 sha256,记录了许可证,没有法律封锁,没有发现事实有误,内容也不在索引中;
222
- 未收录的每个候选都连同原因列在 `meta/not_included.parquet` 中([`docs/access.zh.md`](docs/access.zh.md))。
223
- - **收录添加的内容:** 把候选的文件作为一个单元并给出它的 `unit_id`,材料集的 README.md 和 LICENSE,它的合集(它的任务类别所属的合集),它的形式,以及语言(`language`)、`robots` 和使用条款(`terms`)等标签。
224
- 之后采集步骤再次取回每个单元的每个文件,并在材料集发布之前校验它的 sha256。
225
- - 由语料库行做成的材料集是一个独立的来源,与那个语料库分开:同一个文件既是语料库的一行,又是一个单元的一个文件,所以两层的计数从不相加;
226
- 去重后大小只把它的内容算一次。
227
-
228
- </details>
229
-
230
- <details>
231
- <summary>大小怎样测得</summary>
232
-
233
- 两层分开计数,从不相加;
234
- 只有两层的大小相加,得到整个索引的大小。
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 | 33,745 | 32,867 | 32,740(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 | 2,151 | 2,144 | 2,130(99.3%) | 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。
244
- 估计值针对没有记录大小的行,并注明是估计值;
245
- 在来源之上的层级,它的范围是各来源范围之和,比联合区间更宽。
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��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
-
253
- <details>
254
- <summary>训练阶段表怎样计数</summary>
255
-
256
- 代码仓库的阶段表 `catalog/stages.yaml` 给每个训练阶段一个条件,条件只用索引已有的字段:一行所属来源的种类、形式和领域,一个单元的文件类型,以及有些阶段自己的规则,例如大小上限。
257
- 发布为材料集的每个单元和语料库的每一行记录这些字段,并记录它按以下顺序通过了各阶段共用规则中的几条:基准测试规则(与公开基准测试任务相符的行被去掉)、还没有与基准测试清单比对的行的规则和去重(`meta/stage_fields.parquet`);
258
- 每个阶段在这些规则之前和每条规则之后统计符合它的条件的材料,最后一条规则之后的计数就是它能用的材料。
259
- 训练许可和个人数据标记是每一行的标签,既不是任何阶段的条件,也不是规则。
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多少允许训练(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
-
267
- 用于构建任务的材料集单元:该单元的每个文件为英语或不含自然语言,且最多 20 MB。
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
-
280
- 代码、文档和终端侧的语料库行:仅列条件和数量。
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
-
293
- 纯文本的语料库行,不在此统计中期训练类型:仅列条件和数量。
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
-
306
- <details>
307
- <summary>采集、检查和发布</summary>
308
-
309
- - **数据采集:** 行从每个来源的公开列表、API 和存档中采���,加入时即校验:
310
- 地址必须能访问,大小和哈希值会重新取回并比对。
311
- 材料集来自收录,它接收检索在上游找到的作品和成组的语料库行(见上面的“从语料库到材料集”)。
312
- 每个材料集和每个语料库及其计数:[`docs/sources.zh.md`](docs/sources.zh.md)。
313
- - **发布前的检查:** 每个来源都记录了许可证并归入一个等级,个人信息从样本文件中判读,各行与公开基准测试比对;
314
- 复制了公开基准测试的行会被移除,不满足某项要求的单元会连同原因一起列出:[`docs/access.zh.md`](docs/access.zh.md)。
315
- - **发布:** 每次发布是一个带有自己标签的提交;
316
- 本卡片、`docs/figures/` 下它的图、`cards/*.md`、`meta/catalog.parquet`、`meta/croissant.json` 和 `meta/size_figures.json` 由同一个提交生成,因此数字彼此一致。
317
- 本卡片的每个计数和大小都读自 [`meta/size_figures.json`](meta/size_figures.json),每个数在那里为本次发布只算一次;
318
- 阶段表读自 [`meta/stage_figures.json`](meta/stage_figures.json)。
319
- 每次发布中各个文件的内容:[`docs/fields.zh.md`](docs/fields.zh.md)。
320
- - **中文版本:** 中文卡片及其页面显示的每个名称和描述,都由语言模型对每段文字翻译一次,按英文原文的 sha256 保存在 `meta/translations.parquet` 中。
321
-
322
- </details>
323
 
324
  ## 查阅
325
 
326
- | 想做的事 | 去读 |
327
- |---|---|
328
- | 查看每个行字段和目录的每一列 | [`docs/fields.zh.md`](docs/fields.zh.md) |
329
- | 查看每个材料集和每个语料库,以及它的计数、大小、许可证等级和登录要求 | [`docs/sources.zh.md`](docs/sources.zh.md) |
330
- | 查看哪些没有收录、为什么,以及哪些来源需要登录 | [`docs/access.zh.md`](docs/access.zh.md) |
331
- | 用这些文件构建任务,并把智能体不该看到的内容挡在它的环境之外 | [`docs/for-agents.zh.md`](docs/for-agents.zh.md) |
332
- | 查看每次发布及其改动 | [CHANGELOG.md](CHANGELOG.md) |
333
- | 以 Croissant 元数据读取本数据集 | [由 Hub 生成的 Croissant](https://huggingface.co/api/datasets/Battam/training-data-foundation/croissant);[`meta/croissant.json`](meta/croissant.json) 描述目录和每个来源 |
334
- | 浏览目录,每个材料集或语料库一行,带有它的 `kind` 和 `collection` | [数据集查看器中的 `catalog` 子集](https://huggingface.co/datasets/Battam/training-data-foundation/viewer/catalog) |
335
- | 在浏览器里查询目录;它按发布原样列出每个来源,不应用 `tools/fetch.py` 的默认配置,命令才应用它 | [SQL 控制台](https://huggingface.co/datasets/Battam/training-data-foundation?sql_console=true&sql=--+every+source+of+the+catalog+as+published%3A+the+default+profile+of+tools%2Ffetch.py+is+not+applied+here%0ASELECT+source%2C+kind%2C+collection%2C+title%2C+units%2C+files%0AFROM+train%0AORDER+BY+kind%2C+collection%2C+files+DESC%0ALIMIT+50) |
336
- | 观看本索引的使用示例视频,有中文版和英文版,各带字幕;数据页在“获取数据”部分播放与页面语言相同的视频和字幕 | [数据页上的使用示例视频](https://huggingface.co/spaces/Battam/training-data-foundation#usage-video) |
337
- | 核对本卡片的每个数字 | [`meta/size_figures.json`](meta/size_figures.json) 和 [`meta/stage_figures.json`](meta/stage_figures.json) |
338
-
339
- <details>
340
- <summary>数据集结构:最常用的字段</summary>
341
-
342
- `data/` 下数据文件的每一行指向一个文件或一条记录(见术语中的“行”);
343
- 目录的 `data_files` 列列出每个来源的文件,单个材料集或语料库就从这些文件读取,见上面的查询和 `tools/fetch.py`。
344
- 数据集查看器有两个子集:`catalog`,每个材料集或语料库一行,带有它的 `kind` 和 `collection`;`preview`,每个的前 10 行。
345
- 最常用的字段:
346
-
347
- | 字段 | 类型 | 含义 |
348
- |---|---|---|
349
- | `source` | 文本 | 来源名称;与 data/ 下的文件夹名相同。 |
350
- | `uri` | 文本 | 下载该条目的地址(https、s3、git、rsync 或 hf)。 |
351
- | `size_bytes` | 整数 | 交付时的字节数;来源没有公布时为空。 |
352
- | `sha256` | 32 字节 | 该条目的 SHA-256,32 字节(转成十六进制后比较);未知时为空。 |
353
- | `ext` | 文本 | 取自上游文件名的扩展名,小写。 |
354
- | `media_type` | 文本 | 已知时为该条目的媒体类型。 |
355
- | `license_spdx` | 文本 | 该行自己的许可证(SPDX);沿用所在批次的许可证时为空。 |
356
- | `batch` | 文本 | 该行发布时所在的批次;批次的许可证见 meta/sources.json。 |
357
- | `sensitive` | 文本 | 条目涉及健康或未成年人时为 health 或 minors。 |
358
- | `extra` | 文本 | 该行其余字段的 JSON;材料集的行中,extra.source_meta.unit_group 是该行所属单元的键。 |
359
-
360
- 全部行字段和目录的全部列:[`docs/fields.zh.md`](docs/fields.zh.md)。
361
-
362
- </details>
363
 
364
  ## 许可、隐私与局限
365
 
366
- | 事项 | 对你意味着什么 |
367
- |---|---|
368
- | 许可 | 目录本身以 CC0 1.0 发布。每个文件保留其来源的许可证,写在它所在的行里。 |
369
- | 个人数据 | 目录本身不含任何个人记录。涉及健康或未成年人的行在行列 `sensitive` 中标有 `health` 或 `minors`,可以据此过滤掉。 |
370
- | 不是副本 | 索引为每个公开文件或记录记一行:它的地址、大小、哈希值和许可证。文件留在各自的来源;索引不是它们的副本,不是现成的任务或环境列表,本身也不是训练语料。 |
371
- | 移除 | 如需移除某个来源或更正某个条目,请在本仓库的社区标签页发起讨论。 |
372
-
373
- <details>
374
- <summary>许可详情</summary>
375
-
376
- - 本目录本身(地址、大小、哈希值和描述文字)以 CC0 1.0 发布;
377
- 这一声明不延及它所指向的文件。
378
- - `media/` 下的文件(使用示例视频及其字幕)和 `examples/` 下的文件(完整示例)同样以 CC0 1.0 发布。
379
- - 每个文件保留其来源的许可证:见每一行的许可证列(没有自己许可证的行沿用所在批次的许可证,即 `meta/sources.json` 中的 `batches[].license`)。
380
- - 许可证等级按许可内容对每个来源的许可证分组;
381
- 每个来源的等级以及每个等级的含义:[`docs/sources.zh.md`](docs/sources.zh.md)。
382
- - 许可证等级和原因码由 ScanCode LicenseDB 的许可证类别算出,它以 CC-BY-4.0 许可;
383
- 它的署名:[`NOTICE`](NOTICE)。
384
-
385
- </details>
386
-
387
- <details>
388
- <summary>个人与敏感信息</summary>
389
-
390
- - 目录本身只包含地址、大小、哈希值和许可证信息;
391
- 它本身不含任何个人记录。
392
- - 有些来源指向署有作者姓名的材料,例如邮件列表、问题跟踪系统和论坛。
393
- - 指向健康或未成年人相关数据的行,在行列 `sensitive` 中标有 `health` 或 `minors`,可以据此过滤掉。
394
- - 从样本文件读出的个人信息,按等级,与每个来源的卡片给出的相同(`cards/<source>.yaml` 里的 `risk.pii` 和 `risk.pii_evidence`,附所依据的样本文件):
395
-
396
- | 等级 | 语料库 | 任务材料 |
397
- |---|---:|---:|
398
- | `none_known` | 3 | 0 |
399
- | `present` | 14 | 0 |
400
- | `suspected` | 1 | 0 |
401
- | 尚未判读 | 121 | 1,809 |
402
- | 每个已判读的来源读过的样本文件数 | 9 到 30 | 无 |
403
- | 读取样本文件的日期 | 2026-10-04 和 2026-10-05 | 无 |
404
-
405
- </details>
406
-
407
- <details>
408
- <summary>已知局限</summary>
409
-
410
- | 来源 | 语料库 | 任务材料 |
411
- |---|---:|---:|
412
- | 每一行都有记录的大小 | 139 个中的 92 个 | 1,809 个中的 1,809 个 |
413
- | 已从样本文件读出摘要 | 139 个中的 0 个 | 1,809 个中的 0 个 |
414
- | 已从样本文件读出个人信息 | 139 个中的 18 个 | 1,809 个中的 0 个 |
415
- | 仍在采集中 | 139 个中的 25 个 | 1,809 个中的 0 个 |
416
-
417
- - 来源没有公布大小的行显示为未知,不计入大小合计,也从不按 0 计。
418
- - 尚未判读的字段会注明;
419
- 每个来源的计数、类型、大小和许可证都直接来自索引本身。
420
- - 仍在采集中的来源随每次发布增长。
421
- - 中文名称和描述是机器翻译。
422
-
423
- </details>
424
-
425
- <details>
426
- <summary>用途</summary>
427
-
428
- - **给谁用:** 为构建智能体任务和环境挑选材料、然后自己下载文件的人和智能体。
429
- - **直接用途:** 为构建智能体任务和环境挑选来源和文件,再从各自的来源下载每个文件,并按它所在行记录的大小和哈希值校验。
430
- - **不适用的用途:** 本数据集不是文件的副本,本身也不是训练语料:
431
- 文件留在各自的来源,遵循各自的许可证。
432
- 一道题检查什么,在用材料构建这道题时才决定;
433
- 索引只存材料。
434
- - 如何用这些文件构建任务、`tools/fetch.py` 的默认配置会去掉什么、哪些内容不能放进智能体的环境:[`docs/for-agents.zh.md`](docs/for-agents.zh.md)。
435
-
436
- </details>
437
-
438
- <details>
439
- <summary>发布</summary>
440
-
441
- - 最新发布:`release-2026-11-12.1`(其最新的 CHANGELOG 条目日期为 2026-11-12)。
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-2026-11-12.1`](https://huggingface.co/datasets/Battam/training-data-foundation/tree/release-2026-11-12.1)),
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);
447
- `meta/croissant.json` 描述目录和每个来源。
448
-
449
- 有发布的天数,依据为本卡片读取的提交历史:
450
-
451
- | 时间窗口 | 天数 | 有发布的天数 |
452
- |---|---:|---:|
453
- | 2026-10-14 到 2026-11-12 | 30 | 17 |
454
-
455
- </details>
456
-
457
- <details>
458
- <summary>移除请求</summary>
459
-
460
- 本索引列出的是托管在别处的文件的指针(地址、大小、校验和与许可证信息),并不托管文件本身。
461
-
462
- 如果你维护这里列出的某个来源并希望将其移除,或者发现某个条目有误,请在本仓库的社区标签页发起讨论。
463
- 请写明来源名称;
464
- 如果请求只涉及部分文件,也请写明是哪些文件。
465
-
466
- 我们会定期查看新的讨论。
467
- 移除请求会得到处理:该来源的条目会在下一次索引更新中移除,并记录在 CHANGELOG.md 中。
468
- 对错误条目的报告也会以同样的方式核查和更正。
469
-
470
- 移除覆盖本仓库的当前版本(main 分支)以及 Hugging Face 为数据集查看器自动生成的 Parquet 副本(refs/convert/parquet 分支)。
471
- 移除之后,我们会核查被移除的条目已不在 Parquet 副本中;
472
- 整个来源被移除时,我们会核查它的文件夹已从中消失。
473
- 文件的早期版本仍保留在本仓库的 git 历史中,被固定引用的提交依赖它们;
474
- 如果你也需要从历史中移除,请在请求中说明。
475
-
476
- </details>
477
 
478
  ## 引用
479
 
@@ -481,7 +167,7 @@ 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-2026-11-12.1},
485
  year = {2026}
486
  }
487
  ```
 
8
  - manifest
9
  - urls
10
  - provenance
11
+ - training-data
12
+ - pre-training
13
+ - post-training
14
  ---
15
  <!-- 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. -->
16
 
17
+ # 数据基座
18
 
19
+ 用于预训练、中训练和后训练的真实公开数据,每个文件都带有地址、大小、哈希值和许可证,一条命令即可取回;
20
+ 本数据集是数据基座的公开索引。
21
 
22
  [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)
23
 
24
  <picture class="not-prose">
25
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/stages-narrow-light.zh.svg">
26
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/stages-wide-light.zh.svg" alt="本次发布中每个训练阶段能用的数据。预训练(暂定):去重后大小 30.6 TB,149,118,607 行,其中 29.9 TB 允许用于训练。中训练(暂定):去重后大小 3.86 TB,34,079,767 行,其中 1.42 TB 允许用于训练。后训练:材料集 1,758,来自 1,817 个上游来源,其中 1,405 允许用于训练。">
27
  </picture>
28
  <picture class="not-prose">
29
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/stages-narrow-dark.zh.svg">
30
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/stages-wide-dark.zh.svg" alt="本次发布中每个训练阶段能用的数据。预训练(暂定):去重后大小 30.6 TB,149,118,607 行,其中 29.9 TB 允许用于训练。中训练(暂定):去重后大小 3.86 TB,34,079,767 行,其中 1.42 TB 允许用于训练。后训练:材料集 1,758,来自 1,817 个上游来源,其中 1,405 允许用于训练。">
31
  </picture>
32
 
33
  ## 一条命令取数据
34
 
35
  ```sh
36
+ uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/tools/fetch.py --form spreadsheets --sample 3 --out ./data
37
  ```
38
 
39
  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。
 
43
  <details>
44
  <summary>命令取什么、按什么顺序取</summary>
45
 
46
+ 这条命令和下面的查询都固定在发布标签 `release-2026-11-12.2` 上。
47
  [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。
48
 
49
  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。
 
71
 
72
  ```sql
73
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
74
+ FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-12.2/meta/catalog.parquet'
75
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
76
  ORDER BY collection, units DESC;
77
  ```
 
82
  在一个语料库里挑选文件:
83
 
84
  ```sql
85
+ SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-12.2/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-12.2/meta/catalog.parquet' WHERE source = 'if-archive'));
86
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
87
  FROM read_parquet(getvariable('files'))
88
  WHERE source = 'if-archive' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
 
97
 
98
  ## 内容概览
99
 
 
 
 
 
 
 
100
  <picture class="not-prose">
101
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/inside-narrow-light.zh.svg">
102
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/inside-wide-light.zh.svg" alt="按顶层领域分的各训练阶段,由大到小。预训练(暂定),去重后大小,按领域:数据科学 29.8 TB;科学 531 GB;机器学习 82.8 GB;媒体 48.9 GB;系统管理 35.7 GB;硬件 29.6 GB;多领域 29.6 GB;安全 4.18 GB;未分类 1.43 GB;杂项 376 MB;软件工程 175 MB;运营 4.08 KB。中训练(暂定),去重后大小,按领域:系统管理 1.27 TB;多领域 1.00 TB;机器学习 503 GB;科学 461 GB;数据科学 429 GB;硬件 177 GB;安全 13.9 GB;杂项 3.45 GB;软件工程 2.48 GB;未分类 2.21 GB;媒体 701 MB;运营 13.8 MB。后训练,材料集,按领域:数据科学 347;软件工程 286;媒体 273;运营 206;机器学习 182;科学 174;安全 103;硬件 87;系统管理 71;杂项 29。">
103
  </picture>
104
  <picture class="not-prose">
105
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/inside-narrow-dark.zh.svg">
106
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/inside-wide-dark.zh.svg" alt="按顶层领域分的各训练阶段,由大到小。预训练(暂定),去重后大小,按领域:数据��学 29.8 TB;科学 531 GB;机器学习 82.8 GB;媒体 48.9 GB;系统管理 35.7 GB;硬件 29.6 GB;多领域 29.6 GB;安全 4.18 GB;未分类 1.43 GB;杂项 376 MB;软件工程 175 MB;运营 4.08 KB。中训练(暂定),去重后大小,按领域:系统管理 1.27 TB;多领域 1.00 TB;机器学习 503 GB;科学 461 GB;数据科学 429 GB;硬件 177 GB;安全 13.9 GB;杂项 3.45 GB;软件工程 2.48 GB;未分类 2.21 GB;媒体 701 MB;运营 13.8 MB。后训练,材料集,按领域:数据科学 347;软件工程 286;媒体 273;运营 206;机器学习 182;科学 174;安全 103;硬件 87;系统管理 71;杂项 29。">
107
  </picture>
108
 
 
 
109
  <details>
110
+ <summary>按顶层领域分的各训练阶段(表格)</summary>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
111
 
112
+ | 领域 | 预训练:去重后大小 | 预训练:行 | 中训练:去重后大小 | 中训练:行 | 后训练:材料集 |
113
+ |---|---:|---:|---:|---:|---:|
114
+ | 软件工程 | 175 MB | 58,928 | 2.48 GB | 241,069 | 286 |
115
+ | 系统管理 | 35.7 GB | 1,997,853 | 1.27 TB | 11,372,393 | 71 |
116
+ | 数据科学 | 29.8 TB | 105,523,069 | 429 GB | 264,581 | 347 |
117
+ | 机器学习 | 82.8 GB | 237,230 | 503 GB | 219,905 | 182 |
118
+ | 科学 | 531 GB | 13,352,880 | 461 GB | 197,348 | 174 |
119
+ | 运营 | 4.08 KB | 18,862,045 | 13.8 MB | 423 | 206 |
120
+ | 硬件 | 29.6 GB | 388,557 | 177 GB | 4,945,550 | 87 |
121
+ | 安全 | 4.18 GB | 80,662 | 13.9 GB | 117,484 | 103 |
122
+ | 媒体 | 48.9 GB | 16,365 | 701 MB | 34,670 | 273 |
123
+ | 杂项 | 376 MB | 2,403 | 3.45 GB | 102,989 | 29 |
124
+ | 多领域 | 29.6 GB | 8,262,018 | 1.00 TB | 16,181,781 | 0 |
125
+ | 未分类 | 1.43 GB | 336,597 | 2.21 GB | 401,574 | 0 |
126
 
127
+ 每个来源计入它的主要领域,即它的第一个领域代码所属的顶层领域。
128
+ 每个大小都是去重后的大小,每个计数都是一个阶段在各阶段共用的规则之后保留下来的部分([数据页使用手册的“怎样做成”](https://battam-training-data-foundation.static.hf.space/index.html?lng=zh#manual))。
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
129
 
130
  </details>
131
 
132
+ 在[数据页](https://huggingface.co/spaces/Battam/training-data-foundation)上按领域和训练阶段浏览,并在那里生成下载命令。
133
  每个材料集和每个语料库,以及它的计数、大小、许可证等级和登录要求,都在 [`docs/sources.zh.md`](docs/sources.zh.md) 中。
134
 
135
  ## 怎样做成
136
 
137
  <picture class="not-prose">
138
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/steps-narrow-light.zh.svg">
139
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/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的公开上游,每个文件或每条记录记一行,成为语料库,供预训练和中训练使用。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,供后训练使用。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-2026-11-12.2。">
140
  </picture>
141
  <picture class="not-prose">
142
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/docs/figures/steps-narrow-dark.zh.svg">
143
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-12.2/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的公开上游,每个文件或每条记录记一行,成为语料库,供预训练和中训练使用。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,供后训练使用。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-2026-11-12.2。">
144
  </picture>
145
 
146
+ 采集器整批读取完整的公开上游,记成语料库,供预训练和中训练使用,收录则把检索在上游找到的作品和选出的语料库行做成材料集,供后训练使用。
147
+ 每次发布前,每个来源都要核对许可证、个人数据、与基准测试的重叠和大小,每次发布是一个带标签的提交,它的卡片和图都由它自己的文件算出。
148
+ 术语、每一步、大小和阶段计数怎样测得,以及本次发布的核对结果:[数据页使用手册的“怎样做成”](https://battam-training-data-foundation.static.hf.space/index.html?lng=zh#manual)。
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
149
 
150
  ## 查阅
151
 
152
+ 本数据集的每个页面和文件、最常用的字段、Croissant 元数据、数据集查看器和 SQL 控制台:[数据页使用手册的“数据集的页面和文件”](https://battam-training-data-foundation.static.hf.space/index.html?lng=zh#manual)。
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
153
 
154
  ## 许可、隐私与局限
155
 
156
+ 每一行都带有两个标签:它的文件的训练许可 `license_training_ok`(`yes`、`conditional`、`no` 或 `unknown`),以及有个人数据时的个人数据标记:它的 `sensitive` 列中的 `health` 或 `minors`、它的池字段中的 `personal_data`,或目录中它的来源的 `pii.level`。
157
+ 它们是标签,不是规则:不论它们是什么,`tools/fetch.py` 都取每一行,只在要求时按它们筛选:`--training-only` 去掉许可证不允许随附分发的文件,`--no-personal-data` 去掉标为带有个人数据的行。
158
+ 在数据页上,“训练”筛选按训练许可挑选行,每个来源的卡片写明它有没有个人数据。
159
+ 复制了公开基准测试任务的行,以及还没有与基准测试清单比对的行,不计入任何训练阶段,下载工具默认也不取;
160
+ 只有 `--allow-benchmark` 和 `--allow-not-compared` 会取它们。
161
+ 目录本身以 CC0 1.0 发布,每个文件保留其来源的许可证。
162
+ 完整说明,包括许可详情、从样本文件读出的个人信息、已知局限、用途、发布和移除请求:[数据页使用手册的“许可、隐私与局限”](https://battam-training-data-foundation.static.hf.space/index.html?lng=zh#manual)。
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
163
 
164
  ## 引用
165
 
 
167
  @misc{training_data_foundation,
168
  title = {training-data-foundation: an index of public materials for building agent environments},
169
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
170
+ note = {Release release-2026-11-12.2},
171
  year = {2026}
172
  }
173
  ```
campaign_files/2026-10-a-spreadsheets-with-answers-2/rmi-tools-noassertion/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):** NOASSERTION
 
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):** NOASSERTION
campaign_files/2026-10-b-disk-images-3/rust-fs-ntfs-windows-interrupted/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: corruption_desc.
12
-
13
  **Tools to open and check:** Open with tools for these file types: .gz.
14
 
15
  **Licence (SPDX):** MIT OR Apache-2.0
@@ -20,49 +18,37 @@ Files:
20
 
21
  - `test-disks/windows-interrupted-1.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (1), rust-fs-ntfs test disks.
22
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
23
- Reference result: corruption_desc
24
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
25
  - `test-disks/windows-interrupted-6.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (6), rust-fs-ntfs test disks.
26
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
27
- Reference result: corruption_desc
28
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
29
  - `test-disks/windows-interrupted-wrap-span.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (wrap-span), rust-fs-ntfs test disks.
30
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
31
- Reference result: corruption_desc
32
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
33
  - `test-disks/windows-interrupted-wrap-boundary.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (wrap-boundary), rust-fs-ntfs test disks.
34
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
35
- Reference result: corruption_desc
36
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
37
  - `test-disks/windows-interrupted-wrap-mft-grows.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (wrap-mft-grows), rust-fs-ntfs test disks.
38
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
39
- Reference result: corruption_desc
40
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
41
  - `test-disks/windows-interrupted-index-vcn.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (index-vcn), rust-fs-ntfs test disks.
42
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
43
- Reference result: corruption_desc
44
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
45
  - `test-disks/windows-interrupted-small-cluster.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (small-cluster), rust-fs-ntfs test disks.
46
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
47
- Reference result: corruption_desc
48
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
49
  - `test-disks/windows-interrupted-tail-continuation.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (tail-continuation), rust-fs-ntfs test disks.
50
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
51
- Reference result: corruption_desc
52
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
53
  - `test-disks/windows-interrupted-undo.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (undo), rust-fs-ntfs test disks.
54
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
55
- Reference result: corruption_desc
56
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
57
  - `test-disks/windows-interrupted-open-at-checkpoint.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (open-at-checkpoint), rust-fs-ntfs test disks.
58
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
59
- Reference result: corruption_desc
60
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
61
  - `test-disks/windows-interrupted-quota.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (quota), rust-fs-ntfs test disks.
62
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
63
- Reference result: corruption_desc
64
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
65
  - `test-disks/windows-interrupted-fresh-log.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (fresh-log), rust-fs-ntfs test disks.
66
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
67
- Reference result: corruption_desc
68
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
 
8
 
9
  **Self-contained:** Whether each 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: .gz.
12
 
13
  **Licence (SPDX):** MIT OR Apache-2.0
 
18
 
19
  - `test-disks/windows-interrupted-1.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (1), rust-fs-ntfs test disks.
20
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
21
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
22
  - `test-disks/windows-interrupted-6.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (6), rust-fs-ntfs test disks.
23
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
24
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
25
  - `test-disks/windows-interrupted-wrap-span.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (wrap-span), rust-fs-ntfs test disks.
26
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
27
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
28
  - `test-disks/windows-interrupted-wrap-boundary.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (wrap-boundary), rust-fs-ntfs test disks.
29
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
30
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
31
  - `test-disks/windows-interrupted-wrap-mft-grows.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (wrap-mft-grows), rust-fs-ntfs test disks.
32
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
33
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
34
  - `test-disks/windows-interrupted-index-vcn.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (index-vcn), rust-fs-ntfs test disks.
35
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
36
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
37
  - `test-disks/windows-interrupted-small-cluster.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (small-cluster), rust-fs-ntfs test disks.
38
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
39
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
40
  - `test-disks/windows-interrupted-tail-continuation.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (tail-continuation), rust-fs-ntfs test disks.
41
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
42
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
43
  - `test-disks/windows-interrupted-undo.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (undo), rust-fs-ntfs test disks.
44
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
45
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
46
  - `test-disks/windows-interrupted-open-at-checkpoint.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (open-at-checkpoint), rust-fs-ntfs test disks.
47
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
48
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
49
  - `test-disks/windows-interrupted-quota.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (quota), rust-fs-ntfs test disks.
50
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
51
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
52
  - `test-disks/windows-interrupted-fresh-log.img.gz`: NTFS volume captured mid-write by Windows, $LogFile not replayed (fresh-log), rust-fs-ntfs test disks.
53
  Origin: Upstream test fixture captured from the real Windows NTFS driver mid-write (VSS snapshot of the hosting volume); upstream states nothing was synthesised and no byte was written by the crate.
 
54
  Dependencies: gzip or Python gzip to unpack; only needed to rerun the exercise, not to grade the answer
campaign_files/2026-10-b-disk-images-3/saurav-astro-ctf-fat12-deleted/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: .img.
14
 
15
  **Licence (SPDX):** MIT
@@ -20,5 +18,4 @@ Files:
20
 
21
  - `forensics/medium/2_Disk_FAT12_Deleted/player/floppy.img`: FAT12 floppy image with one deleted, XOR-masked FLAG.TXT (Disk FAT12 Deleted challenge).
22
  Origin: CTF image made with mkfs.fat by the challenge author for a FAT12 deleted-entry exercise.
23
- Reference result: deleted_files
24
  Dependencies: 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: .img.
12
 
13
  **Licence (SPDX):** MIT
 
18
 
19
  - `forensics/medium/2_Disk_FAT12_Deleted/player/floppy.img`: FAT12 floppy image with one deleted, XOR-masked FLAG.TXT (Disk FAT12 Deleted challenge).
20
  Origin: CTF image made with mkfs.fat by the challenge author for a FAT12 deleted-entry exercise.
 
21
  Dependencies: only needed to rerun the exercise, not to grade the answer
campaign_files/2026-10-c-repros-2/sbw-covariate-adjustment-repro/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether the unit needs other upstream files is not stated. 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: .r.
14
 
15
  **Licence (SPDX):** MIT
 
8
 
9
  **Self-contained:** Whether the unit needs other upstream files is not stated. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Open with tools for these file types: .r.
12
 
13
  **Licence (SPDX):** MIT
campaign_files/2026-10-g-solver-instances-2/sato7f-puzzle-generation-sat/LICENSE ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ Licence: MIT
2
+
3
+ - MIT: https://github.com/sato7f/puzzle-generation-sat/blob/66cad7a41d13c2fbe73eddfe216bca54ce6aac03/LICENSE (publisher: sato7f)
campaign_files/2026-10-g-solver-instances-2/sato7f-puzzle-generation-sat/README.md ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # sato7f-puzzle-generation-sat
2
+
3
+ **What this is:** 2 units (10 files) published by sato7f. Collection theme: Seeded puzzle generators (Simon Tatham puzzle collection class, with generation command, seed and solution).
4
+
5
+ **What one unit is:** Each unit is a few named upstream files, delivered together, kept at its upstream path. Format: JSONL of generated region boards with the star set, plus a generation record; JSONL of generated puzzles with solutions, plus a generation record.
6
+
7
+ **Where it comes from:** Upstream: sato7f/puzzle-generation-sat. Kind: generated data.
8
+
9
+ **Self-contained:** Each unit is self-contained. 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:** Check: Unique solution: exact match of the star set; equivalently the upstream rule check similar.starbattle.check (similar/src/similar/starbattle.py lines 51-62: k stars per row, column and region, no two stars adjacent including diagonals). Wrong-answer controls must fail. Unique solution: exact match of the grid; equivalently the upstream similar.tango.check (similar/src/similar/tango.py lines 48-67: each line has n/2 ones, no three equal in a row, givens kept, = and x signs respected). Wrong-answer controls must fail.
12
+
13
+ **Licence (SPDX):** MIT
14
+
15
+ **Provenance:** https://github.com/sato7f/puzzle-generation-sat
16
+
17
+ Units:
18
+
19
+ - `/` (8 files): sato7f/puzzle-generation-sat region puzzles: seeded SAT-checked Star Battle (k=2) and Queens/Meowdoku (k=1) boards with unique solutions.
20
+ Format: JSONL of generated region boards with the star set, plus a generation record
21
+ Check: Unique solution: exact match of the star set; equivalently the upstream rule check similar.starbattle.check (similar/src/similar/starbattle.py lines 51-62: k stars per row, column and region, no two stars adjacent including diagonals). Wrong-answer controls must fail.
22
+ Companions (not included): similar/src/similar/starbattle.py; similar/src/similar/__init__.py; generation/scripts/run.py; generation/pyproject.toml; similar/pyproject.toml; LICENSE; generation/README.md
23
+ Dependencies: Python >= 3.12 (generation/pyproject.toml line 8); python-sat>=1.9.dev15 (CaDiCaL 1.9.5 backend; exact version recorded at generation); similar at git commit 66cad7a4 (local path package similar/ of the same repository); sato7f/puzzle-generation-sat at git commit 66cad7a4; not packed
24
+ - `/` (2 files): sato7f/puzzle-generation-sat Tango (LinkedIn binary puzzle with = and x signs): seeded, unique and hint-minimal puzzles with solutions.
25
+ Format: JSONL of generated puzzles with solutions, plus a generation record
26
+ Check: Unique solution: exact match of the grid; equivalently the upstream similar.tango.check (similar/src/similar/tango.py lines 48-67: each line has n/2 ones, no three equal in a row, givens kept, = and x signs respected). Wrong-answer controls must fail.
27
+ Companions (not included): similar/src/similar/tango.py; similar/src/similar/__init__.py; generation/scripts/run.py; generation/pyproject.toml; similar/pyproject.toml; LICENSE; generation/README.md
28
+ Dependencies: Python >= 3.12; python-sat>=1.9.dev15 (exact version recorded at generation); similar at git commit 66cad7a4 (local path package similar/ of the same repository); sato7f/puzzle-generation-sat at git commit 66cad7a4; not packed
campaign_files/2026-10-g-solver-instances-2/schedulingbenchmarks-matsp/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: Best known objective value (not proven optimal). 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: read the instance (SECTION_HORIZON, SECTION_TASKS, SECTION_STAFF, SECTION_COVER) and check every hard constraint of the format guide instances1_225.html (companion): horizon from 06:00 on day 1 to 06:00 on the day after the last, in 15-minute intervals; shift starts, ends and task changes on 15-minute boundaries; at most one shift started per employee per calendar day; at most 5 consecutive working days, with the days before and after the horizon taken as days off; total minutes per employee between MinTotalMinutes and MaxTotalMinutes of SECTION_STAFF (shift duration = start to end); at least 14 hours rest after every shift; every task duration within a shift at least 1 hour; shift length 6 to 10 hours; shift starts only within 06:00-10:00, 14:00-18:00 or 20:00-00:00; Min of SECTION_COVER met for every interval and task. Objective: the sum over every interval and task of (x - max) * (x - max) * weight where x exceeds Max of SECTION_COVER, weight 1. Pass when every hard constraint holds and the objective is at most 383 (best known, not proven optimal). The instance file itself states no constraints or objective.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -22,7 +20,5 @@ Units:
22
  Read out of the archive: https://www.schedulingbenchmarks.org/matsp/data/instances1_225.zip
23
  Format: text with SECTION_HORIZON, SECTION_TASKS, SECTION_STAFF and further sections; one case = one instance; answer = shifts and activities per employee and the objective
24
  Origin: instances by Curtois; best values from the cited reference TRA26. Standard benchmark suite or competition instance as published by the library, recorded tool_fixture.
25
- Reference result: best known objective value (not proven optimal)
26
- Reference source: matsp.html line 129 row "1 7 10 1 383 TRA26"
27
  Check: Program: read the instance (SECTION_HORIZON, SECTION_TASKS, SECTION_STAFF, SECTION_COVER) and check every hard constraint of the format guide instances1_225.html (companion): horizon from 06:00 on day 1 to 06:00 on the day after the last, in 15-minute intervals; shift starts, ends and task changes on 15-minute boundaries; at most one shift started per employee per calendar day; at most 5 consecutive working days, with the days before and after the horizon taken as days off; total minutes per employee between MinTotalMinutes and MaxTotalMinutes of SECTION_STAFF (shift duration = start to end); at least 14 hours rest after every shift; every task duration within a shift at least 1 hour; shift length 6 to 10 hours; shift starts only within 06:00-10:00, 14:00-18:00 or 20:00-00:00; Min of SECTION_COVER met for every interval and task. Objective: the sum over every interval and task of (x - max) * (x - max) * weight where x exceeds Max of SECTION_COVER, weight 1. Pass when every hard constraint holds and the objective is at most 383 (best known, not proven optimal). The instance file itself states no constraints or objective.
28
  Companions (not included): matsp/instances1_225.html
 
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: Program: read the instance (SECTION_HORIZON, SECTION_TASKS, SECTION_STAFF, SECTION_COVER) and check every hard constraint of the format guide instances1_225.html (companion): horizon from 06:00 on day 1 to 06:00 on the day after the last, in 15-minute intervals; shift starts, ends and task changes on 15-minute boundaries; at most one shift started per employee per calendar day; at most 5 consecutive working days, with the days before and after the horizon taken as days off; total minutes per employee between MinTotalMinutes and MaxTotalMinutes of SECTION_STAFF (shift duration = start to end); at least 14 hours rest after every shift; every task duration within a shift at least 1 hour; shift length 6 to 10 hours; shift starts only within 06:00-10:00, 14:00-18:00 or 20:00-00:00; Min of SECTION_COVER met for every interval and task. Objective: the sum over every interval and task of (x - max) * (x - max) * weight where x exceeds Max of SECTION_COVER, weight 1. Pass when every hard constraint holds and the objective is at most 383 (best known, not proven optimal). The instance file itself states no constraints or objective.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
20
  Read out of the archive: https://www.schedulingbenchmarks.org/matsp/data/instances1_225.zip
21
  Format: text with SECTION_HORIZON, SECTION_TASKS, SECTION_STAFF and further sections; one case = one instance; answer = shifts and activities per employee and the objective
22
  Origin: instances by Curtois; best values from the cited reference TRA26. Standard benchmark suite or competition instance as published by the library, recorded tool_fixture.
 
 
23
  Check: Program: read the instance (SECTION_HORIZON, SECTION_TASKS, SECTION_STAFF, SECTION_COVER) and check every hard constraint of the format guide instances1_225.html (companion): horizon from 06:00 on day 1 to 06:00 on the day after the last, in 15-minute intervals; shift starts, ends and task changes on 15-minute boundaries; at most one shift started per employee per calendar day; at most 5 consecutive working days, with the days before and after the horizon taken as days off; total minutes per employee between MinTotalMinutes and MaxTotalMinutes of SECTION_STAFF (shift duration = start to end); at least 14 hours rest after every shift; every task duration within a shift at least 1 hour; shift length 6 to 10 hours; shift starts only within 06:00-10:00, 14:00-18:00 or 20:00-00:00; Min of SECTION_COVER met for every interval and task. Objective: the sum over every interval and task of (x - max) * (x - max) * weight where x exceeds Max of SECTION_COVER, weight 1. Pass when every hard constraint holds and the objective is at most 383 (best known, not proven optimal). The instance file itself states no constraints or objective.
24
  Companions (not included): matsp/instances1_225.html
campaign_files/2026-10-g-solver-instances-2/schedulingbenchmarks-nrp-literature/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained. 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 objective value plus a reference roster. 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 roster, check every hard rule of Musa.ros, compute the weighted penalty of soft rules as defined in the XML; pass when feasible and the penalty equals 175.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,8 +19,6 @@ Files:
21
  - `nrp/data/Musa.ros`: Musa nurse scheduling instance in the RosterBooster XML format with proven optimal penalty 175.
22
  Format: XML SchedulingPeriod document (shift types, contracts, employees, cover requirements, requests); one case = one instance; answer = a roster and its penalty
23
  Origin: instance from Musa and Saxena 1984, converted to XML by Curtois. Standard benchmark suite or competition instance as published by the library, recorded tool_fixture.
24
- Reference result: optimal objective value plus a reference roster
25
- Reference source: https://www.schedulingbenchmarks.org/nrp/bounds.html line 1161: "The optimal solution is 175 (solved using this IP model)"; reference file Musa.Solution.175.roster
26
  Check: Program: parse the agent roster, check every hard rule of Musa.ros, compute the weighted penalty of soft rules as defined in the XML; pass when feasible and the penalty equals 175
27
  Companions (not included): nrp/data/solutions/xml/Musa.Solution.175.roster
28
  Dependencies: the site's RosterViewer GUI is not used
 
8
 
9
  **Self-contained:** The file is self-contained. 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: Program: parse the agent roster, check every hard rule of Musa.ros, compute the weighted penalty of soft rules as defined in the XML; pass when feasible and the penalty equals 175.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `nrp/data/Musa.ros`: Musa nurse scheduling instance in the RosterBooster XML format with proven optimal penalty 175.
20
  Format: XML SchedulingPeriod document (shift types, contracts, employees, cover requirements, requests); one case = one instance; answer = a roster and its penalty
21
  Origin: instance from Musa and Saxena 1984, converted to XML by Curtois. Standard benchmark suite or competition instance as published by the library, recorded tool_fixture.
 
 
22
  Check: Program: parse the agent roster, check every hard rule of Musa.ros, compute the weighted penalty of soft rules as defined in the XML; pass when feasible and the penalty equals 175
23
  Companions (not included): nrp/data/solutions/xml/Musa.Solution.175.roster
24
  Dependencies: the site's RosterViewer GUI is not used
campaign_files/2026-10-g-solver-instances-2/schedulingbenchmarks-nrp/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** 1 of 4 files are self-contained; the others are used together with other upstream files. 4 of 4 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: Optimal objective value (lower bound equals best known solution) plus a reference roster; proven optimum (penalty), best known lower bound equals best known solution. 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:** Tools are named in the check written per file in the notes.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,28 +19,20 @@ Files:
21
  - `nrp/data/Instance1.txt`: Shift scheduling benchmark Instance1 (2 weeks, 8 employees, 1 shift type) with proven optimum 607.
22
  Format: plain text with SECTION_HORIZON, SECTION_SHIFTS, SECTION_STAFF, SECTION_DAYS_OFF, SECTION_SHIFT_ON_REQUESTS, SECTION_SHIFT_OFF_REQUESTS, SECTION_COVER; one case = one instance file; the answer is a roster (shift per employee per day) and its penalty
23
  Origin: instances generated by Curtois and Qu to reflect real rostering requirements (Technical Report, University of Nottingham, 2014); bounds from the authors' IP model with Gurobi. Standard benchmark suite or competition instance as published by the library, recorded tool_fixture.
24
- Reference result: optimal objective value (lower bound equals best known solution) plus a reference roster
25
- Reference source: nrp index table (https://www.schedulingbenchmarks.org/nrp/, line 120 onward) row "Instance1 txt xml 2 8 1 607 607" (columns Weeks, Employees, Shift types, Best known lower bound, Best known solution); reference solution file Instance1.Solution.607.roster from instances1_24solutions.zip
26
  Check: Program: check the agent roster against every hard constraint of the instance (one shift per day, max shifts per type, min and max total minutes, max and min consecutive shifts, min consecutive days off, max weekends, days off, shift rotation), then compute the penalty from shift on and off requests and cover under or over staffing weights given in the file; pass when feasible and the penalty equals 607, the proven optimum
27
  Companions (not included): nrp/data/instances1_24solutions/Solutions/XML/Instance1.Solution.607.roster; nrp/verify.html
28
  - `nrp/data/Instance2.txt`: Nurse rostering Instance2: 2 weeks, 14 employees, 2 shift types; proven optimal penalty 828.
29
  Format: plain text with SECTION_ blocks (horizon, shifts, staff limits, days off, shift on and off requests with weights, cover with under and over weights), Instance2.txt lines 2 to 143
30
  Origin: Curtois and Qu nurse rostering benchmark instances 1 to 24 (schedulingbenchmarks.org/nrp)
31
- Reference result: proven optimum (penalty), best known lower bound equals best known solution
32
- Reference source: nrp/index.html lines 129 to 134: row 'Instance2' with Weeks 2, Employees 14, Shift types 2, Best known lower bound 828, Best known solution 828; reference roster Instance2.Solution.828.roster in instances1_24solutions.zip
33
  Check: Program: check the agent roster against every hard constraint of the instance (one shift per day, max shifts per type, min and max total minutes, max and min consecutive shifts, min consecutive days off, max weekends, days off, shift rotation), then compute the penalty from shift on and off requests and cover under or over staffing weights given in the file; pass when feasible and the penalty is at most 828, the proven optimum
34
  Companions (not included): nrp/data/instances1_24solutions/Solutions/XML/Instance2.Solution.828.roster; nrp/index.html
35
  - `nrp/data/Instance3.txt`: Nurse rostering Instance3: 2 weeks, 20 employees, 3 shift types; proven optimal penalty 1001.
36
  Format: plain text with SECTION_ blocks (horizon, shifts, staff limits, days off, shift on and off requests with weights, cover with under and over weights)
37
  Origin: Curtois and Qu nurse rostering benchmark instances 1 to 24 (schedulingbenchmarks.org/nrp)
38
- Reference result: proven optimum (penalty), best known lower bound equals best known solution
39
- Reference source: nrp/index.html lines 138 to 143: row 'Instance3' with Weeks 2, Employees 20, Shift types 3, Best known lower bound 1001, Best known solution 1001; reference roster Instance3.Solution.1001.roster in instances1_24solutions.zip
40
  Check: Program: check the agent roster against every hard constraint of the instance (one shift per day, max shifts per type, min and max total minutes, max and min consecutive shifts, min consecutive days off, max weekends, days off, shift rotation), then compute the penalty from shift on and off requests and cover under or over staffing weights given in the file; pass when feasible and the penalty is at most 1001, the proven optimum
41
  Companions (not included): nrp/data/instances1_24solutions/Solutions/XML/Instance3.Solution.1001.roster; nrp/index.html
42
  - `nrp/data/Instance4.txt`: Nurse rostering Instance4: 4 weeks, 10 employees, 2 shift types; proven optimal penalty 1716.
43
  Format: plain text with SECTION_ blocks (horizon, shifts, staff limits, days off, shift on and off requests with weights, cover with under and over weights)
44
  Origin: Curtois and Qu nurse rostering benchmark instances 1 to 24 (schedulingbenchmarks.org/nrp)
45
- Reference result: proven optimum (penalty), best known lower bound equals best known solution
46
- Reference source: nrp/index.html lines 147 to 152: row 'Instance4' with Weeks 4, Employees 10, Shift types 2, Best known lower bound 1716, Best known solution 1716; reference roster Instance4.Solution.1716.roster in instances1_24solutions.zip
47
  Check: Program: check the agent roster against every hard constraint of the instance (one shift per day, max shifts per type, min and max total minutes, max and min consecutive shifts, min consecutive days off, max weekends, days off, shift rotation), then compute the penalty from shift on and off requests and cover under or over staffing weights given in the file; pass when feasible and the penalty is at most 1716, the proven optimum
48
  Companions (not included): nrp/data/instances1_24solutions/Solutions/XML/Instance4.Solution.1716.roster; nrp/index.html
 
8
 
9
  **Self-contained:** 1 of 4 files are self-contained; the others are used together with other upstream files. 4 of 4 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:** Tools are named in the check written per file in the notes.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `nrp/data/Instance1.txt`: Shift scheduling benchmark Instance1 (2 weeks, 8 employees, 1 shift type) with proven optimum 607.
20
  Format: plain text with SECTION_HORIZON, SECTION_SHIFTS, SECTION_STAFF, SECTION_DAYS_OFF, SECTION_SHIFT_ON_REQUESTS, SECTION_SHIFT_OFF_REQUESTS, SECTION_COVER; one case = one instance file; the answer is a roster (shift per employee per day) and its penalty
21
  Origin: instances generated by Curtois and Qu to reflect real rostering requirements (Technical Report, University of Nottingham, 2014); bounds from the authors' IP model with Gurobi. Standard benchmark suite or competition instance as published by the library, recorded tool_fixture.
 
 
22
  Check: Program: check the agent roster against every hard constraint of the instance (one shift per day, max shifts per type, min and max total minutes, max and min consecutive shifts, min consecutive days off, max weekends, days off, shift rotation), then compute the penalty from shift on and off requests and cover under or over staffing weights given in the file; pass when feasible and the penalty equals 607, the proven optimum
23
  Companions (not included): nrp/data/instances1_24solutions/Solutions/XML/Instance1.Solution.607.roster; nrp/verify.html
24
  - `nrp/data/Instance2.txt`: Nurse rostering Instance2: 2 weeks, 14 employees, 2 shift types; proven optimal penalty 828.
25
  Format: plain text with SECTION_ blocks (horizon, shifts, staff limits, days off, shift on and off requests with weights, cover with under and over weights), Instance2.txt lines 2 to 143
26
  Origin: Curtois and Qu nurse rostering benchmark instances 1 to 24 (schedulingbenchmarks.org/nrp)
 
 
27
  Check: Program: check the agent roster against every hard constraint of the instance (one shift per day, max shifts per type, min and max total minutes, max and min consecutive shifts, min consecutive days off, max weekends, days off, shift rotation), then compute the penalty from shift on and off requests and cover under or over staffing weights given in the file; pass when feasible and the penalty is at most 828, the proven optimum
28
  Companions (not included): nrp/data/instances1_24solutions/Solutions/XML/Instance2.Solution.828.roster; nrp/index.html
29
  - `nrp/data/Instance3.txt`: Nurse rostering Instance3: 2 weeks, 20 employees, 3 shift types; proven optimal penalty 1001.
30
  Format: plain text with SECTION_ blocks (horizon, shifts, staff limits, days off, shift on and off requests with weights, cover with under and over weights)
31
  Origin: Curtois and Qu nurse rostering benchmark instances 1 to 24 (schedulingbenchmarks.org/nrp)
 
 
32
  Check: Program: check the agent roster against every hard constraint of the instance (one shift per day, max shifts per type, min and max total minutes, max and min consecutive shifts, min consecutive days off, max weekends, days off, shift rotation), then compute the penalty from shift on and off requests and cover under or over staffing weights given in the file; pass when feasible and the penalty is at most 1001, the proven optimum
33
  Companions (not included): nrp/data/instances1_24solutions/Solutions/XML/Instance3.Solution.1001.roster; nrp/index.html
34
  - `nrp/data/Instance4.txt`: Nurse rostering Instance4: 4 weeks, 10 employees, 2 shift types; proven optimal penalty 1716.
35
  Format: plain text with SECTION_ blocks (horizon, shifts, staff limits, days off, shift on and off requests with weights, cover with under and over weights)
36
  Origin: Curtois and Qu nurse rostering benchmark instances 1 to 24 (schedulingbenchmarks.org/nrp)
 
 
37
  Check: Program: check the agent roster against every hard constraint of the instance (one shift per day, max shifts per type, min and max total minutes, max and min consecutive shifts, min consecutive days off, max weekends, days off, shift rotation), then compute the penalty from shift on and off requests and cover under or over staffing weights given in the file; pass when feasible and the penalty is at most 1716, the proven optimum
38
  Companions (not included): nrp/data/instances1_24solutions/Solutions/XML/Instance4.Solution.1716.roster; nrp/index.html
campaign_files/2026-10-g-solver-instances-2/schedulinglab-fjsp-instances/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: Proven optimum makespan. 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: check each operation is on an eligible machine with its listed time, job precedence holds and machines do not overlap; makespan must be at most 2503.
14
 
15
  **Licence (SPDX):** MIT
@@ -21,7 +19,5 @@ Files:
21
  - `dauzere/04a.txt`: Flexible job shop instance dpp04 (Dauzere-Peres and Paulli 04a, 10 jobs, 5 machines) with proven optimum 2503.
22
  Format: plain text Brandimarte FJSP format: line 1 'jobs machines', then per job the number of operations and for each operation the count of eligible machines followed by (machine, time) pairs; one case
23
  Origin: Dauzere-Peres and Paulli (1997) FJSP benchmark, curated with bounds by SchedulingLab
24
- Reference result: proven optimum makespan
25
- Reference source: instances.json lines 1935 to 1939: name dpp04, jobs 10, machines 5, "optimum" : 2503, path dauzere/04a.txt
26
  Check: Program: check each operation is on an eligible machine with its listed time, job precedence holds and machines do not overlap; makespan must be at most 2503
27
  Companions (not included): instances.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.
10
 
 
 
11
  **Tools to open and check:** Check: Program: check each operation is on an eligible machine with its listed time, job precedence holds and machines do not overlap; makespan must be at most 2503.
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `dauzere/04a.txt`: Flexible job shop instance dpp04 (Dauzere-Peres and Paulli 04a, 10 jobs, 5 machines) with proven optimum 2503.
20
  Format: plain text Brandimarte FJSP format: line 1 'jobs machines', then per job the number of operations and for each operation the count of eligible machines followed by (machine, time) pairs; one case
21
  Origin: Dauzere-Peres and Paulli (1997) FJSP benchmark, curated with bounds by SchedulingLab
 
 
22
  Check: Program: check each operation is on an eligible machine with its listed time, job precedence holds and machines do not overlap; makespan must be at most 2503
23
  Companions (not included): instances.json
campaign_files/2026-10-g-solver-instances/sdplib/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: Published optimal objective 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: Compare the agent's optimal objective with the README value at the 6 significant digits printed there.
14
 
15
  **Licence (SPDX):** GPL-3.0-only
@@ -21,8 +19,6 @@ Files:
21
  - `data/arch0.dat-s`: SDPLIB instance arch0 (truss topology semidefinite program, m=174, n=335) with the published optimal objective 5.66517e-01.
22
  Format: SDPA sparse format (.dat-s); answer in companion README.md table (Problem, m, n, Optimal Objective Value, Notes)
23
  Origin: truss topology design SDP from the SDPLIB 1.2 collection
24
- Reference result: published optimal objective value
25
- Reference source: README.md line 32: "| arch0 | 174 | 335 | 5.66517e-01 | 1 |" (table header line 30, SDPA sign conventions per the README).
26
  Check: Compare the agent's optimal objective with the README value at the 6 significant digits printed there.
27
  Companions (not included): README.md
28
  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: Compare the agent's optimal objective with the README value at the 6 significant digits printed there.
12
 
13
  **Licence (SPDX):** GPL-3.0-only
 
19
  - `data/arch0.dat-s`: SDPLIB instance arch0 (truss topology semidefinite program, m=174, n=335) with the published optimal objective 5.66517e-01.
20
  Format: SDPA sparse format (.dat-s); answer in companion README.md table (Problem, m, n, Optimal Objective Value, Notes)
21
  Origin: truss topology design SDP from the SDPLIB 1.2 collection
 
 
22
  Check: Compare the agent's optimal objective with the README value at the 6 significant digits printed there.
23
  Companions (not included): README.md
24
  Dependencies: not packed
campaign_files/2026-10-gt-browser-tests-2/ruby-icalendar-fixtures/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: 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:** Ruby icalendar gem (Icalendar::Calendar.parse then to_ical), or Python icalendar from_ical then to_ical; the script writes the re-serialized calendar to a file
14
 
15
  **Licence (SPDX):** BSD-2-Clause OR GPL-3.0-only OR LicenseRef-icalendar
@@ -21,7 +19,5 @@ Files:
21
  - `spec/fixtures/timezone.ics`: Round-trip a calendar with a VTIMEZONE through parse and serialize without changing a byte.
22
  Format: iCalendar (.ics)
23
  Origin: Ruby icalendar spec fixture (tool_fixture)
24
- Reference result: golden_output_files
25
- Reference source: the input file itself: spec/roundtrip_spec.rb lines 36-41 require Calendar.parse(source).first.to_ical to equal the source
26
  Check: Final file state: the re-serialized file must equal timezone.ics byte for byte (expect(ical).to eq source). No screenshot
27
  Tools: Ruby icalendar gem (Icalendar::Calendar.parse then to_ical), or Python icalendar from_ical then to_ical; the script writes the re-serialized calendar to a file
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Ruby icalendar gem (Icalendar::Calendar.parse then to_ical), or Python icalendar from_ical then to_ical; the script writes the re-serialized calendar to a file
12
 
13
  **Licence (SPDX):** BSD-2-Clause OR GPL-3.0-only OR LicenseRef-icalendar
 
19
  - `spec/fixtures/timezone.ics`: Round-trip a calendar with a VTIMEZONE through parse and serialize without changing a byte.
20
  Format: iCalendar (.ics)
21
  Origin: Ruby icalendar spec fixture (tool_fixture)
 
 
22
  Check: Final file state: the re-serialized file must equal timezone.ics byte for byte (expect(ical).to eq source). No screenshot
23
  Tools: Ruby icalendar gem (Icalendar::Calendar.parse then to_ical), or Python icalendar from_ical then to_ical; the script writes the re-serialized calendar to a file
campaign_files/2026-10-gt-browser-tests-2/ruby-mail-dsn-report-530/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained. 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_values_in_test_code. 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:** Ruby mail gem (Mail.read, bounced?, action, final_recipient, error_status, diagnostic_code, remote_mta, retryable?) or Python email with the message/delivery-status part; a script parses the bounce and writes the fields to JSON
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `spec/fixtures/emails/multipart_report_emails/report_530.eml`: Extract the delivery status fields from a permanent-failure bounce.
22
  Format: RFC 822 message (.eml) with a multipart/report delivery status; expected values in an RSpec file
23
  Origin: Ruby mail gem multipart report fixture (tool_fixture)
24
- Reference result: expected_values_in_test_code
25
- Reference source: spec/mail/multipart_report_spec.rb lines 91-124 (permanent failure): bounced true, action 'failed', final_recipient 'RFC822; edwin@zzzzzzz.com', error_status '5.3.0', diagnostic_code 'SMTP; 553 5.3.0 <edwin@zzzzzzz.com>... Unknown E-Mail Address', remote_mta 'DNS; mail.zzzzzz.com', not retryable
26
  Check: Final file state: the agent writes a JSON object with bounced, action, final_recipient, error_status, diagnostic_code, remote_mta and retryable; each string must equal the expected value exactly and the two booleans must match. No screenshot
27
  Tools: Ruby mail gem (Mail.read, bounced?, action, final_recipient, error_status, diagnostic_code, remote_mta, retryable?) or Python email with the message/delivery-status part; a script parses the bounce and writes the fields to JSON
28
  Companions (not included): spec/mail/multipart_report_spec.rb
 
8
 
9
  **Self-contained:** The file is self-contained. 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:** Ruby mail gem (Mail.read, bounced?, action, final_recipient, error_status, diagnostic_code, remote_mta, retryable?) or Python email with the message/delivery-status part; a script parses the bounce and writes the fields to JSON
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `spec/fixtures/emails/multipart_report_emails/report_530.eml`: Extract the delivery status fields from a permanent-failure bounce.
20
  Format: RFC 822 message (.eml) with a multipart/report delivery status; expected values in an RSpec file
21
  Origin: Ruby mail gem multipart report fixture (tool_fixture)
 
 
22
  Check: Final file state: the agent writes a JSON object with bounced, action, final_recipient, error_status, diagnostic_code, remote_mta and retryable; each string must equal the expected value exactly and the two booleans must match. No screenshot
23
  Tools: Ruby mail gem (Mail.read, bounced?, action, final_recipient, error_status, diagnostic_code, remote_mta, retryable?) or Python email with the message/delivery-status part; a script parses the bounce and writes the fields to JSON
24
  Companions (not included): spec/mail/multipart_report_spec.rb
campaign_files/2026-10-gt-browser-tests-2/sabre-vobject-itip/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: inline_expected_values. 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:** sabre/vobject ITip Broker processMessage (PHP), or Python icalendar editing ATTENDEE PARTSTAT and SCHEDULE-STATUS by hand; a script applies each REPLY to the old calendar and writes the updated .ics
14
 
15
  **Licence (SPDX):** BSD-3-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `tests/VObject/ITip/BrokerProcessReplyTest.php`: Apply iTIP REPLY messages to stored events and update attendee participation status.
22
  Format: PHP test file holding iCalendar texts inline
23
  Origin: sabre/vobject iTIP broker tests (tool_fixture)
24
- Reference result: inline_expected_values
25
- Reference source: 14 test methods of BrokerProcessReplyTest.php (lines 7-711), each with the REPLY message, the old calendar and the expected calendar inline; BrokerTester.php process() lines 56-107 runs them
26
  Check: Final file state: for each case the updated calendar must equal the expected one after removing PRODID and CALSCALE:GREGORIAN, with **ANY** in the expected text matching any value; an expected null means the reply must be ignored. No screenshot
27
  Tools: sabre/vobject ITip Broker processMessage (PHP), or Python icalendar editing ATTENDEE PARTSTAT and SCHEDULE-STATUS by hand; a script applies each REPLY to the old calendar and writes the updated .ics
28
  Companions (not included): tests/VObject/ITip/BrokerTester.php; lib/PHPUnitAssertions.php
 
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:** sabre/vobject ITip Broker processMessage (PHP), or Python icalendar editing ATTENDEE PARTSTAT and SCHEDULE-STATUS by hand; a script applies each REPLY to the old calendar and writes the updated .ics
12
 
13
  **Licence (SPDX):** BSD-3-Clause
 
19
  - `tests/VObject/ITip/BrokerProcessReplyTest.php`: Apply iTIP REPLY messages to stored events and update attendee participation status.
20
  Format: PHP test file holding iCalendar texts inline
21
  Origin: sabre/vobject iTIP broker tests (tool_fixture)
 
 
22
  Check: Final file state: for each case the updated calendar must equal the expected one after removing PRODID and CALSCALE:GREGORIAN, with **ANY** in the expected text matching any value; an expected null means the reply must be ignored. No screenshot
23
  Tools: sabre/vobject ITip Broker processMessage (PHP), or Python icalendar editing ATTENDEE PARTSTAT and SCHEDULE-STATUS by hand; a script applies each REPLY to the old calendar and writes the updated .ics
24
  Companions (not included): tests/VObject/ITip/BrokerTester.php; lib/PHPUnitAssertions.php
campaign_files/2026-10-hw-eda-tests-2/rp-rs-rp2040-pac/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: generated_code_register_values. 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: Task: apply rp2040.yaml to the SVD and run svd2rust, or hand-write the TIMER access code. Judge: compare the TIMER block pointer and every register offset and field bit offset and width in the generated Rust against the SVD by exact integer equality; compile check of the crate optional.
14
 
15
  **Licence (SPDX):** BSD-3-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `svd/rp2040.svd`: RP2040 PAC: generate the Rust register access code for TIMER from the SVD with svd2rust; addresses and field bits must match.
22
  Format: CMSIS-SVD XML; svdtools YAML patch; Rust source
23
  Origin: community crate generated from the vendor SVD; README and lib.rs disagree on the svd2rust version (0.29.0 vs 0.32.0), flagged
24
- Reference result: generated_code_register_values
25
- Reference source: src/lib.rs line 1 `generated using svd2rust v0.32.0`; src/lib.rs line 1630 `pub const PTR: *const timer::RegisterBlock = 0x4005_4000 as *const _;` against svd/rp2040.svd line 28366 `<baseAddress>0x40054000</baseAddress>`; src/timer/inte.rs line 5 field ALARM_0 against svd/rp2040.svd line 28550
26
  Check: Task: apply rp2040.yaml to the SVD and run svd2rust, or hand-write the TIMER access code. Judge: compare the TIMER block pointer and every register offset and field bit offset and width in the generated Rust against the SVD by exact integer equality; compile check of the crate optional.
27
  Companions (not included): Cargo.toml; README.md; src/lib.rs; src/timer.rs; src/timer/alarm0.rs; src/timer/inte.rs; svd/rp2040.yaml
28
  Dependencies: svd2rust 0.32.0 (named in src/lib.rs line 1; README.md line 9 says 0.29.0, a mismatch): only needed to regenerate; svdtools (version not pinned upstream, recorded as pinned to the crate commit) for applying rp2040.yaml; Python 3.9 or later standard library for the comparison judge
 
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: Task: apply rp2040.yaml to the SVD and run svd2rust, or hand-write the TIMER access code. Judge: compare the TIMER block pointer and every register offset and field bit offset and width in the generated Rust against the SVD by exact integer equality; compile check of the crate optional.
12
 
13
  **Licence (SPDX):** BSD-3-Clause
 
19
  - `svd/rp2040.svd`: RP2040 PAC: generate the Rust register access code for TIMER from the SVD with svd2rust; addresses and field bits must match.
20
  Format: CMSIS-SVD XML; svdtools YAML patch; Rust source
21
  Origin: community crate generated from the vendor SVD; README and lib.rs disagree on the svd2rust version (0.29.0 vs 0.32.0), flagged
 
 
22
  Check: Task: apply rp2040.yaml to the SVD and run svd2rust, or hand-write the TIMER access code. Judge: compare the TIMER block pointer and every register offset and field bit offset and width in the generated Rust against the SVD by exact integer equality; compile check of the crate optional.
23
  Companions (not included): Cargo.toml; README.md; src/lib.rs; src/timer.rs; src/timer/alarm0.rs; src/timer/inte.rs; svd/rp2040.yaml
24
  Dependencies: svd2rust 0.32.0 (named in src/lib.rs line 1; README.md line 9 says 0.29.0, a mismatch): only needed to regenerate; svdtools (version not pinned upstream, recorded as pinned to the crate commit) for applying rp2040.yaml; Python 3.9 or later standard library for the comparison judge
campaign_files/2026-10-hw-eda-tests-2/rust-embedded-cortex-m-testsuite-github/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: emulator_selfcheck_pass. 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: Task: given the crates and the testsuite specification, write (or complete) the tests so that they pass on the emulated Cortex-M. Judge: cargo test --features qemu --target thumbv7em-none-eabihf (and thumbv6m-none-eabi --release) in testsuite/ builds the defmt-test binary and runs it with qemu-run on the board named in .cargo/config.toml; a zero exit status passes. cycle_count runs on hardware only and sau_set_get_region only on Armv8-M.
14
 
15
  **Licence (SPDX):** MIT OR Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `testsuite/src/lib.rs`: Rust cortex-m testsuite on QEMU (lm3s6965evb Cortex-M3, microbit Cortex-M0): peripheral singleton, nested critical sections with PendSV, interrupt-free nesting, stack handles, inline assembly, semihosting write, running on the process stack with SVCall and a final breakpoint fault.
22
  Format: Rust test module (no_std, defmt-test) built for thumbv7em-none-eabihf or thumbv6m-none-eabi and run with qemu-run; verdict is the semihosting exit status
23
  Origin: the cortex-m crates' own testsuite (defmt-test), run under QEMU with qemu-run; written by the project
24
- Reference result: emulator_selfcheck_pass
25
- Reference source: the assertions in testsuite/src/lib.rs (for example double_take line 75, critical_section_nesting lines 110-114, check_stack_handles line 136, user_fn lines 206-212); the HardFault handler ends the run with semihosting EXIT_SUCCESS only when the final breakpoint was wanted (lines 38-56, exit() at lines 222-225), a panic or an unexpected fault ends with EXIT_FAILURE (fail() at lines 228-231)
26
  Check: Task: given the crates and the testsuite specification, write (or complete) the tests so that they pass on the emulated Cortex-M. Judge: cargo test --features qemu --target thumbv7em-none-eabihf (and thumbv6m-none-eabi --release) in testsuite/ builds the defmt-test binary and runs it with qemu-run on the board named in .cargo/config.toml; a zero exit status passes. cycle_count runs on hardware only and sau_set_get_region only on Armv8-M.
27
  Companions (not included): testsuite/Cargo.toml; testsuite/build.rs; testsuite/memory_m3.x; testsuite/src/main.rs; testsuite/README.md; testsuite/.cargo/config.toml; LICENSE-MIT; LICENSE-APACHE
28
  Dependencies: crates cortex-m, cortex-m-rt and cortex-m-semihosting from rust-embedded/cortex-m at commit f259e379 (MIT OR Apache-2.0; path dependencies; not packed); crates defmt 1, defmt-test 0.4, defmt-semihosting 0.3 from crates.io (MIT OR Apache-2.0; not packed); Rust toolchain 1.85 or later with thumbv7em-none-eabihf and thumbv6m-none-eabi (testsuite/Cargo.toml line 6 `rust-version = "1.85"`; not packed); qemu-run and qemu-system-arm with machines lm3s6965evb and microbit
 
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: Task: given the crates and the testsuite specification, write (or complete) the tests so that they pass on the emulated Cortex-M. Judge: cargo test --features qemu --target thumbv7em-none-eabihf (and thumbv6m-none-eabi --release) in testsuite/ builds the defmt-test binary and runs it with qemu-run on the board named in .cargo/config.toml; a zero exit status passes. cycle_count runs on hardware only and sau_set_get_region only on Armv8-M.
12
 
13
  **Licence (SPDX):** MIT OR Apache-2.0
 
19
  - `testsuite/src/lib.rs`: Rust cortex-m testsuite on QEMU (lm3s6965evb Cortex-M3, microbit Cortex-M0): peripheral singleton, nested critical sections with PendSV, interrupt-free nesting, stack handles, inline assembly, semihosting write, running on the process stack with SVCall and a final breakpoint fault.
20
  Format: Rust test module (no_std, defmt-test) built for thumbv7em-none-eabihf or thumbv6m-none-eabi and run with qemu-run; verdict is the semihosting exit status
21
  Origin: the cortex-m crates' own testsuite (defmt-test), run under QEMU with qemu-run; written by the project
 
 
22
  Check: Task: given the crates and the testsuite specification, write (or complete) the tests so that they pass on the emulated Cortex-M. Judge: cargo test --features qemu --target thumbv7em-none-eabihf (and thumbv6m-none-eabi --release) in testsuite/ builds the defmt-test binary and runs it with qemu-run on the board named in .cargo/config.toml; a zero exit status passes. cycle_count runs on hardware only and sau_set_get_region only on Armv8-M.
23
  Companions (not included): testsuite/Cargo.toml; testsuite/build.rs; testsuite/memory_m3.x; testsuite/src/main.rs; testsuite/README.md; testsuite/.cargo/config.toml; LICENSE-MIT; LICENSE-APACHE
24
  Dependencies: crates cortex-m, cortex-m-rt and cortex-m-semihosting from rust-embedded/cortex-m at commit f259e379 (MIT OR Apache-2.0; path dependencies; not packed); crates defmt 1, defmt-test 0.4, defmt-semihosting 0.3 from crates.io (MIT OR Apache-2.0; not packed); Rust toolchain 1.85 or later with thumbv7em-none-eabihf and thumbv6m-none-eabi (testsuite/Cargo.toml line 6 `rust-version = "1.85"`; not packed); qemu-run and qemu-system-arm with machines lm3s6965evb and microbit
campaign_files/2026-10-hw-eda-tests-2/rust-embedded-riscv-tests-qemu-examples-and-ci-expected-golden-files-github/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 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: emulator_stdout_exact. 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: Task: given the program (or its specification), produce the exact serial text, or write a program whose run matches it. Judge: cargo xtask qemu --target <riscv target> --example multi_hart --features multi-hart builds the example, runs it under QEMU virt with 2 harts, filters and trims stdout and compares it exactly with ci/expected/multi_hart.run. Task: given the program (or its specification), produce the exact serial text, or write a program whose run matches it. Judge: cargo xtask qemu --target <riscv target> --example qemu_uart (optionally --features s-mode) builds the example, runs it under QEMU virt, filters and trims stdout and compares it exactly with ci/expected/qemu_uart.run.
14
 
15
  **Licence (SPDX):** MIT OR Apache-2.0
@@ -21,16 +19,12 @@ Files:
21
  - `tests-qemu/examples/multi_hart.rs`: Rust on QEMU RISC-V virt: two harts synchronise through shared atomics and inter-processor interrupts; hart 0 initialises, hart 1 runs, hart 0 reports both done; produce the exact serial output.
22
  Format: Rust source file (no_std, no_main example of tests-qemu) built for a bare-metal RISC-V target and run under QEMU virt; reference output is plain text stdout
23
  Origin: example program of the riscv-rt runtime crates, run under QEMU in the project's CI against a golden output file; written by the project as a regression test
24
- Reference result: emulator_stdout_exact
25
- Reference source: ci/expected/multi_hart.run (3 lines, 61 bytes), the golden file the xtask runner compares with QEMU stdout; written by the project
26
  Check: Task: given the program (or its specification), produce the exact serial text, or write a program whose run matches it. Judge: cargo xtask qemu --target <riscv target> --example multi_hart --features multi-hart builds the example, runs it under QEMU virt with 2 harts, filters and trims stdout and compares it exactly with ci/expected/multi_hart.run.
27
  Companions (not included): ci/expected/multi_hart.run; xtask/src/main.rs; tests-qemu/Cargo.toml; tests-qemu/build.rs; tests-qemu/memory.x; tests-qemu/memory-multihart.x; tests-qemu/memory-s-mode.x; .github/workflows/qemu.yaml; riscv/Cargo.toml; README.md; LICENSE-MIT.md; LICENSE-APACHE.md
28
  Dependencies: crates riscv, riscv-rt and riscv-semihosting from rust-embedded/riscv at commit 1e0c91df (MIT OR Apache-2.0; path dependencies of tests-qemu; not packed); Rust stable toolchain with the riscv32 and riscv64 bare-metal targets; crate panic-halt 1.0 from crates.io (MIT OR Apache-2.0; not packed); QEMU qemu-system-riscv32 and qemu-system-riscv64 with machine virt, and OpenSBI for the s-mode run
29
  - `tests-qemu/examples/qemu_uart.rs`: Rust on QEMU RISC-V virt: drive the NS16550 UART at 0x1000_0000 directly (in machine mode with -bios none, and in supervisor mode under OpenSBI); produce the exact serial output.
30
  Format: Rust source file (no_std, no_main example of tests-qemu) built for a bare-metal RISC-V target and run under QEMU virt; reference output is plain text stdout
31
  Origin: example program of the riscv-rt runtime crates, run under QEMU in the project's CI against a golden output file; written by the project as a regression test
32
- Reference result: emulator_stdout_exact
33
- Reference source: ci/expected/qemu_uart.run (1 line), the golden file the xtask runner compares with QEMU stdout; written by the project
34
  Check: Task: given the program (or its specification), produce the exact serial text, or write a program whose run matches it. Judge: cargo xtask qemu --target <riscv target> --example qemu_uart (optionally --features s-mode) builds the example, runs it under QEMU virt, filters and trims stdout and compares it exactly with ci/expected/qemu_uart.run.
35
  Companions (not included): ci/expected/qemu_uart.run; xtask/src/main.rs; tests-qemu/Cargo.toml; tests-qemu/build.rs; tests-qemu/memory.x; tests-qemu/memory-multihart.x; tests-qemu/memory-s-mode.x; .github/workflows/qemu.yaml; riscv/Cargo.toml; README.md; LICENSE-MIT.md; LICENSE-APACHE.md
36
  Dependencies: crates riscv, riscv-rt and riscv-semihosting from rust-embedded/riscv at commit 1e0c91df (MIT OR Apache-2.0; path dependencies of tests-qemu; not packed); Rust stable toolchain with the riscv32 and riscv64 bare-metal targets; crate panic-halt 1.0 from crates.io (MIT OR Apache-2.0; not packed); QEMU qemu-system-riscv32 and qemu-system-riscv64 with machine virt, and OpenSBI for the s-mode run
 
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 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: Task: given the program (or its specification), produce the exact serial text, or write a program whose run matches it. Judge: cargo xtask qemu --target <riscv target> --example multi_hart --features multi-hart builds the example, runs it under QEMU virt with 2 harts, filters and trims stdout and compares it exactly with ci/expected/multi_hart.run. Task: given the program (or its specification), produce the exact serial text, or write a program whose run matches it. Judge: cargo xtask qemu --target <riscv target> --example qemu_uart (optionally --features s-mode) builds the example, runs it under QEMU virt, filters and trims stdout and compares it exactly with ci/expected/qemu_uart.run.
12
 
13
  **Licence (SPDX):** MIT OR Apache-2.0
 
19
  - `tests-qemu/examples/multi_hart.rs`: Rust on QEMU RISC-V virt: two harts synchronise through shared atomics and inter-processor interrupts; hart 0 initialises, hart 1 runs, hart 0 reports both done; produce the exact serial output.
20
  Format: Rust source file (no_std, no_main example of tests-qemu) built for a bare-metal RISC-V target and run under QEMU virt; reference output is plain text stdout
21
  Origin: example program of the riscv-rt runtime crates, run under QEMU in the project's CI against a golden output file; written by the project as a regression test
 
 
22
  Check: Task: given the program (or its specification), produce the exact serial text, or write a program whose run matches it. Judge: cargo xtask qemu --target <riscv target> --example multi_hart --features multi-hart builds the example, runs it under QEMU virt with 2 harts, filters and trims stdout and compares it exactly with ci/expected/multi_hart.run.
23
  Companions (not included): ci/expected/multi_hart.run; xtask/src/main.rs; tests-qemu/Cargo.toml; tests-qemu/build.rs; tests-qemu/memory.x; tests-qemu/memory-multihart.x; tests-qemu/memory-s-mode.x; .github/workflows/qemu.yaml; riscv/Cargo.toml; README.md; LICENSE-MIT.md; LICENSE-APACHE.md
24
  Dependencies: crates riscv, riscv-rt and riscv-semihosting from rust-embedded/riscv at commit 1e0c91df (MIT OR Apache-2.0; path dependencies of tests-qemu; not packed); Rust stable toolchain with the riscv32 and riscv64 bare-metal targets; crate panic-halt 1.0 from crates.io (MIT OR Apache-2.0; not packed); QEMU qemu-system-riscv32 and qemu-system-riscv64 with machine virt, and OpenSBI for the s-mode run
25
  - `tests-qemu/examples/qemu_uart.rs`: Rust on QEMU RISC-V virt: drive the NS16550 UART at 0x1000_0000 directly (in machine mode with -bios none, and in supervisor mode under OpenSBI); produce the exact serial output.
26
  Format: Rust source file (no_std, no_main example of tests-qemu) built for a bare-metal RISC-V target and run under QEMU virt; reference output is plain text stdout
27
  Origin: example program of the riscv-rt runtime crates, run under QEMU in the project's CI against a golden output file; written by the project as a regression test
 
 
28
  Check: Task: given the program (or its specification), produce the exact serial text, or write a program whose run matches it. Judge: cargo xtask qemu --target <riscv target> --example qemu_uart (optionally --features s-mode) builds the example, runs it under QEMU virt, filters and trims stdout and compares it exactly with ci/expected/qemu_uart.run.
29
  Companions (not included): ci/expected/qemu_uart.run; xtask/src/main.rs; tests-qemu/Cargo.toml; tests-qemu/build.rs; tests-qemu/memory.x; tests-qemu/memory-multihart.x; tests-qemu/memory-s-mode.x; .github/workflows/qemu.yaml; riscv/Cargo.toml; README.md; LICENSE-MIT.md; LICENSE-APACHE.md
30
  Dependencies: crates riscv, riscv-rt and riscv-semihosting from rust-embedded/riscv at commit 1e0c91df (MIT OR Apache-2.0; path dependencies of tests-qemu; not packed); Rust stable toolchain with the riscv32 and riscv64 bare-metal targets; crate panic-halt 1.0 from crates.io (MIT OR Apache-2.0; not packed); QEMU qemu-system-riscv32 and qemu-system-riscv64 with machine virt, and OpenSBI for the s-mode run
campaign_files/2026-10-hw-eda-tests-2/rust-embedded-svdtools-test-fixture-res-cluster/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: patched_svd_register_geometry. 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: Task: apply patch.yaml to esp32s3.svd. Judge: flatten both the produced SVD and expected.svd to (cluster, register, absolute offset, field, bit offset, bit width) tuples and compare the sets by exact equality, plus check that every renamed field (MPS, XFERCOMPL) exists.
14
 
15
  **Licence (SPDX):** MIT OR Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `res/cluster/esp32s3.svd`: Apply an svdtools _cluster patch to an ESP32-S3 USB SVD excerpt: group endpoint registers into clusters and rename fields, matching the expected SVD.
22
  Format: CMSIS-SVD XML; svdtools YAML patch
23
  Origin: regression fixture of svdtools built from a real Espressif SVD excerpt
24
- Reference result: patched_svd_register_geometry
25
- Reference source: res/cluster/expected.svd (committed expected output of the svdtools test for this patch); e.g. expected.svd line 18 `<cluster>` holding DIEPCTL with field MPS renamed from D_MPS0 per patch.yaml lines 7-10
26
  Check: Task: apply patch.yaml to esp32s3.svd. Judge: flatten both the produced SVD and expected.svd to (cluster, register, absolute offset, field, bit offset, bit width) tuples and compare the sets by exact equality, plus check that every renamed field (MPS, XFERCOMPL) exists.
27
  Companions (not included): LICENSE-MIT; LICENSE-APACHE; res/cluster/patch.yaml; res/cluster/expected.svd
28
  Dependencies: svdtools: only needed to regenerate; Python 3.9 or later standard library (xml.etree) for the comparison judge; Ruby 3.x standard library Psych to read the YAML patch in the judge
 
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: Task: apply patch.yaml to esp32s3.svd. Judge: flatten both the produced SVD and expected.svd to (cluster, register, absolute offset, field, bit offset, bit width) tuples and compare the sets by exact equality, plus check that every renamed field (MPS, XFERCOMPL) exists.
12
 
13
  **Licence (SPDX):** MIT OR Apache-2.0
 
19
  - `res/cluster/esp32s3.svd`: Apply an svdtools _cluster patch to an ESP32-S3 USB SVD excerpt: group endpoint registers into clusters and rename fields, matching the expected SVD.
20
  Format: CMSIS-SVD XML; svdtools YAML patch
21
  Origin: regression fixture of svdtools built from a real Espressif SVD excerpt
 
 
22
  Check: Task: apply patch.yaml to esp32s3.svd. Judge: flatten both the produced SVD and expected.svd to (cluster, register, absolute offset, field, bit offset, bit width) tuples and compare the sets by exact equality, plus check that every renamed field (MPS, XFERCOMPL) exists.
23
  Companions (not included): LICENSE-MIT; LICENSE-APACHE; res/cluster/patch.yaml; res/cluster/expected.svd
24
  Dependencies: svdtools: only needed to regenerate; Python 3.9 or later standard library (xml.etree) for the comparison judge; Ruby 3.x standard library Psych to read the YAML patch in the judge
campaign_files/2026-10-hw-eda-tests-2/secworks-chacha/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: self_checking_testbench_known_vectors. 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 make sim-top and make sim-core in toolruns/ (targets at toolruns/Makefile:72 and 68: iverilog -Wall then vvp) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the logs: the unit passes when tb_chacha prints "*** All NN test cases completed successfully" (tb_chacha.v:412-414) and tb_chacha_core prints "*** All test cases completed successfully" (tb_chacha_core.v:447-449), both printed only when error_ctr == 0. For a write-the-RTL task, hide src/rtl/*.v and keep the testbenches.
14
 
15
  **Licence (SPDX):** BSD-2-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `src/tb/tb_chacha.v`: Secworks chacha: ChaCha stream cipher core (8, 12 and 20 rounds, 128 and 256 bit keys) with memory-mapped top level, verified by self-checking Verilog testbenches.
22
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
23
  Origin: self-checking testbench by the core author, with expected outputs taken from published standard test vectors (NIST, IETF draft)
24
- Reference result: self_checking_testbench_known_vectors
25
- Reference source: tb_chacha.v and tb_chacha_core.v hard-code test cases (key, key length, IV, round count 8, 12 or 20, and the expected 512-bit keystream block) and compare the produced block with the expected one, counting errors; the testbenches do not name the origin of the vectors, so the expected blocks are the author's own and can be recomputed with any independent ChaCha implementation (for example the Python model src/model/python/chacha.py in the repo or a library one)
26
  Check: Run make sim-top and make sim-core in toolruns/ (targets at toolruns/Makefile:72 and 68: iverilog -Wall then vvp) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the logs: the unit passes when tb_chacha prints "*** All NN test cases completed successfully" (tb_chacha.v:412-414) and tb_chacha_core prints "*** All test cases completed successfully" (tb_chacha_core.v:447-449), both printed only when error_ctr == 0. For a write-the-RTL task, hide src/rtl/*.v and keep the testbenches.
27
  Companions (not included): src/rtl/chacha.v; src/rtl/chacha_core.v; src/rtl/chacha_qr.v; src/tb/tb_chacha_core.v; toolruns/Makefile; LICENSE
28
  Dependencies: Icarus Verilog v12_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2026-01-17, GitHub release published 2023-06-11, the next one v13_0 came 2026-03-02; https://github.com/steveicarus/iverilog/releases/tag/v12_0); open source, not packed; GNU make (runs toolruns/Makefile; any current version); open source, 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: Run make sim-top and make sim-core in toolruns/ (targets at toolruns/Makefile:72 and 68: iverilog -Wall then vvp) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the logs: the unit passes when tb_chacha prints "*** All NN test cases completed successfully" (tb_chacha.v:412-414) and tb_chacha_core prints "*** All test cases completed successfully" (tb_chacha_core.v:447-449), both printed only when error_ctr == 0. For a write-the-RTL task, hide src/rtl/*.v and keep the testbenches.
12
 
13
  **Licence (SPDX):** BSD-2-Clause
 
19
  - `src/tb/tb_chacha.v`: Secworks chacha: ChaCha stream cipher core (8, 12 and 20 rounds, 128 and 256 bit keys) with memory-mapped top level, verified by self-checking Verilog testbenches.
20
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
21
  Origin: self-checking testbench by the core author, with expected outputs taken from published standard test vectors (NIST, IETF draft)
 
 
22
  Check: Run make sim-top and make sim-core in toolruns/ (targets at toolruns/Makefile:72 and 68: iverilog -Wall then vvp) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the logs: the unit passes when tb_chacha prints "*** All NN test cases completed successfully" (tb_chacha.v:412-414) and tb_chacha_core prints "*** All test cases completed successfully" (tb_chacha_core.v:447-449), both printed only when error_ctr == 0. For a write-the-RTL task, hide src/rtl/*.v and keep the testbenches.
23
  Companions (not included): src/rtl/chacha.v; src/rtl/chacha_core.v; src/rtl/chacha_qr.v; src/tb/tb_chacha_core.v; toolruns/Makefile; LICENSE
24
  Dependencies: Icarus Verilog v12_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2026-01-17, GitHub release published 2023-06-11, the next one v13_0 came 2026-03-02; https://github.com/steveicarus/iverilog/releases/tag/v12_0); open source, not packed; GNU make (runs toolruns/Makefile; any current version); open source, not packed
campaign_files/2026-10-hw-eda-tests-2/secworks-cmac/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: self_checking_testbench_known_vectors. 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 make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_cmac.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "07 test completed. All test cases completed successfully." (printed only when error_ctr == 0) (tb_cmac.v:227) and fails on "test cases did not complete successfully" (tb_cmac.v:231). For a write-the-RTL task, hide src/rtl/*.v and keep tb_cmac.v.
14
 
15
  **Licence (SPDX):** BSD-2-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `src/tb/tb_cmac.v`: Secworks cmac: AES-CMAC message authentication core with memory-mapped top level, verified by the self-checking Verilog testbench tb_cmac.v.
22
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
23
  Origin: self-checking testbench by the core author, with expected outputs hard-coded in the testbench (published standard vectors where the file cites them)
24
- Reference result: self_checking_testbench_known_vectors
25
- Reference source: tb_cmac.v runs 7 test cases (tasks tc1 to tc7 at lines 445-816) with the AES-CMAC vectors of NIST SP 800-38B Appendix D.1 and RFC 4493 (cited at tb_cmac.v:559, 605, 650, 699, 753, 810) and compares each tag with the hard-coded expected value
26
  Check: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_cmac.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "07 test completed. All test cases completed successfully." (printed only when error_ctr == 0) (tb_cmac.v:227) and fails on "test cases did not complete successfully" (tb_cmac.v:231). For a write-the-RTL task, hide src/rtl/*.v and keep tb_cmac.v.
27
  Companions (not included): src/rtl/cmac.v; src/rtl/cmac_core.v; toolruns/Makefile; LICENSE
28
  Dependencies: Icarus Verilog v12_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2025-04-02, GitHub release published 2023-06-11, the next one v13_0 came 2026-03-02; https://github.com/steveicarus/iverilog/releases/tag/v12_0); open source, not packed; secworks/aes at commit 80dc4718e1dcbbdb4b0dd1bdb393d8f7b98981dc checked out as sibling directory ../../aes relative to toolruns/; open source, not packed; GNU make (runs toolruns/Makefile; upstream does not pin a version; version pinned when the pool is built); open source, 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: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_cmac.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "07 test completed. All test cases completed successfully." (printed only when error_ctr == 0) (tb_cmac.v:227) and fails on "test cases did not complete successfully" (tb_cmac.v:231). For a write-the-RTL task, hide src/rtl/*.v and keep tb_cmac.v.
12
 
13
  **Licence (SPDX):** BSD-2-Clause
 
19
  - `src/tb/tb_cmac.v`: Secworks cmac: AES-CMAC message authentication core with memory-mapped top level, verified by the self-checking Verilog testbench tb_cmac.v.
20
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
21
  Origin: self-checking testbench by the core author, with expected outputs hard-coded in the testbench (published standard vectors where the file cites them)
 
 
22
  Check: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_cmac.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "07 test completed. All test cases completed successfully." (printed only when error_ctr == 0) (tb_cmac.v:227) and fails on "test cases did not complete successfully" (tb_cmac.v:231). For a write-the-RTL task, hide src/rtl/*.v and keep tb_cmac.v.
23
  Companions (not included): src/rtl/cmac.v; src/rtl/cmac_core.v; toolruns/Makefile; LICENSE
24
  Dependencies: Icarus Verilog v12_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2025-04-02, GitHub release published 2023-06-11, the next one v13_0 came 2026-03-02; https://github.com/steveicarus/iverilog/releases/tag/v12_0); open source, not packed; secworks/aes at commit 80dc4718e1dcbbdb4b0dd1bdb393d8f7b98981dc checked out as sibling directory ../../aes relative to toolruns/; open source, not packed; GNU make (runs toolruns/Makefile; upstream does not pin a version; version pinned when the pool is built); open source, not packed
campaign_files/2026-10-hw-eda-tests-2/secworks-gift/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: self_checking_testbench_known_vectors. 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 make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_gift.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 03 test cases completed successfully" (printed only when error_ctr == 0) (tb_gift.v:166) and fails on "test cases did not complete successfully" (tb_gift.v:170). For a write-the-RTL task, hide src/rtl/*.v and keep tb_gift.v.
14
 
15
  **Licence (SPDX):** BSD-2-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `src/tb/tb_gift.v`: Secworks gift: GIFT-128 block cipher core with memory-mapped top level, verified by the self-checking Verilog testbench tb_gift.v.
22
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
23
  Origin: self-checking testbench by the core author, with expected outputs hard-coded in the testbench (published standard vectors where the file cites them)
24
- Reference result: self_checking_testbench_known_vectors
25
- Reference source: tb_gift.v calls enctest 3 times (lines 336, 339, 343) with key, plaintext and expected ciphertext; the task (line 267) compares encres == expected (line 310) and increments error_ctr on a mismatch (line 315); the file does not name the source of the vectors
26
  Check: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_gift.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 03 test cases completed successfully" (printed only when error_ctr == 0) (tb_gift.v:166) and fails on "test cases did not complete successfully" (tb_gift.v:170). For a write-the-RTL task, hide src/rtl/*.v and keep tb_gift.v.
27
  Companions (not included): src/rtl/gift.v; src/rtl/gift_core.v; src/rtl/gift_api.vh; src/rtl/gift_round_functions.vh; toolruns/Makefile; LICENSE
28
  Dependencies: Icarus Verilog v11_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2021-08-31, GitHub release published 2020-10-11, the next one v12_0 came 2023-06-11; https://github.com/steveicarus/iverilog/releases/tag/v11_0); open source, not packed; GNU make (runs toolruns/Makefile; upstream does not pin a version; version pinned when the pool is built); open source, 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: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_gift.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 03 test cases completed successfully" (printed only when error_ctr == 0) (tb_gift.v:166) and fails on "test cases did not complete successfully" (tb_gift.v:170). For a write-the-RTL task, hide src/rtl/*.v and keep tb_gift.v.
12
 
13
  **Licence (SPDX):** BSD-2-Clause
 
19
  - `src/tb/tb_gift.v`: Secworks gift: GIFT-128 block cipher core with memory-mapped top level, verified by the self-checking Verilog testbench tb_gift.v.
20
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
21
  Origin: self-checking testbench by the core author, with expected outputs hard-coded in the testbench (published standard vectors where the file cites them)
 
 
22
  Check: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_gift.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 03 test cases completed successfully" (printed only when error_ctr == 0) (tb_gift.v:166) and fails on "test cases did not complete successfully" (tb_gift.v:170). For a write-the-RTL task, hide src/rtl/*.v and keep tb_gift.v.
23
  Companions (not included): src/rtl/gift.v; src/rtl/gift_core.v; src/rtl/gift_api.vh; src/rtl/gift_round_functions.vh; toolruns/Makefile; LICENSE
24
  Dependencies: Icarus Verilog v11_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2021-08-31, GitHub release published 2020-10-11, the next one v12_0 came 2023-06-11; https://github.com/steveicarus/iverilog/releases/tag/v11_0); open source, not packed; GNU make (runs toolruns/Makefile; upstream does not pin a version; version pinned when the pool is built); open source, not packed
campaign_files/2026-10-hw-eda-tests-2/secworks-md5/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: self_checking_testbench_known_vectors. 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 make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_md5.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 04 test cases completed successfully" (printed only when error_ctr == 0) (tb_md5.v:180) and fails on "test cases did not complete successfully" (tb_md5.v:184). For a write-the-RTL task, hide src/rtl/*.v and keep tb_md5.v.
14
 
15
  **Licence (SPDX):** BSD-2-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `src/tb/tb_md5.v`: Secworks md5: MD5 hash core with memory-mapped top level, verified by the self-checking Verilog testbench tb_md5.v.
22
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
23
  Origin: self-checking testbench by the core author, with expected outputs hard-coded in the testbench (published standard vectors where the file cites them)
24
- Reference result: self_checking_testbench_known_vectors
25
- Reference source: tb_md5.v hashes the empty string, "a", "The quick brown fox jumps over the lazy dog" and a dual-block string of 'a' characters (tasks tc1 to tc4 at lines 279-495) and compares the 128-bit digest with the hard-coded expected value
26
  Check: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_md5.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 04 test cases completed successfully" (printed only when error_ctr == 0) (tb_md5.v:180) and fails on "test cases did not complete successfully" (tb_md5.v:184). For a write-the-RTL task, hide src/rtl/*.v and keep tb_md5.v.
27
  Companions (not included): src/rtl/md5.v; src/rtl/md5_core.v; toolruns/Makefile; LICENSE
28
  Dependencies: Icarus Verilog v11_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2021-05-23, GitHub release published 2020-10-11, the next one v12_0 came 2023-06-11; https://github.com/steveicarus/iverilog/releases/tag/v11_0); open source, not packed; GNU make (runs toolruns/Makefile; upstream does not pin a version; version pinned when the pool is built); open source, 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: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_md5.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 04 test cases completed successfully" (printed only when error_ctr == 0) (tb_md5.v:180) and fails on "test cases did not complete successfully" (tb_md5.v:184). For a write-the-RTL task, hide src/rtl/*.v and keep tb_md5.v.
12
 
13
  **Licence (SPDX):** BSD-2-Clause
 
19
  - `src/tb/tb_md5.v`: Secworks md5: MD5 hash core with memory-mapped top level, verified by the self-checking Verilog testbench tb_md5.v.
20
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
21
  Origin: self-checking testbench by the core author, with expected outputs hard-coded in the testbench (published standard vectors where the file cites them)
 
 
22
  Check: Run make sim-top in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_md5.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 04 test cases completed successfully" (printed only when error_ctr == 0) (tb_md5.v:180) and fails on "test cases did not complete successfully" (tb_md5.v:184). For a write-the-RTL task, hide src/rtl/*.v and keep tb_md5.v.
23
  Companions (not included): src/rtl/md5.v; src/rtl/md5_core.v; toolruns/Makefile; LICENSE
24
  Dependencies: Icarus Verilog v11_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2021-05-23, GitHub release published 2020-10-11, the next one v12_0 came 2023-06-11; https://github.com/steveicarus/iverilog/releases/tag/v11_0); open source, not packed; GNU make (runs toolruns/Makefile; upstream does not pin a version; version pinned when the pool is built); open source, not packed
campaign_files/2026-10-hw-eda-tests-2/secworks-modexp/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: self_checking_testbench_known_vectors. 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 make sim-modexp in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_modexp.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 01 test cases completed successfully" (printed only when error_ctr == 0) (tb_modexp.v:293) and fails on "test cases did not complete successfully" (tb_modexp.v:297). For a write-the-RTL task, hide src/rtl/*.v and keep tb_modexp.v.
14
 
15
  **Licence (SPDX):** BSD-2-Clause AND BSD-3-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `src/tb/tb_modexp.v`: Secworks modexp: modular exponentiation core (Montgomery multiplication, for RSA) with memory-mapped top level, verified by the self-checking Verilog testbench tb_modexp.v.
22
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
23
  Origin: self-checking testbench by the core author, with expected outputs hard-coded in the testbench (published standard vectors where the file cites them)
24
- Reference result: self_checking_testbench_known_vectors
25
- Reference source: tb_modexp.v runs one active test, rob_enc_1024 (task at line 1182, called at line 1381 from the main block at line 1360): a 1024-bit modular exponentiation whose 32 result words are compared with hard-coded expected words (assertEquals calls up to line 1341; success checked at line 1343, error_ctr incremented at line 1346); the other tests in the file are commented out
26
  Check: Run make sim-modexp in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_modexp.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 01 test cases completed successfully" (printed only when error_ctr == 0) (tb_modexp.v:293) and fails on "test cases did not complete successfully" (tb_modexp.v:297). For a write-the-RTL task, hide src/rtl/*.v and keep tb_modexp.v.
27
  Companions (not included): src/rtl/modexp.v; src/rtl/modexp_core.v; src/rtl/montprod.v; src/rtl/residue.v; src/rtl/adder32.v; src/rtl/shl32.v; src/rtl/shr32.v; src/rtl/blockmem1r1w.v; src/rtl/blockmem2r1w.v; src/rtl/blockmem2r1wptr.v; src/rtl/blockmem2rptr1w.v; toolruns/Makefile; LICENSE
28
  Dependencies: Icarus Verilog v10_3 (iverilog and vvp; upstream does not pin it; v11_0 was published on GitHub 2020-10-11, after the unit commit date 2020-10-08, so the latest release before the commit is the tag v10_3 (tag date not read; tag list fetched from api.github.com/repos/steveicarus/iverilog/tags); https://github.com/steveicarus/iverilog/releases/tag/v10_3); open source, not packed; GNU make (runs toolruns/Makefile; upstream does not pin a version; version pinned when the pool is built); open source, 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: Run make sim-modexp in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_modexp.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 01 test cases completed successfully" (printed only when error_ctr == 0) (tb_modexp.v:293) and fails on "test cases did not complete successfully" (tb_modexp.v:297). For a write-the-RTL task, hide src/rtl/*.v and keep tb_modexp.v.
12
 
13
  **Licence (SPDX):** BSD-2-Clause AND BSD-3-Clause
 
19
  - `src/tb/tb_modexp.v`: Secworks modexp: modular exponentiation core (Montgomery multiplication, for RSA) with memory-mapped top level, verified by the self-checking Verilog testbench tb_modexp.v.
20
  Format: plain Verilog-2001 self-checking testbench (no cocotb) plus RTL sources and the toolruns Makefile that compiles and runs them
21
  Origin: self-checking testbench by the core author, with expected outputs hard-coded in the testbench (published standard vectors where the file cites them)
 
 
22
  Check: Run make sim-modexp in toolruns/ (iverilog -Wall builds the simulation from the RTL and tb_modexp.v, then vvp runs it) in a sandbox with the versions listed in dependencies; vvp exits 0 after $finish whatever the outcome, so the pass is read from the log: the unit passes on log line "*** All 01 test cases completed successfully" (printed only when error_ctr == 0) (tb_modexp.v:293) and fails on "test cases did not complete successfully" (tb_modexp.v:297). For a write-the-RTL task, hide src/rtl/*.v and keep tb_modexp.v.
23
  Companions (not included): src/rtl/modexp.v; src/rtl/modexp_core.v; src/rtl/montprod.v; src/rtl/residue.v; src/rtl/adder32.v; src/rtl/shl32.v; src/rtl/shr32.v; src/rtl/blockmem1r1w.v; src/rtl/blockmem2r1w.v; src/rtl/blockmem2r1wptr.v; src/rtl/blockmem2rptr1w.v; toolruns/Makefile; LICENSE
24
  Dependencies: Icarus Verilog v10_3 (iverilog and vvp; upstream does not pin it; v11_0 was published on GitHub 2020-10-11, after the unit commit date 2020-10-08, so the latest release before the commit is the tag v10_3 (tag date not read; tag list fetched from api.github.com/repos/steveicarus/iverilog/tags); https://github.com/steveicarus/iverilog/releases/tag/v10_3); open source, not packed; GNU make (runs toolruns/Makefile; upstream does not pin a version; version pinned when the pool is built); open source, not packed
campaign_files/2026-10-m-issue-trackers-2/rtcsnv-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: https://developer.rtcsnv.com/transitData/google_transit.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 Regional Transportation Commission of Southern Nevada (RTC) at https://developer.rtcsnv.com/transitData/google_transit.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: https://developer.rtcsnv.com/transitData/google_transit.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 Regional Transportation Commission of Southern Nevada (RTC) at https://developer.rtcsnv.com/transitData/google_transit.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/rtems-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
  - `m329-01/2020-September.txt.gz`: How to use lvgl on pc386 BSP. (9 messages, 2020-09-07 to 2020-09-09).
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 'RTEMS 4.10 i386 undefined reference to `rtems_bdbuf_configuration'' (8 messages, 3 people, root <FD2713B2-258D-461E-A612-8C848528319D@fhi-berlin.mpg.de>); message count, Message-ID set, last message time and first-reply seconds 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
  - `m329-01/2020-September.txt.gz`: How to use lvgl on pc386 BSP. (9 messages, 2020-09-07 to 2020-09-09).
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 'RTEMS 4.10 i386 undefined reference to `rtems_bdbuf_configuration'' (8 messages, 3 people, root <FD2713B2-258D-461E-A612-8C848528319D@fhi-berlin.mpg.de>); message count, Message-ID set, last message time and first-reply seconds 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/ruby-redmine/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained. 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: Facts computed from the ticket's own timeline (status transitions, assignee changes, 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 Redmine REST API (/issues/N.json?include=journals, /issue_statuses.json) and read it with jq or Python json; python-redmine also works.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,16 +19,12 @@ Files:
21
  - `issues/15000.json`: Ruby Redmine Feature #15000: Prevent to initialize MonitorMixin twice (Closed).
22
  Format: Redmine REST API issue JSON with include=journals: issue fields plus journals[] (user, created_on, notes, details[] with property, name, old_value, new_value); companion issue_statuses.json maps status ids to names.
23
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: copies fetched apart were byte-identical. Unit files kept only outside the repository for checking and then deleted.
24
- Reference result: facts computed from the ticket's own timeline (status transitions, assignee changes, comments)
25
- Reference source: The pinned response(s): issues/15000.json, issue_statuses.json.
26
  Check: Exact match against values read from the pinned JSON: journals[] (user.name, created_on, notes, details[] with name 'status_id' or 'assigned_to_id', old_value, new_value); status ids are mapped to names through the companion issue_statuses.json; assignees in journal details are user ids (names only where the issue's author or assigned_to field gives them). Deterministic, no model judgement. No bot accounts seen. Timeline events = journals. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
27
  Tools: curl the Redmine REST API (/issues/N.json?include=journals, /issue_statuses.json) and read it with jq or Python json; python-redmine also works.
28
  Companions (not included): issue_statuses.json
29
  - `issues/16500.json`: Ruby Redmine Bug #16500: Argument is added to both splat and last &block argument (Closed).
30
  Format: Redmine REST API issue JSON with include=journals: issue fields plus journals[] (user, created_on, notes, details[] with property, name, old_value, new_value); companion issue_statuses.json maps status ids to names.
31
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: copies fetched apart were byte-identical. Unit files kept only outside the repository for checking and then deleted.
32
- Reference result: facts computed from the ticket's own timeline (status transitions, assignee changes, comments)
33
- Reference source: The pinned response(s): issues/16500.json, issue_statuses.json.
34
  Check: Exact match against values read from the pinned JSON: journals[] (user.name, created_on, notes, details[] with name 'status_id' or 'assigned_to_id', old_value, new_value); status ids are mapped to names through the companion issue_statuses.json; assignees in journal details are user ids (names only where the issue's author or assigned_to field gives them). Deterministic, no model judgement. No bot accounts seen. Timeline events = journals. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
35
  Tools: curl the Redmine REST API (/issues/N.json?include=journals, /issue_statuses.json) and read it with jq or Python json; python-redmine also works.
36
  Companions (not included): issue_statuses.json
 
8
 
9
  **Self-contained:** Each file is self-contained. 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:** curl the Redmine REST API (/issues/N.json?include=journals, /issue_statuses.json) and read it with jq or Python json; python-redmine also works.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `issues/15000.json`: Ruby Redmine Feature #15000: Prevent to initialize MonitorMixin twice (Closed).
20
  Format: Redmine REST API issue JSON with include=journals: issue fields plus journals[] (user, created_on, notes, details[] with property, name, old_value, new_value); companion issue_statuses.json maps status ids to names.
21
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: copies fetched apart were byte-identical. Unit files kept only outside the repository for checking and then deleted.
 
 
22
  Check: Exact match against values read from the pinned JSON: journals[] (user.name, created_on, notes, details[] with name 'status_id' or 'assigned_to_id', old_value, new_value); status ids are mapped to names through the companion issue_statuses.json; assignees in journal details are user ids (names only where the issue's author or assigned_to field gives them). Deterministic, no model judgement. No bot accounts seen. Timeline events = journals. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
23
  Tools: curl the Redmine REST API (/issues/N.json?include=journals, /issue_statuses.json) and read it with jq or Python json; python-redmine also works.
24
  Companions (not included): issue_statuses.json
25
  - `issues/16500.json`: Ruby Redmine Bug #16500: Argument is added to both splat and last &block argument (Closed).
26
  Format: Redmine REST API issue JSON with include=journals: issue fields plus journals[] (user, created_on, notes, details[] with property, name, old_value, new_value); companion issue_statuses.json maps status ids to names.
27
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: copies fetched apart were byte-identical. Unit files kept only outside the repository for checking and then deleted.
 
 
28
  Check: Exact match against values read from the pinned JSON: journals[] (user.name, created_on, notes, details[] with name 'status_id' or 'assigned_to_id', old_value, new_value); status ids are mapped to names through the companion issue_statuses.json; assignees in journal details are user ids (names only where the issue's author or assigned_to field gives them). Deterministic, no model judgement. No bot accounts seen. Timeline events = journals. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
29
  Tools: curl the Redmine REST API (/issues/N.json?include=journals, /issue_statuses.json) and read it with jq or Python json; python-redmine also works.
30
  Companions (not included): issue_statuses.json
campaign_files/2026-10-m-issue-trackers-2/ruby-talk/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:** Fetch or open the thread JSON and process it with jq or Python's json module (group by sender.mailman_id, sort by date in UTC, follow parent links); the month export opens as mbox in mutt -f or Python's mailbox module.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m259-11/thread_O4I562JD35ECVDIK2E76NPCGIKANJJML_emails.json`: [ruby-talk:444294] Alternative logging solutions (11 messages, 2023-08-04 to 2023-08-04).
22
  Format: HyperKitty REST API JSON (Mailman 3): list of the thread's messages, each with message_id, message_id_hash, subject, sender (address obfuscated, mailman_id), sender_name, date (ISO 8601 with zone), parent (URL of the parent message or null for the root) and children. The thread id is the base32 SHA-1 of the root Message-ID.
23
  Origin: Real list traffic in a closed past month of the list's archive. The unit is the HyperKitty thread JSON of one thread; HyperKitty assigns each message to the thread, so the JSON already holds the whole thread.
24
- Reference result: facts computed from the thread headers
25
- Reference source: message_id, sender, sender_name, date and parent fields of the entries in the pinned thread JSON.
26
  Check: Exact match against values computed from the pinned thread JSON: one entry per message; the root is the entry with parent null; direct replies are entries whose parent is the root's URL; dates are converted to UTC from their ISO 8601 offsets; people are counted by sender.mailman_id (one Mailman sender record per address). The first-reply interval is the exact difference in whole seconds between the root date and the earliest reply date; the dates carry one-second resolution, so no rounding is involved. Deterministic. Hand check on 2 threads of this host (general rule 5): the chosen thread and '[ruby-talk:444309] Rails Already Supports View Components!' (root <CAAZkKDL6jFxKSNxc+H2=ihycgoryDC6VLDi++_XWjQML0OJ1eg@mail.gmail.com>); for each, the thread JSON was compared with the same thread rebuilt from the month mbox export by two independent parsers (Python mailbox module, and a regular-expression parser); message count, people, last message time and first-reply seconds agreed on the chosen thread.
27
  Tools: Fetch or open the thread JSON and process it with jq or Python's json module (group by sender.mailman_id, sort by date in UTC, follow parent links); the month export opens as mbox in mutt -f or Python's mailbox module.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Fetch or open the thread JSON and process it with jq or Python's json module (group by sender.mailman_id, sort by date in UTC, follow parent links); the month export opens as mbox in mutt -f or Python's mailbox module.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m259-11/thread_O4I562JD35ECVDIK2E76NPCGIKANJJML_emails.json`: [ruby-talk:444294] Alternative logging solutions (11 messages, 2023-08-04 to 2023-08-04).
20
  Format: HyperKitty REST API JSON (Mailman 3): list of the thread's messages, each with message_id, message_id_hash, subject, sender (address obfuscated, mailman_id), sender_name, date (ISO 8601 with zone), parent (URL of the parent message or null for the root) and children. The thread id is the base32 SHA-1 of the root Message-ID.
21
  Origin: Real list traffic in a closed past month of the list's archive. The unit is the HyperKitty thread JSON of one thread; HyperKitty assigns each message to the thread, so the JSON already holds the whole thread.
 
 
22
  Check: Exact match against values computed from the pinned thread JSON: one entry per message; the root is the entry with parent null; direct replies are entries whose parent is the root's URL; dates are converted to UTC from their ISO 8601 offsets; people are counted by sender.mailman_id (one Mailman sender record per address). The first-reply interval is the exact difference in whole seconds between the root date and the earliest reply date; the dates carry one-second resolution, so no rounding is involved. Deterministic. Hand check on 2 threads of this host (general rule 5): the chosen thread and '[ruby-talk:444309] Rails Already Supports View Components!' (root <CAAZkKDL6jFxKSNxc+H2=ihycgoryDC6VLDi++_XWjQML0OJ1eg@mail.gmail.com>); for each, the thread JSON was compared with the same thread rebuilt from the month mbox export by two independent parsers (Python mailbox module, and a regular-expression parser); message count, people, last message time and first-reply seconds agreed on the chosen thread.
23
  Tools: Fetch or open the thread JSON and process it with jq or Python's json module (group by sender.mailman_id, sort by date in UTC, follow parent links); the month export opens as mbox in mutt -f or Python's mailbox module.
campaign_files/2026-10-m-issue-trackers-2/rust-github-issues/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: Facts computed from the issue's own timeline (status changes, assignments, comments and their authors). 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 -H 'Accept: application/vnd.github+json' https://api.github.com/repos/rust-lang/rust/issues/83318/events?per_page=100 and .../issues/83318/comments?per_page=100, merge by created_at with jq or Python json; 'gh api' reaches the same endpoints
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,8 +19,6 @@ Files:
21
  - `repos/rust-lang/rust/issues/83318/events.json`: Rust-lang/rust issue #83318: Migration for or patterns.
22
  Format: Two GitHub REST API JSON files fetched from the same issue: events.json (GET /repos/{owner}/{repo}/issues/{number}/events, per_page=100: actor, event, created_at, assignee/assigner, requested_reviewer/review_requester, label, commit_id; 27 entries) and the companion comments.json (GET /repos/{owner}/{repo}/issues/{number}/comments, per_page=100: user, created_at, updated_at, author_association, body; 12 entries). Each is one page with fewer than 100 entries, so both are complete. Together they hold the issue timeline without 'cross-referenced' entries, which embed live repository counters and are not byte-stable (see report).
23
  Origin: Live GitHub REST interface of a closed issue, collected as a NOASSERTION pointer row under spec general rule 4: the events and comments JSON were each fetched twice apart and were byte-identical. Unit files kept only outside the repository for checking and then deleted.
24
- Reference result: facts computed from the issue's own timeline (status changes, assignments, comments and their authors)
25
- Reference source: The pinned responses: repos/rust-lang/rust/issues/83318/events.json and its companion repos/rust-lang/rust/issues/83318/comments.json.
26
  Check: Exact match against values read from the pinned JSON: closed/reopened/merged entries (actor, created_at), assigned/unassigned/review_requested entries (assignee or requested_reviewer, assigner or review_requester, created_at), and comments (author login, created_at). A human comment is a 'commented' entry whose author is not a bot. Bot accounts excluded from the comment count: accounts with type 'Bot', logins ending in '[bot]', and the named automation accounts azure-pipelines, bors, codecov, dotnet-bot, dotnet-issue-labeler, dotnet-maestro, dotnet-policy-service, facebook-github-bot, flutter-dashboard, fluttergithubbot, google-cla, grafanabot, homeassistant, k8s-ci-robot, k8s-triage-robot, linux-foundation-easycla, msftbot, netlify, probot-home-assistant, pytorch-bot, pytorch-probot, rust-highfive, rustbot, typescript-bot, vs-code-engineering, vscode-triage-bot, vscodenpm. Bot accounts seen in this record: bors, rustbot. 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 -H 'Accept: application/vnd.github+json' https://api.github.com/repos/rust-lang/rust/issues/83318/events?per_page=100 and .../issues/83318/comments?per_page=100, merge by created_at with jq or Python json; 'gh api' reaches the same endpoints
28
  Companions (not included): repos/rust-lang/rust/issues/83318/comments.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.
10
 
 
 
11
  **Tools to open and check:** curl -H 'Accept: application/vnd.github+json' https://api.github.com/repos/rust-lang/rust/issues/83318/events?per_page=100 and .../issues/83318/comments?per_page=100, merge by created_at with jq or Python json; 'gh api' reaches the same endpoints
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `repos/rust-lang/rust/issues/83318/events.json`: Rust-lang/rust issue #83318: Migration for or patterns.
20
  Format: Two GitHub REST API JSON files fetched from the same issue: events.json (GET /repos/{owner}/{repo}/issues/{number}/events, per_page=100: actor, event, created_at, assignee/assigner, requested_reviewer/review_requester, label, commit_id; 27 entries) and the companion comments.json (GET /repos/{owner}/{repo}/issues/{number}/comments, per_page=100: user, created_at, updated_at, author_association, body; 12 entries). Each is one page with fewer than 100 entries, so both are complete. Together they hold the issue timeline without 'cross-referenced' entries, which embed live repository counters and are not byte-stable (see report).
21
  Origin: Live GitHub REST interface of a closed issue, collected as a NOASSERTION pointer row under spec general rule 4: the events and comments JSON were each fetched twice apart and were byte-identical. Unit files kept only outside the repository for checking and then deleted.
 
 
22
  Check: Exact match against values read from the pinned JSON: closed/reopened/merged entries (actor, created_at), assigned/unassigned/review_requested entries (assignee or requested_reviewer, assigner or review_requester, created_at), and comments (author login, created_at). A human comment is a 'commented' entry whose author is not a bot. Bot accounts excluded from the comment count: accounts with type 'Bot', logins ending in '[bot]', and the named automation accounts azure-pipelines, bors, codecov, dotnet-bot, dotnet-issue-labeler, dotnet-maestro, dotnet-policy-service, facebook-github-bot, flutter-dashboard, fluttergithubbot, google-cla, grafanabot, homeassistant, k8s-ci-robot, k8s-triage-robot, linux-foundation-easycla, msftbot, netlify, probot-home-assistant, pytorch-bot, pytorch-probot, rust-highfive, rustbot, typescript-bot, vs-code-engineering, vscode-triage-bot, vscodenpm. Bot accounts seen in this record: bors, rustbot. 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 -H 'Accept: application/vnd.github+json' https://api.github.com/repos/rust-lang/rust/issues/83318/events?per_page=100 and .../issues/83318/comments?per_page=100, merge by created_at with jq or Python json; 'gh api' reaches the same endpoints
24
  Companions (not included): repos/rust-lang/rust/issues/83318/comments.json
campaign_files/2026-10-m-issue-trackers-2/salisburyreds-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/salisburyreds/gtfs/salisburyreds_1790764726.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
24
  Origin: Live static GTFS feed published by Salisbury Reds (published through Passenger, data.discoverpassenger.com) at https://s3-eu-west-1.amazonaws.com/passenger-sources/salisburyreds/gtfs/salisburyreds_1790764726.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/salisburyreds/gtfs/salisburyreds_1790764726.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
22
  Origin: Live static GTFS feed published by Salisbury Reds (published through Passenger, data.discoverpassenger.com) at https://s3-eu-west-1.amazonaws.com/passenger-sources/salisburyreds/gtfs/salisburyreds_1790764726.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/samba-samba/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
  - `m209-06/2019-March.txt.gz`: [Samba] Backup of shared data-files (8 messages, 2019-03-07 to 2019-03-14).
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 '[Samba] How to automatically store the macAddress in AD' (9 messages, root <CABC5mi5vVi9HtpfdoSvm9euiXfSHon=Cmc0i8RKG918S+TAq-w@mail.gmail.com>); 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
  - `m209-06/2019-March.txt.gz`: [Samba] Backup of shared data-files (8 messages, 2019-03-07 to 2019-03-14).
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 '[Samba] How to automatically store the macAddress in AD' (9 messages, root <CABC5mi5vVi9HtpfdoSvm9euiXfSHon=Cmc0i8RKG918S+TAq-w@mail.gmail.com>); 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/sanderscoaches-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/sanderscoaches/gtfs/sanderscoaches_1789675114.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
24
  Origin: Live static GTFS feed published by Sanders Coaches (published through Passenger, data.discoverpassenger.com) at https://s3-eu-west-1.amazonaws.com/passenger-sources/sanderscoaches/gtfs/sanderscoaches_1789675114.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/sanderscoaches/gtfs/sanderscoaches_1789675114.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
22
  Origin: Live static GTFS feed published by Sanders Coaches (published through Passenger, data.discoverpassenger.com) at https://s3-eu-west-1.amazonaws.com/passenger-sources/sanderscoaches/gtfs/sanderscoaches_1789675114.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/saskatoon-transit-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-Saskatoon-Open-Data
@@ -22,8 +20,6 @@ Units:
22
  Read out of the archive: https://apps5.saskatoon.ca/data/gtfs/google_transit.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
24
  Origin: Live static GTFS feed published by City of Saskatoon (Saskatoon Transit) at https://apps5.saskatoon.ca/data/gtfs/google_transit.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):** LicenseRef-Saskatoon-Open-Data
 
20
  Read out of the archive: https://apps5.saskatoon.ca/data/gtfs/google_transit.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
22
  Origin: Live static GTFS feed published by City of Saskatoon (Saskatoon Transit) at https://apps5.saskatoon.ca/data/gtfs/google_transit.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/sbmtd-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://sbmtd.gov/google_transit/feed.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 Santa Barbara Metropolitan Transit District (MTD) at http://sbmtd.gov/google_transit/feed.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://sbmtd.gov/google_transit/feed.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 Santa Barbara Metropolitan Transit District (MTD) at http://sbmtd.gov/google_transit/feed.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/ronn-ng-definition-list-syntax/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 fragment and roff man page. 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:** ronn command line: ronn --pipe --html --fragment page.ronn and ronn --pipe --roff --date=1979-01-01 page.ronn; man -l to view the roff
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/definition_list_syntax.ronn`: Ronn-ng test "definition_list_syntax": a ronn Markdown man page converted to an HTML fragment and to roff.
22
  Format: ronn-format Markdown file with expected outputs definition_list_syntax.html and definition_list_syntax.roff
23
  Origin: test page written by the ronn maintainers for the definition list syntax of ronn-format Markdown
24
- Reference result: expected HTML fragment and roff man page
25
- Reference source: definition_list_syntax.html and definition_list_syntax.roff
26
  Check: show the agent the .ronn file (hide the .html and .roff); HTML: the agent's fragment must equal the .html file exactly; roff: render with date 1979-01-01, drop the first two lines (the ronn version comments) and the rest must equal the .roff file exactly
27
  Tools: ronn command line: ronn --pipe --html --fragment page.ronn and ronn --pipe --roff --date=1979-01-01 page.ronn; man -l to view the roff
28
  Companions (not included): test/definition_list_syntax.html; test/definition_list_syntax.roff; test/test_ronn.rb; test/contest.rb
 
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:** ronn command line: ronn --pipe --html --fragment page.ronn and ronn --pipe --roff --date=1979-01-01 page.ronn; man -l to view the roff
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/definition_list_syntax.ronn`: Ronn-ng test "definition_list_syntax": a ronn Markdown man page converted to an HTML fragment and to roff.
20
  Format: ronn-format Markdown file with expected outputs definition_list_syntax.html and definition_list_syntax.roff
21
  Origin: test page written by the ronn maintainers for the definition list syntax of ronn-format Markdown
 
 
22
  Check: show the agent the .ronn file (hide the .html and .roff); HTML: the agent's fragment must equal the .html file exactly; roff: render with date 1979-01-01, drop the first two lines (the ronn version comments) and the rest must equal the .roff file exactly
23
  Tools: ronn command line: ronn --pipe --html --fragment page.ronn and ronn --pipe --roff --date=1979-01-01 page.ronn; man -l to view the roff
24
  Companions (not included): test/definition_list_syntax.html; test/definition_list_syntax.roff; test/test_ronn.rb; test/contest.rb
campaign_files/2026-10-md-media-and-documents/rst-to-myst-fixtures-render/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 MyST Markdown for each reStructuredText example. 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:** rst2myst command line (rst2myst convert page.rst) or the Python API rst_to_myst(text); pandoc -f rst -t markdown gives a comparable conversion
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `tests/fixtures/render.txt`: Rst-to-myst render fixtures: reStructuredText snippets converted to MyST Markdown.
22
  Format: pytest-param-files fixture text file: a title line, ".", the reStructuredText input, ".", the expected MyST Markdown, "." per example (102 "." lines, 34 examples)
23
  Origin: fixture file written by the rst-to-myst maintainers; the front-matter case (about lines 355 to 380) quotes a docinfo block naming the docutils authors, taken from the public docutils documentation
24
- Reference result: expected MyST Markdown for each reStructuredText example
25
- Reference source: the part between the second and third "." of each example
26
  Check: show only the reStructuredText of each example (hide the expected MyST); the agent writes MyST Markdown; apply rstrip to both the agent's text and the expected text and pass when they are equal; the upstream also requires an empty warning stream except for examples whose title holds EXPECT_WARNING, which the pool may drop or keep as a second check
27
  Tools: rst2myst command line (rst2myst convert page.rst) or the Python API rst_to_myst(text); pandoc -f rst -t markdown gives a comparable conversion
28
  Companions (not included): tests/test_fixtures.py
 
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:** rst2myst command line (rst2myst convert page.rst) or the Python API rst_to_myst(text); pandoc -f rst -t markdown gives a comparable conversion
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `tests/fixtures/render.txt`: Rst-to-myst render fixtures: reStructuredText snippets converted to MyST Markdown.
20
  Format: pytest-param-files fixture text file: a title line, ".", the reStructuredText input, ".", the expected MyST Markdown, "." per example (102 "." lines, 34 examples)
21
  Origin: fixture file written by the rst-to-myst maintainers; the front-matter case (about lines 355 to 380) quotes a docinfo block naming the docutils authors, taken from the public docutils documentation
 
 
22
  Check: show only the reStructuredText of each example (hide the expected MyST); the agent writes MyST Markdown; apply rstrip to both the agent's text and the expected text and pass when they are equal; the upstream also requires an empty warning stream except for examples whose title holds EXPECT_WARNING, which the pool may drop or keep as a second check
23
  Tools: rst2myst command line (rst2myst convert page.rst) or the Python API rst_to_myst(text); pandoc -f rst -t markdown gives a comparable conversion
24
  Companions (not included): tests/test_fixtures.py
campaign_files/2026-10-md-media-and-documents/rst2pdf-rst2pdf-tests-input-and-tests-reference/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: Reference rendering (PDF) compared by text blocks, coordinates and 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:** pytest tests -k test_issue_114 (rst2pdf renders the .rst headless with ReportLab; PyMuPDF compares with the reference PDF)
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `tests/input/test_issue_114.rst`: rst2pdf test test_issue_114: section titles with ampersands in the PDF outline, with reference PDF.
22
  Format: reStructuredText (.rst) plus reference PDF
23
  Origin: regression test for issue 114 written by the rst2pdf maintainers; the reference PDF is the accepted rendering
24
- Reference result: reference rendering (PDF) compared by text blocks, coordinates and metadata
25
- Reference source: tests/reference/test_issue_114.pdf, the rendering the maintainers accepted
26
  Check: render the .rst with rst2pdf, open both PDFs with PyMuPDF, compare metadata without dates and producer, the page count, the number of text blocks per page, each block bounding box within the coordinate threshold, the words of each block, and block type and number exactly; hide the reference PDF when posing the task
27
  Tools: pytest tests -k test_issue_114 (rst2pdf renders the .rst headless with ReportLab; PyMuPDF compares with the reference PDF)
28
  Companions (not included): tests/reference/test_issue_114.pdf
 
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:** pytest tests -k test_issue_114 (rst2pdf renders the .rst headless with ReportLab; PyMuPDF compares with the reference PDF)
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `tests/input/test_issue_114.rst`: rst2pdf test test_issue_114: section titles with ampersands in the PDF outline, with reference PDF.
20
  Format: reStructuredText (.rst) plus reference PDF
21
  Origin: regression test for issue 114 written by the rst2pdf maintainers; the reference PDF is the accepted rendering
 
 
22
  Check: render the .rst with rst2pdf, open both PDFs with PyMuPDF, compare metadata without dates and producer, the page count, the number of text blocks per page, each block bounding box within the coordinate threshold, the words of each block, and block type and number exactly; hide the reference PDF when posing the task
23
  Tools: pytest tests -k test_issue_114 (rst2pdf renders the .rst headless with ReportLab; PyMuPDF compares with the reference PDF)
24
  Companions (not included): tests/reference/test_issue_114.pdf
campaign_files/2026-10-md-media-and-documents/s3-symbolic-symphony-set/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained. 4 of 4 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: chord_segment_table. 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:** music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
14
 
15
  **Licence (SPDX):** MIT AND CC0-1.0
@@ -21,8 +19,6 @@ Files:
21
  - `symbolic_symphony_set/mo1/sheet.xml`: Roman-numeral harmony, cadence and phrase annotation of Mozart Symphony No. 41 in C major, K. 551 ("Jupiter"), first movement (Allegro vivace): full orchestral MusicXML score.
22
  Format: 4 files, 7,980,970 bytes: sheet.xml (MusicXML 4.0 full orchestral score, 17 parts, exported by MuseScore 4.3.2), harmony.csv (onset, offset, key, degree, quality, inversion, roman number), cadence.csv (onset, resolve time, offset, cadence type), form_analysis.csv (phrase onset, phrase offset, section, theme)
23
  Origin: author annotation (method not stated in the README) of a public-domain symphony on a CC0 OpenScore edition
24
- Reference result: chord_segment_table
25
- Reference source: harmony.csv written by the S3 authors; the README does not say whether the chord labels were made by hand or checked after an automatic pass, so the annotation method is unverified (see the report); cadence.csv and form_analysis.csv come from the same annotation
26
  Check: give sheet.xml; hide harmony.csv, cadence.csv and form_analysis.csv; the agent outputs chord segments (onset, offset in crotchet beats from the start of the written score, key, degree, quality, inversion, Roman numeral); score by time-weighted overlap per field, each field reported separately; cadence.csv (onset, type AC, HC, PC, DC) and form_analysis.csv (phrase spans, section, theme) serve as separate cadence and phrase tasks scored by exact onset and label match; no pass line is fixed by the source (see the report)
27
  Tools: music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
28
  Companions (not included): symbolic_symphony_set/README.md (column definitions and Roman numeral notation of harmony.csv)
@@ -30,8 +26,6 @@ Files:
30
  - `symbolic_symphony_set/mo2/sheet.xml`: Roman-numeral harmony, cadence and phrase annotation of Mozart Symphony No. 41 in C major, K. 551 ("Jupiter"), second movement (Andante cantabile): full orchestral MusicXML score.
31
  Format: 4 files, 3,897,678 bytes: sheet.xml (MusicXML 4.0 full orchestral score, 17 parts, exported by MuseScore 4.3.2), harmony.csv (onset, offset, key, degree, quality, inversion, roman number), cadence.csv (onset, resolve time, offset, cadence type), form_analysis.csv (phrase onset, phrase offset, section, theme)
32
  Origin: author annotation (method not stated in the README) of a public-domain symphony on a CC0 OpenScore edition
33
- Reference result: chord_segment_table
34
- Reference source: harmony.csv written by the S3 authors; the README does not say whether the chord labels were made by hand or checked after an automatic pass, so the annotation method is unverified (see the report); cadence.csv and form_analysis.csv come from the same annotation
35
  Check: give sheet.xml; hide harmony.csv, cadence.csv and form_analysis.csv; the agent outputs chord segments (onset, offset in crotchet beats from the start of the written score, key, degree, quality, inversion, Roman numeral); score by time-weighted overlap per field, each field reported separately; cadence.csv (onset, type AC, HC, PC, DC) and form_analysis.csv (phrase spans, section, theme) serve as separate cadence and phrase tasks scored by exact onset and label match; no pass line is fixed by the source (see the report)
36
  Tools: music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
37
  Companions (not included): symbolic_symphony_set/README.md (column definitions and Roman numeral notation of harmony.csv)
@@ -39,8 +33,6 @@ Files:
39
  - `symbolic_symphony_set/mo3/sheet.xml`: Roman-numeral harmony, cadence and phrase annotation of Mozart Symphony No. 41 in C major, K. 551 ("Jupiter"), third movement (Menuetto: Allegretto, with Trio): full orchestral MusicXML score.
40
  Format: 4 files, 2,079,490 bytes: sheet.xml (MusicXML 4.0 full orchestral score, 17 parts, exported by MuseScore 4.3.2), harmony.csv (onset, offset, key, degree, quality, inversion, roman number), cadence.csv (onset, resolve time, offset, cadence type), form_analysis.csv (phrase onset, phrase offset, section, theme)
41
  Origin: author annotation (method not stated in the README) of a public-domain symphony on a CC0 OpenScore edition
42
- Reference result: chord_segment_table
43
- Reference source: harmony.csv written by the S3 authors; the README does not say whether the chord labels were made by hand or checked after an automatic pass, so the annotation method is unverified (see the report); cadence.csv and form_analysis.csv come from the same annotation
44
  Check: give sheet.xml; hide harmony.csv, cadence.csv and form_analysis.csv; the agent outputs chord segments (onset, offset in crotchet beats from the start of the written score, key, degree, quality, inversion, Roman numeral); score by time-weighted overlap per field, each field reported separately; cadence.csv (onset, type AC, HC, PC, DC) and form_analysis.csv (phrase spans, section, theme) serve as separate cadence and phrase tasks scored by exact onset and label match; no pass line is fixed by the source (see the report)
45
  Tools: music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
46
  Companions (not included): symbolic_symphony_set/README.md (column definitions and Roman numeral notation of harmony.csv)
@@ -48,8 +40,6 @@ Files:
48
  - `symbolic_symphony_set/mo4/sheet.xml`: Roman-numeral harmony, cadence and phrase annotation of Mozart Symphony No. 41 in C major, K. 551 ("Jupiter"), fourth movement (Molto allegro): full orchestral MusicXML score.
49
  Format: 4 files, 10,316,445 bytes: sheet.xml (MusicXML 4.0 full orchestral score, 17 parts, exported by MuseScore 4.3.2), harmony.csv (onset, offset, key, degree, quality, inversion, roman number), cadence.csv (onset, resolve time, offset, cadence type), form_analysis.csv (phrase onset, phrase offset, section, theme)
50
  Origin: author annotation (method not stated in the README) of a public-domain symphony on a CC0 OpenScore edition
51
- Reference result: chord_segment_table
52
- Reference source: harmony.csv written by the S3 authors; the README does not say whether the chord labels were made by hand or checked after an automatic pass, so the annotation method is unverified (see the report); cadence.csv and form_analysis.csv come from the same annotation
53
  Check: give sheet.xml; hide harmony.csv, cadence.csv and form_analysis.csv; the agent outputs chord segments (onset, offset in crotchet beats from the start of the written score, key, degree, quality, inversion, Roman numeral); score by time-weighted overlap per field, each field reported separately; cadence.csv (onset, type AC, HC, PC, DC) and form_analysis.csv (phrase spans, section, theme) serve as separate cadence and phrase tasks scored by exact onset and label match; no pass line is fixed by the source (see the report)
54
  Tools: music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
55
  Companions (not included): symbolic_symphony_set/README.md (column definitions and Roman numeral notation of harmony.csv)
 
8
 
9
  **Self-contained:** Each file is self-contained. 4 of 4 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:** music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
12
 
13
  **Licence (SPDX):** MIT AND CC0-1.0
 
19
  - `symbolic_symphony_set/mo1/sheet.xml`: Roman-numeral harmony, cadence and phrase annotation of Mozart Symphony No. 41 in C major, K. 551 ("Jupiter"), first movement (Allegro vivace): full orchestral MusicXML score.
20
  Format: 4 files, 7,980,970 bytes: sheet.xml (MusicXML 4.0 full orchestral score, 17 parts, exported by MuseScore 4.3.2), harmony.csv (onset, offset, key, degree, quality, inversion, roman number), cadence.csv (onset, resolve time, offset, cadence type), form_analysis.csv (phrase onset, phrase offset, section, theme)
21
  Origin: author annotation (method not stated in the README) of a public-domain symphony on a CC0 OpenScore edition
 
 
22
  Check: give sheet.xml; hide harmony.csv, cadence.csv and form_analysis.csv; the agent outputs chord segments (onset, offset in crotchet beats from the start of the written score, key, degree, quality, inversion, Roman numeral); score by time-weighted overlap per field, each field reported separately; cadence.csv (onset, type AC, HC, PC, DC) and form_analysis.csv (phrase spans, section, theme) serve as separate cadence and phrase tasks scored by exact onset and label match; no pass line is fixed by the source (see the report)
23
  Tools: music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
24
  Companions (not included): symbolic_symphony_set/README.md (column definitions and Roman numeral notation of harmony.csv)
 
26
  - `symbolic_symphony_set/mo2/sheet.xml`: Roman-numeral harmony, cadence and phrase annotation of Mozart Symphony No. 41 in C major, K. 551 ("Jupiter"), second movement (Andante cantabile): full orchestral MusicXML score.
27
  Format: 4 files, 3,897,678 bytes: sheet.xml (MusicXML 4.0 full orchestral score, 17 parts, exported by MuseScore 4.3.2), harmony.csv (onset, offset, key, degree, quality, inversion, roman number), cadence.csv (onset, resolve time, offset, cadence type), form_analysis.csv (phrase onset, phrase offset, section, theme)
28
  Origin: author annotation (method not stated in the README) of a public-domain symphony on a CC0 OpenScore edition
 
 
29
  Check: give sheet.xml; hide harmony.csv, cadence.csv and form_analysis.csv; the agent outputs chord segments (onset, offset in crotchet beats from the start of the written score, key, degree, quality, inversion, Roman numeral); score by time-weighted overlap per field, each field reported separately; cadence.csv (onset, type AC, HC, PC, DC) and form_analysis.csv (phrase spans, section, theme) serve as separate cadence and phrase tasks scored by exact onset and label match; no pass line is fixed by the source (see the report)
30
  Tools: music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
31
  Companions (not included): symbolic_symphony_set/README.md (column definitions and Roman numeral notation of harmony.csv)
 
33
  - `symbolic_symphony_set/mo3/sheet.xml`: Roman-numeral harmony, cadence and phrase annotation of Mozart Symphony No. 41 in C major, K. 551 ("Jupiter"), third movement (Menuetto: Allegretto, with Trio): full orchestral MusicXML score.
34
  Format: 4 files, 2,079,490 bytes: sheet.xml (MusicXML 4.0 full orchestral score, 17 parts, exported by MuseScore 4.3.2), harmony.csv (onset, offset, key, degree, quality, inversion, roman number), cadence.csv (onset, resolve time, offset, cadence type), form_analysis.csv (phrase onset, phrase offset, section, theme)
35
  Origin: author annotation (method not stated in the README) of a public-domain symphony on a CC0 OpenScore edition
 
 
36
  Check: give sheet.xml; hide harmony.csv, cadence.csv and form_analysis.csv; the agent outputs chord segments (onset, offset in crotchet beats from the start of the written score, key, degree, quality, inversion, Roman numeral); score by time-weighted overlap per field, each field reported separately; cadence.csv (onset, type AC, HC, PC, DC) and form_analysis.csv (phrase spans, section, theme) serve as separate cadence and phrase tasks scored by exact onset and label match; no pass line is fixed by the source (see the report)
37
  Tools: music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
38
  Companions (not included): symbolic_symphony_set/README.md (column definitions and Roman numeral notation of harmony.csv)
 
40
  - `symbolic_symphony_set/mo4/sheet.xml`: Roman-numeral harmony, cadence and phrase annotation of Mozart Symphony No. 41 in C major, K. 551 ("Jupiter"), fourth movement (Molto allegro): full orchestral MusicXML score.
41
  Format: 4 files, 10,316,445 bytes: sheet.xml (MusicXML 4.0 full orchestral score, 17 parts, exported by MuseScore 4.3.2), harmony.csv (onset, offset, key, degree, quality, inversion, roman number), cadence.csv (onset, resolve time, offset, cadence type), form_analysis.csv (phrase onset, phrase offset, section, theme)
42
  Origin: author annotation (method not stated in the README) of a public-domain symphony on a CC0 OpenScore edition
 
 
43
  Check: give sheet.xml; hide harmony.csv, cadence.csv and form_analysis.csv; the agent outputs chord segments (onset, offset in crotchet beats from the start of the written score, key, degree, quality, inversion, Roman numeral); score by time-weighted overlap per field, each field reported separately; cadence.csv (onset, type AC, HC, PC, DC) and form_analysis.csv (phrase spans, section, theme) serve as separate cadence and phrase tasks scored by exact onset and label match; no pass line is fixed by the source (see the report)
44
  Tools: music21 (Python) or any XML parser reads sheet.xml headlessly (MuseScore also runs headless with -o for conversion); the csv module or pandas reads harmony.csv, cadence.csv and form_analysis.csv
45
  Companions (not included): symbolic_symphony_set/README.md (column definitions and Roman numeral notation of harmony.csv)
campaign_files/2026-10-ml-serving-configs-2/salesforce-merlion/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 3 of 3 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: labelled_anomaly_points. 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: Give the agent the timestamp and value columns only (the anomaly column removed); the agent returns one 0/1 flag per row. Compare with the hidden anomaly column row by row and report precision, recall and F1 of the positive class; exact equality per row, no tolerance and no score cut-off.
14
 
15
  **Licence (SPDX):** BSD-3-Clause
@@ -21,24 +19,18 @@ Files:
21
  - `data/synthetic_anomaly/seasonal_level_anomaly.csv`: Merlion synthetic anomaly seasonal_level_anomaly.csv: 10000 rows at 300 s steps, 600 rows labelled anomaly=1; level shifts (LevelShift, pos_prob=0.5, sd_range=(3, 6)) injected into a two-period sine series.
22
  Format: CSV, header timestamp,seasonal,anomaly; anomaly is 0.0 or 1.0 per row
23
  Origin: synthetic data written by Merlion's generator (np.random.seed(12345) and n = 10000 at lines 22 and 23, per-file seed np.random.seed(1000 * i + j) at line 69) and committed to the repository as a benchmark set
24
- Reference result: labelled_anomaly_points
25
- Reference source: the anomaly column of the file itself (600 rows equal 1.0), written by the generator: anomalies injected by level=LevelShift(pos_prob=0.5, sd_range=(3, 6), **anomalize_kwargs) (examples/misc/generate_synthetic_tsad_dataset.py line 55) with the shared kwargs anom_prob=0.002, anom_width_range=(20, 200), alpha=0.5 (line 49); the Merlion loader reads the label as df["anomaly"].astype(bool) (ts_datasets/ts_datasets/anomaly/synthetic.py line 58)
26
  Check: Give the agent the timestamp and value columns only (the anomaly column removed); the agent returns one 0/1 flag per row. Compare with the hidden anomaly column row by row and report precision, recall and F1 of the positive class; exact equality per row, no tolerance and no score cut-off
27
  Companions (not included): examples/misc/generate_synthetic_tsad_dataset.py; LICENSE
28
  Dependencies: none to read the data (CSV with columns timestamp,seasonal,anomaly; timestamp in seconds); the generator script is provenance only and is never run.
29
  - `data/synthetic_anomaly/upward_downward_trend_anomaly.csv`: Merlion synthetic anomaly upward_downward_trend_anomaly.csv: 10000 rows at 300 s steps, 92 rows labelled anomaly=1; trend changes (TrendChange, anom_prob=0.01, pos_prob=0.5, scale_range=(2.5, 5)) injected into an up-then-down trend series.
30
  Format: CSV, header timestamp,upward_downward,anomaly; anomaly is 0.0 or 1.0 per row
31
  Origin: synthetic data written by Merlion's generator (np.random.seed(12345) and n = 10000 at lines 22 and 23, per-file seed np.random.seed(1000 * i + j) at line 69) and committed to the repository as a benchmark set
32
- Reference result: labelled_anomaly_points
33
- Reference source: the anomaly column of the file itself (92 rows equal 1.0), written by the generator: anomalies injected by trend=TrendChange(anom_prob=0.01, pos_prob=0.5, scale_range=(2.5, 5)) (examples/misc/generate_synthetic_tsad_dataset.py line 56) with the shared kwargs anom_prob=0.002, anom_width_range=(20, 200), alpha=0.5 (line 49); the Merlion loader reads the label as df["anomaly"].astype(bool) (ts_datasets/ts_datasets/anomaly/synthetic.py line 58)
34
  Check: Give the agent the timestamp and value columns only (the anomaly column removed); the agent returns one 0/1 flag per row. Compare with the hidden anomaly column row by row and report precision, recall and F1 of the positive class; exact equality per row, no tolerance and no score cut-off
35
  Companions (not included): examples/misc/generate_synthetic_tsad_dataset.py; LICENSE
36
  Dependencies: none to read the data (CSV with columns timestamp,upward_downward,anomaly; timestamp in seconds); the generator script is provenance only and is never run.
37
  - `data/synthetic_anomaly/horizontal_shock_anomaly.csv`: Merlion synthetic anomaly horizontal_shock_anomaly.csv: 10000 rows at 300 s steps, 1957 rows labelled anomaly=1; shocks of both signs (Shock, pos_prob=0.5, sd_range=(4, 8)) injected into white noise.
38
  Format: CSV, header timestamp,horizontal,anomaly; anomaly is 0.0 or 1.0 per row
39
  Origin: synthetic data written by Merlion's generator (np.random.seed(12345) and n = 10000 at lines 22 and 23, per-file seed np.random.seed(1000 * i + j) at line 69) and committed to the repository as a benchmark set
40
- Reference result: labelled_anomaly_points
41
- Reference source: the anomaly column of the file itself (1957 rows equal 1.0), written by the generator: anomalies injected by shock=Shock(pos_prob=0.5, sd_range=(4, 8), **anomalize_kwargs) (examples/misc/generate_synthetic_tsad_dataset.py line 52) with the shared kwargs anom_prob=0.002, anom_width_range=(20, 200), alpha=0.5 (line 49); the Merlion loader reads the label as df["anomaly"].astype(bool) (ts_datasets/ts_datasets/anomaly/synthetic.py line 58)
42
  Check: Give the agent the timestamp and value columns only (the anomaly column removed); the agent returns one 0/1 flag per row. Compare with the hidden anomaly column row by row and report precision, recall and F1 of the positive class; exact equality per row, no tolerance and no score cut-off
43
  Companions (not included): examples/misc/generate_synthetic_tsad_dataset.py; LICENSE
44
  Dependencies: none to read the data (CSV with columns timestamp,horizontal,anomaly; timestamp in seconds); the generator script is provenance only and is never run.
 
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 3 of 3 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: Give the agent the timestamp and value columns only (the anomaly column removed); the agent returns one 0/1 flag per row. Compare with the hidden anomaly column row by row and report precision, recall and F1 of the positive class; exact equality per row, no tolerance and no score cut-off.
12
 
13
  **Licence (SPDX):** BSD-3-Clause
 
19
  - `data/synthetic_anomaly/seasonal_level_anomaly.csv`: Merlion synthetic anomaly seasonal_level_anomaly.csv: 10000 rows at 300 s steps, 600 rows labelled anomaly=1; level shifts (LevelShift, pos_prob=0.5, sd_range=(3, 6)) injected into a two-period sine series.
20
  Format: CSV, header timestamp,seasonal,anomaly; anomaly is 0.0 or 1.0 per row
21
  Origin: synthetic data written by Merlion's generator (np.random.seed(12345) and n = 10000 at lines 22 and 23, per-file seed np.random.seed(1000 * i + j) at line 69) and committed to the repository as a benchmark set
 
 
22
  Check: Give the agent the timestamp and value columns only (the anomaly column removed); the agent returns one 0/1 flag per row. Compare with the hidden anomaly column row by row and report precision, recall and F1 of the positive class; exact equality per row, no tolerance and no score cut-off
23
  Companions (not included): examples/misc/generate_synthetic_tsad_dataset.py; LICENSE
24
  Dependencies: none to read the data (CSV with columns timestamp,seasonal,anomaly; timestamp in seconds); the generator script is provenance only and is never run.
25
  - `data/synthetic_anomaly/upward_downward_trend_anomaly.csv`: Merlion synthetic anomaly upward_downward_trend_anomaly.csv: 10000 rows at 300 s steps, 92 rows labelled anomaly=1; trend changes (TrendChange, anom_prob=0.01, pos_prob=0.5, scale_range=(2.5, 5)) injected into an up-then-down trend series.
26
  Format: CSV, header timestamp,upward_downward,anomaly; anomaly is 0.0 or 1.0 per row
27
  Origin: synthetic data written by Merlion's generator (np.random.seed(12345) and n = 10000 at lines 22 and 23, per-file seed np.random.seed(1000 * i + j) at line 69) and committed to the repository as a benchmark set
 
 
28
  Check: Give the agent the timestamp and value columns only (the anomaly column removed); the agent returns one 0/1 flag per row. Compare with the hidden anomaly column row by row and report precision, recall and F1 of the positive class; exact equality per row, no tolerance and no score cut-off
29
  Companions (not included): examples/misc/generate_synthetic_tsad_dataset.py; LICENSE
30
  Dependencies: none to read the data (CSV with columns timestamp,upward_downward,anomaly; timestamp in seconds); the generator script is provenance only and is never run.
31
  - `data/synthetic_anomaly/horizontal_shock_anomaly.csv`: Merlion synthetic anomaly horizontal_shock_anomaly.csv: 10000 rows at 300 s steps, 1957 rows labelled anomaly=1; shocks of both signs (Shock, pos_prob=0.5, sd_range=(4, 8)) injected into white noise.
32
  Format: CSV, header timestamp,horizontal,anomaly; anomaly is 0.0 or 1.0 per row
33
  Origin: synthetic data written by Merlion's generator (np.random.seed(12345) and n = 10000 at lines 22 and 23, per-file seed np.random.seed(1000 * i + j) at line 69) and committed to the repository as a benchmark set
 
 
34
  Check: Give the agent the timestamp and value columns only (the anomaly column removed); the agent returns one 0/1 flag per row. Compare with the hidden anomaly column row by row and report precision, recall and F1 of the positive class; exact equality per row, no tolerance and no score cut-off
35
  Companions (not included): examples/misc/generate_synthetic_tsad_dataset.py; LICENSE
36
  Dependencies: none to read the data (CSV with columns timestamp,horizontal,anomaly; timestamp in seconds); the generator script is provenance only and is never run.
campaign_files/2026-10-op-rule-calculations-2/saleor-warehouse-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 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: test_assertions. 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:** pytest on the test module (headless, PostgreSQL test database), or the Saleor GraphQL API to read stocks and allocations
14
 
15
  **Licence (SPDX):** BSD-3-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `saleor/warehouse/tests/test_stock_management.py`: Saleor stock management tests: allocating, deallocating, increasing and decreasing stock for order lines across warehouses.
22
  Format: pytest module, 1712 lines, with 2 fixture files read for the stock quantities
23
  Origin: test written by the project authors; products, quantities and prices are made up
24
- Reference result: test_assertions
25
- Reference source: stock quantity and quantity_allocated after each allocate, deallocate, increase or decrease call, and the allocation rows per stock (for example lines 43-46)
26
  Check: give the stocks per warehouse (quantity, allocated) and the order lines, the allocation strategy and the operation, and ask the resulting quantity, quantity_allocated and allocations; compare exactly as the asserts do
27
  Tools: pytest on the test module (headless, PostgreSQL test database), or the Saleor GraphQL API to read stocks and allocations
28
  Companions (not included): saleor/warehouse/tests/fixtures/stock.py; saleor/warehouse/tests/fixtures/allocation.py
@@ -30,8 +26,6 @@ Files:
30
  - `saleor/warehouse/tests/test_stock_availability.py`: Saleor stock availability tests: available quantity per variant and country after allocations and checkout reservations.
31
  Format: pytest module, 656 lines, with 1 fixture file read
32
  Origin: test written by the project authors; products, quantities and prices are made up
33
- Reference result: test_assertions
34
- Reference source: available quantity asserted per case: 7 without allocations (line 148), 3 with allocations (line 233), 2 with reservations (line 249), and the parametrised totals at lines 153-155
35
  Check: give the stocks per warehouse with their allocations and reservations and ask the quantity still available to sell in the channel; compare exactly
36
  Tools: pytest on the test module (headless, PostgreSQL test database), or the Saleor GraphQL API to read stocks and allocations
37
  Companions (not included): saleor/warehouse/tests/fixtures/stock.py
 
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 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:** pytest on the test module (headless, PostgreSQL test database), or the Saleor GraphQL API to read stocks and allocations
12
 
13
  **Licence (SPDX):** BSD-3-Clause
 
19
  - `saleor/warehouse/tests/test_stock_management.py`: Saleor stock management tests: allocating, deallocating, increasing and decreasing stock for order lines across warehouses.
20
  Format: pytest module, 1712 lines, with 2 fixture files read for the stock quantities
21
  Origin: test written by the project authors; products, quantities and prices are made up
 
 
22
  Check: give the stocks per warehouse (quantity, allocated) and the order lines, the allocation strategy and the operation, and ask the resulting quantity, quantity_allocated and allocations; compare exactly as the asserts do
23
  Tools: pytest on the test module (headless, PostgreSQL test database), or the Saleor GraphQL API to read stocks and allocations
24
  Companions (not included): saleor/warehouse/tests/fixtures/stock.py; saleor/warehouse/tests/fixtures/allocation.py
 
26
  - `saleor/warehouse/tests/test_stock_availability.py`: Saleor stock availability tests: available quantity per variant and country after allocations and checkout reservations.
27
  Format: pytest module, 656 lines, with 1 fixture file read
28
  Origin: test written by the project authors; products, quantities and prices are made up
 
 
29
  Check: give the stocks per warehouse with their allocations and reservations and ask the quantity still available to sell in the channel; compare exactly
30
  Tools: pytest on the test module (headless, PostgreSQL test database), or the Saleor GraphQL API to read stocks and allocations
31
  Companions (not included): saleor/warehouse/tests/fixtures/stock.py
campaign_files/2026-10-s-security-detection-2/securinets-insat-friendly-ctf-2026-public-challenges/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each unit is self-contained. 4 of 4 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: flag_string. 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: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Variant: name attacker IP 45.83.220.14 and identity mkedri@agilogix.tn, exact match against solution.md line 12. Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Variant: report the three ZipCrypto internal keys, exact match against solution.md line 94. Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Red herrings named in solution.md line 56 (reset?token=RESET_7c1e94ab) are wrong answers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,32 +19,24 @@ Units:
21
  - `forensics/wave3/trailhead/` (3 files): Securinets Friendly CTF 2026 forensics/Trailhead: stitch Entra ID sign-ins, AWS CloudTrail and S3 access logs into a cloud account-takeover timeline and recover the exfiltrated secret.
22
  Format: three cloud logs: CloudTrail management events JSON (173,673 B), Entra ID sign-in log JSON (22,042 B), S3 server access log text (191,413 B)
23
  Origin: official competition instance built by the organisers for a public CTF and released with its flag and solution (fixture of a competition, not recorded production telemetry)
24
- Reference result: flag_string
25
- Reference source: challenge.yml line 24 and flag.txt (Securinets{cl0udtr41l_t1m3l1n3_st1tch3d}); intended path in solution.md (impossible travel, CreateAccessKey at 2019-06-14T02:27:58Z, base64 flag in the exfiltrated S3 object key)
26
  Check: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Variant: name attacker IP 45.83.220.14 and identity mkedri@agilogix.tn, exact match against solution.md line 12.
27
  Companions (not included): challenge.yml; flag.txt; solution.md; Read.md
28
  Dependencies: challenge.yml, flag.txt, solution.md hold the flag, answers or official solution (hidden from the solver); solver tools are free choices (for example an EVTX, SQLite, pcap or zip tool), version fixed when the pool is built
29
  - `forensics/wave3/zip-it/` (2 files): Securinets Friendly CTF 2026 forensics/Zip It: open a ZipCrypto-protected vault export with a known-plaintext attack using the bundled logo file.
30
  Format: ZipCrypto-encrypted zip (12,033 B; members securinets.jpg 11,411 B and notes.txt 243 B, both with the encryption bit set) plus the plaintext securinets.jpg (11,411 B)
31
  Origin: official competition instance built by the organisers for a public CTF and released with its flag and solution (fixture of a competition, not recorded production telemetry)
32
- Reference result: flag_string
33
- Reference source: challenge.yml line 22 and flag.txt (Securinets{kn0wn_pl41nt3xt_b34ts_brut3_f0rc3}); solution.md lines 74 to 100 give the bkcrack commands, the recovered keys edd5f7a3 ba9605d6 194c46bb and the vault_unseal_key line
34
  Check: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Variant: report the three ZipCrypto internal keys, exact match against solution.md line 94.
35
  Companions (not included): challenge.yml; flag.txt; solution.md; Read.md
36
  Dependencies: challenge.yml, flag.txt, solution.md hold the flag, answers or official solution (hidden from the solver); solver tools are free choices (for example an EVTX, SQLite, pcap or zip tool), version fixed when the pool is built
37
  - `forensics/wave2/force-push/attachments/repo.zip`: Securinets Friendly CTF 2026 forensics/Force Push: recover production credentials that a developer committed and then erased by a force-push, from the reflog and dangling objects of a local clone.
38
  Format: zip (44,891 B, 125 members) of a git working copy logi-api/ with its .git directory (objects, logs/HEAD, refs, lost-found)
39
  Origin: official competition instance built by the organisers for a public CTF and released with its flag and solution (fixture of a competition, not recorded production telemetry)
40
- Reference result: flag_string
41
- Reference source: challenge.yml line 23 and flag.txt (Securinets{g1t_r3fl0g_n3v3r_f0rg3ts}); solution.md lines 65 to 120 give reflog path A and fsck path B, the flag is the master_token value in the recovered config (line 96)
42
  Check: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker).
43
  Companions (not included): challenge.yml; flag.txt; solution.md; Read.md
44
  Dependencies: challenge.yml, flag.txt, solution.md hold the flag, answers or official solution (hidden from the solver); solver tools are free choices (for example an EVTX, SQLite, pcap or zip tool), version fixed when the pool is built
45
  - `forensics/wave2/cookie-jar/attachments/chrome_profile.zip`: Securinets Friendly CTF 2026 forensics/Cookie Jar: recover a cleared browser history entry holding a session token from a Chromium History database and its uncheckpointed WAL.
46
  Format: zip (2,763 B) of a Chromium Default profile: Bookmarks, Cookies (SQLite), History (SQLite, 16,384 B), History-shm and History-wal (8,272 B)
47
  Origin: official competition instance built by the organisers for a public CTF and released with its flag and solution (fixture of a competition, not recorded production telemetry)
48
- Reference result: flag_string
49
- Reference source: challenge.yml line 20 and flag.txt (Securinets{h1st0ry_l1v3s_1n_th3_w4l}); solution.md lines 26 to 45: open History without the WAL, the session= parameter of https://vault.agilogix.tn/unlock is the flag
50
  Check: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Red herrings named in solution.md line 56 (reset?token=RESET_7c1e94ab) are wrong answers.
51
  Companions (not included): challenge.yml; flag.txt; solution.md; Read.md
52
  Dependencies: challenge.yml, flag.txt, solution.md hold the flag, answers or official solution (hidden from the solver); solver tools are free choices (for example an EVTX, SQLite, pcap or zip tool), version fixed when the pool is built
 
8
 
9
  **Self-contained:** Each unit is self-contained. 4 of 4 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: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Variant: name attacker IP 45.83.220.14 and identity mkedri@agilogix.tn, exact match against solution.md line 12. Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Variant: report the three ZipCrypto internal keys, exact match against solution.md line 94. Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Red herrings named in solution.md line 56 (reset?token=RESET_7c1e94ab) are wrong answers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `forensics/wave3/trailhead/` (3 files): Securinets Friendly CTF 2026 forensics/Trailhead: stitch Entra ID sign-ins, AWS CloudTrail and S3 access logs into a cloud account-takeover timeline and recover the exfiltrated secret.
20
  Format: three cloud logs: CloudTrail management events JSON (173,673 B), Entra ID sign-in log JSON (22,042 B), S3 server access log text (191,413 B)
21
  Origin: official competition instance built by the organisers for a public CTF and released with its flag and solution (fixture of a competition, not recorded production telemetry)
 
 
22
  Check: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Variant: name attacker IP 45.83.220.14 and identity mkedri@agilogix.tn, exact match against solution.md line 12.
23
  Companions (not included): challenge.yml; flag.txt; solution.md; Read.md
24
  Dependencies: challenge.yml, flag.txt, solution.md hold the flag, answers or official solution (hidden from the solver); solver tools are free choices (for example an EVTX, SQLite, pcap or zip tool), version fixed when the pool is built
25
  - `forensics/wave3/zip-it/` (2 files): Securinets Friendly CTF 2026 forensics/Zip It: open a ZipCrypto-protected vault export with a known-plaintext attack using the bundled logo file.
26
  Format: ZipCrypto-encrypted zip (12,033 B; members securinets.jpg 11,411 B and notes.txt 243 B, both with the encryption bit set) plus the plaintext securinets.jpg (11,411 B)
27
  Origin: official competition instance built by the organisers for a public CTF and released with its flag and solution (fixture of a competition, not recorded production telemetry)
 
 
28
  Check: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Variant: report the three ZipCrypto internal keys, exact match against solution.md line 94.
29
  Companions (not included): challenge.yml; flag.txt; solution.md; Read.md
30
  Dependencies: challenge.yml, flag.txt, solution.md hold the flag, answers or official solution (hidden from the solver); solver tools are free choices (for example an EVTX, SQLite, pcap or zip tool), version fixed when the pool is built
31
  - `forensics/wave2/force-push/attachments/repo.zip`: Securinets Friendly CTF 2026 forensics/Force Push: recover production credentials that a developer committed and then erased by a force-push, from the reflog and dangling objects of a local clone.
32
  Format: zip (44,891 B, 125 members) of a git working copy logi-api/ with its .git directory (objects, logs/HEAD, refs, lost-found)
33
  Origin: official competition instance built by the organisers for a public CTF and released with its flag and solution (fixture of a competition, not recorded production telemetry)
 
 
34
  Check: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker).
35
  Companions (not included): challenge.yml; flag.txt; solution.md; Read.md
36
  Dependencies: challenge.yml, flag.txt, solution.md hold the flag, answers or official solution (hidden from the solver); solver tools are free choices (for example an EVTX, SQLite, pcap or zip tool), version fixed when the pool is built
37
  - `forensics/wave2/cookie-jar/attachments/chrome_profile.zip`: Securinets Friendly CTF 2026 forensics/Cookie Jar: recover a cleared browser history entry holding a session token from a Chromium History database and its uncheckpointed WAL.
38
  Format: zip (2,763 B) of a Chromium Default profile: Bookmarks, Cookies (SQLite), History (SQLite, 16,384 B), History-shm and History-wal (8,272 B)
39
  Origin: official competition instance built by the organisers for a public CTF and released with its flag and solution (fixture of a competition, not recorded production telemetry)
 
 
40
  Check: Exact string match of the submitted flag against challenge.yml (static flag, no case-insensitive marker). Red herrings named in solution.md line 56 (reset?token=RESET_7c1e94ab) are wrong answers.
41
  Companions (not included): challenge.yml; flag.txt; solution.md; Read.md
42
  Dependencies: challenge.yml, flag.txt, solution.md hold the flag, answers or official solution (hidden from the solver); solver tools are free choices (for example an EVTX, SQLite, pcap or zip tool), version fixed when the pool is built
campaign_files/2026-10-sc-science-reference-data-2/rutgers-gsl-monthly-snow-cover/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.
10
 
11
- **Answers:** No answer files are included. Reference result: published_aggregate_in_companion_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:** Check: Area = round(km2/1000); normal = round(mean of the same month 1991 to 2020 / 1000); departure = round(area_unrounded/1000 - normal_unrounded/1000); rank among all years of the same calendar month by area rounded to thousands, ties shown as a-b; n = years of that month in the record; max and min with year. The record is cut at 2026-08, the page's latest month (moncov already holds 2026-09). Checked this round: 56 of 56 months, all 7 fields equal.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `snow_cover/moncov.nhland.txt`: Northern Hemisphere monthly snow cover extent, 1966-11 to 2026-08: recompute area, 1991-2020 normal, departure, rank, record length and extremes.
22
  Format: whitespace table, one row per month: year, month, area in square km (integer)
23
  Origin: NOAA snow charts digitized by the Rutgers Global Snow Lab
24
- Reference result: published_aggregate_in_companion_file
25
- Reference source: rutgers_rank.html Northern Hemisphere table (line 140 onward; line 130 "All snow cover extents are shown in thousands of square kilometers"): area, 1991-2020 normal, departure, rank, years of record, maximum and minimum with year, for 56 months 2022-01 to 2026-08
26
  Check: Area = round(km2/1000); normal = round(mean of the same month 1991 to 2020 / 1000); departure = round(area_unrounded/1000 - normal_unrounded/1000); rank among all years of the same calendar month by area rounded to thousands, ties shown as a-b; n = years of that month in the record; max and min with year. The record is cut at 2026-08, the page's latest month (moncov already holds 2026-09). Checked this round: 56 of 56 months, all 7 fields equal
27
  Companions (not included): pages/rutgers_rank.html (answer: Monthly Snow Area Rankings, Northern Hemisphere block; hidden from the agent)
 
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.
10
 
 
 
11
  **Tools to open and check:** Check: Area = round(km2/1000); normal = round(mean of the same month 1991 to 2020 / 1000); departure = round(area_unrounded/1000 - normal_unrounded/1000); rank among all years of the same calendar month by area rounded to thousands, ties shown as a-b; n = years of that month in the record; max and min with year. The record is cut at 2026-08, the page's latest month (moncov already holds 2026-09). Checked this round: 56 of 56 months, all 7 fields equal.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `snow_cover/moncov.nhland.txt`: Northern Hemisphere monthly snow cover extent, 1966-11 to 2026-08: recompute area, 1991-2020 normal, departure, rank, record length and extremes.
20
  Format: whitespace table, one row per month: year, month, area in square km (integer)
21
  Origin: NOAA snow charts digitized by the Rutgers Global Snow Lab
 
 
22
  Check: Area = round(km2/1000); normal = round(mean of the same month 1991 to 2020 / 1000); departure = round(area_unrounded/1000 - normal_unrounded/1000); rank among all years of the same calendar month by area rounded to thousands, ties shown as a-b; n = years of that month in the record; max and min with year. The record is cut at 2026-08, the page's latest month (moncov already holds 2026-09). Checked this round: 56 of 56 months, all 7 fields equal
23
  Companions (not included): pages/rutgers_rank.html (answer: Monthly Snow Area Rankings, Northern Hemisphere block; hidden from the agent)
campaign_files/2026-10-sc-science-reference-data-2/scilab-cacsd-module-unit-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: routh_table_and_sign_changes; peak_frequencies_hz; riccati_solution_matrix; kalman_gain_matrix; gain_phase_margins_and_crossover_frequencies; natural_frequencies_and_damping_ratios; time_response_samples. 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:** Tools are named in the check written per file in the notes.
14
 
15
  **Licence (SPDX):** GPL-2.0-only OR CeCILL-2.1
@@ -21,49 +19,35 @@ Files:
21
  - `scilab/modules/cacsd/tests/unit_tests/routh_t.tst`: Routh-Hurwitz tables of polynomials, including a zero first-column entry, a row of zeros and a symbolic gain (Scilab routh_t unit test).
22
  Format: Scilab test script .tst (1928 bytes); asserts at lines 19 to 47; error-message checks at lines 50 to 54 are not part of the task
23
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
24
- Reference result: routh_table_and_sign_changes
25
- Reference source: scilab/modules/cacsd/tests/unit_tests/routh_t.tst: tables at lines 19, 26 to 33, 37 and 44, sign-change counts num at lines 22, 40 and 47
26
  Check: four cases posed one by one: P1 = s(s+7)(s+11), P2 = 2s^2 - 24, P3 = s^4 - 6s^3 + 10s^2 - 6s + 9, and the closed loop of 1/P1 with a symbolic gain k; the agent reports the Routh table (rows highest power first, a zero row replaced by the coefficients of the derivative of the auxiliary polynomial) and the number of sign changes in the first column; compared exactly: P1 [1 77; 18 0; 77 0; 0 0] with 0 changes, P2 [2 -24; 4 0; -24 0] with 1, P3 [1 10 9; -6 -6 0; 9 9 0; 18 0 0; 9 0 0] with 2, gain case [1 77; 18 k; (1386-k)/18 0; k 0]; the test file is withheld when posed
27
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
28
  - `scilab/modules/cacsd/tests/unit_tests/freson.tst`: Resonance (peak) frequencies of a continuous-time transfer function and of its zero-order-hold discretisation (Scilab freson unit test).
29
  Format: Scilab test script .tst (610 bytes); asserts at lines 12 and 17
30
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
31
- Reference result: peak_frequencies_hz
32
- Reference source: scilab/modules/cacsd/tests/unit_tests/freson.tst: [1.12527981797; 0.26162097704] at line 12 (continuous), [1.12527923264; 0.26155702493] at line 17 (discretised with sample time 0.05)
33
  Check: pose: h(s) = (s - 1)/((s^2 + 2s + 3)(s^2 + 0.1s + 50)) (line 10); the agent reports the frequencies in Hz at which |h(j 2 pi f)| has a local maximum, in decreasing order; then the same for the zero-order-hold discretisation with sample time 0.05 (line 15), using |h(exp(j 2 pi f 0.05))|; compared element-wise with the default relative tolerance sqrt(%eps) for the continuous case and 3e-8 for the discrete case; the test file is withheld when posed
34
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
35
  - `scilab/modules/cacsd/tests/unit_tests/ricc.tst`: Stabilising solution of a continuous-time algebraic Riccati equation, 3 by 3 (Scilab ricc unit test).
36
  Format: Scilab test script .tst (659 bytes); assert at line 19
37
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
38
- Reference result: riccati_solution_matrix
39
- Reference source: scilab/modules/cacsd/tests/unit_tests/ricc.tst: X_ref at line 18
40
  Check: pose: A = [8,8,9;1,8,6;2,5,9], B0 = [0,6;7,8;4,0], C0 = [8,5,5;9,8,1;5,0,7], C = C0*C0', R0 = [2,9;5,7], R = R0*R0' + I, B = B0*inv(R)*B0' (lines 8 to 16); the agent solves A'X + XA - XBX + C = 0 for the stabilising X (A - BX stable) and reports X; compared element-wise with X_ref within the relative tolerance 1e-6; the test file is withheld when posed
41
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
42
  - `scilab/modules/cacsd/tests/unit_tests/lqe.tst`: Kalman filter gain with correlated process and measurement noise, 5 states and 2 outputs (Scilab lqe unit test).
43
  Format: Scilab test script .tst (3274 bytes); the posed case is lines 10 to 70; later cases (default S, discrete time) have only residual and stability checks
44
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
45
- Reference result: kalman_gain_matrix
46
- Reference source: scilab/modules/cacsd/tests/unit_tests/lqe.tst: K_ref at lines 65 to 69
47
  Check: pose: the 5-state system and noise model at lines 10 to 50 (A, C, G, H, Q_e = I, R_v = diag(1,2), S_ev(1,1) = 0.8, Q = G Q_e G', R = H Q_e H' + S_ev'H' + H S_ev + R_v, S = G(Q_e H' + S_ev)); the agent computes the continuous-time steady-state filter gain K for the estimator with A + KC stable, sign convention K = -(X C' + S) R^-1 with X the stabilising solution of (A - S R^-1 C)X + X(A - S R^-1 C)' - X C' R^-1 C X + Q - S R^-1 S' = 0 (line 64); compared by norm(K - K_ref, 1) < 1e-10; the test file is withheld when posed (only the long-syntax case of lines 59 to 70 is posed)
48
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
49
  - `scilab/modules/cacsd/tests/unit_tests/margins.tst`: Gain and phase margins of continuous and discrete-time transfer functions, Maple reference values (Scilab margins unit test).
50
  Format: Scilab test script .tst (6051 bytes); asserts at lines 42 to 174
51
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
52
- Reference result: gain_phase_margins_and_crossover_frequencies
53
- Reference source: scilab/modules/cacsd/tests/unit_tests/margins.tst: f_ref and g_ref literals at lines 40 to 105 (gain margin), f_ref and p_ref literals at lines 135 to 171 (phase margin); the comments at lines 30 to 37, 73 to 80, 128 to 133 and 156 to 161 give the Maple commands that produced them
54
  Check: cases posed one by one: four discrete-time (sample time 0.1 or 1) and four continuous-time transfer functions for the gain margin (lines 39 to 108), and three discrete-time (lines 134, 142 and 148; line 147 is overwritten by line 148) and two continuous-time (lines 162 and 169) transfer functions for the phase margin; the agent reports the gain margin in dB with its phase-crossover frequency in Hz, and the phase margin in degrees with its gain-crossover frequency in Hz (the first crossing, as the reference values take it); each value is compared with the reference by relative error < 1e-10 (gain margin) or < 1e-9 (phase margin), and by absolute error when the reference is 0; the test file is withheld when posed
55
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
56
  - `scilab/modules/cacsd/tests/unit_tests/damp.tst`: Natural frequencies and damping ratios of the poles of a 7th-degree polynomial, continuous and discrete time, as roots, polynomial, transfer function, zero-pole-gain and state-space forms (Scilab damp unit test).
57
  Format: Scilab test script .tst (4625 bytes); asserts at lines 13 to 128; invalid-call checks at lines 132 to 135 are not part of the task
58
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
59
- Reference result: natural_frequencies_and_damping_ratios
60
- Reference source: scilab/modules/cacsd/tests/unit_tests/damp.tst: closed-form references wnref and zetaref at lines 45 to 46 and 58 to 61 and the repeats down to line 126 (built from the known roots 2, 1+i, 1-i, 2-3i, 2+3i, 0 and 1 of P at line 43)
61
  Check: pose: P(s) = s(s-1)(s-2)(s^2-2s+2)(s^2-4s+13) (line 43, the product of its entries); the agent reports, for each pole r sorted by increasing natural frequency, wn = |r| and zeta = -Re(r)/|r| in continuous time, and wn = |log(r)/dt| and zeta = -Re(log(r)/dt)/wn in discrete time with dt = 1 (a pole at 0 gives wn = inf and zeta = nan; a pole with wn = 0 gives zeta = -1); expected continuous wn = [0; 1; sqrt(2); sqrt(2); 2; sqrt(13); sqrt(13)] and zeta = [-1; -1; -sqrt(2)/2; -sqrt(2)/2; -1; -2/sqrt(13); -2/sqrt(13)], discrete wn = [0; log 2; |log(1+i)|; |log(1+i)|; |log(2-3i)|; |log(2-3i)|; inf]; compared element-wise as in the test; the first damping ratio is checked only up to sign (abs(abs(zeta(1)) - 1) < 100 %eps); infinite and nan entries must sit at the same places; the test file is withheld when posed
62
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
63
  - `scilab/modules/cacsd/tests/unit_tests/csim.tst`: Step, impulse and user-input time responses of a first-order integrator model with zero and non-zero initial state (Scilab csim unit test).
64
  Format: Scilab test script .tst (3811 bytes); asserts at lines 22 to 158
65
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
66
- Reference result: time_response_samples
67
- Reference source: scilab/modules/cacsd/tests/unit_tests/csim.tst: xrefstep and yrefstep at lines 17 and 18, yrefimp and xrefimp at lines 58 and 59, yrefstep2 and xrefstep2 at lines 135 and 136
68
  Check: pose: the continuous-time model x' = u, y = 2x (a = 0, b = 1, c = 2, lines 13 and 15) sampled at t = 0, 0.1, ..., 1 (line 14); the agent reports y and x for a unit step from x(0) = 0, for a unit impulse, and for a unit step from x(0) = 1, in state-space, transfer-function and zero-pole-gain form and with the input given as a function; expected step y = 0:0.2:2 and x = 0:0.1:1, impulse y = 2 and x = 1 at every sample, step from x(0) = 1 y = 2:0.2:4 and x = 1:0.1:2; compared element-wise with the default relative tolerance sqrt(%eps); the test file is withheld when posed
69
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
 
8
 
9
  **Self-contained:** Each file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Tools are named in the check written per file in the notes.
12
 
13
  **Licence (SPDX):** GPL-2.0-only OR CeCILL-2.1
 
19
  - `scilab/modules/cacsd/tests/unit_tests/routh_t.tst`: Routh-Hurwitz tables of polynomials, including a zero first-column entry, a row of zeros and a symbolic gain (Scilab routh_t unit test).
20
  Format: Scilab test script .tst (1928 bytes); asserts at lines 19 to 47; error-message checks at lines 50 to 54 are not part of the task
21
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
 
 
22
  Check: four cases posed one by one: P1 = s(s+7)(s+11), P2 = 2s^2 - 24, P3 = s^4 - 6s^3 + 10s^2 - 6s + 9, and the closed loop of 1/P1 with a symbolic gain k; the agent reports the Routh table (rows highest power first, a zero row replaced by the coefficients of the derivative of the auxiliary polynomial) and the number of sign changes in the first column; compared exactly: P1 [1 77; 18 0; 77 0; 0 0] with 0 changes, P2 [2 -24; 4 0; -24 0] with 1, P3 [1 10 9; -6 -6 0; 9 9 0; 18 0 0; 9 0 0] with 2, gain case [1 77; 18 k; (1386-k)/18 0; k 0]; the test file is withheld when posed
23
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
24
  - `scilab/modules/cacsd/tests/unit_tests/freson.tst`: Resonance (peak) frequencies of a continuous-time transfer function and of its zero-order-hold discretisation (Scilab freson unit test).
25
  Format: Scilab test script .tst (610 bytes); asserts at lines 12 and 17
26
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
 
 
27
  Check: pose: h(s) = (s - 1)/((s^2 + 2s + 3)(s^2 + 0.1s + 50)) (line 10); the agent reports the frequencies in Hz at which |h(j 2 pi f)| has a local maximum, in decreasing order; then the same for the zero-order-hold discretisation with sample time 0.05 (line 15), using |h(exp(j 2 pi f 0.05))|; compared element-wise with the default relative tolerance sqrt(%eps) for the continuous case and 3e-8 for the discrete case; the test file is withheld when posed
28
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
29
  - `scilab/modules/cacsd/tests/unit_tests/ricc.tst`: Stabilising solution of a continuous-time algebraic Riccati equation, 3 by 3 (Scilab ricc unit test).
30
  Format: Scilab test script .tst (659 bytes); assert at line 19
31
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
 
 
32
  Check: pose: A = [8,8,9;1,8,6;2,5,9], B0 = [0,6;7,8;4,0], C0 = [8,5,5;9,8,1;5,0,7], C = C0*C0', R0 = [2,9;5,7], R = R0*R0' + I, B = B0*inv(R)*B0' (lines 8 to 16); the agent solves A'X + XA - XBX + C = 0 for the stabilising X (A - BX stable) and reports X; compared element-wise with X_ref within the relative tolerance 1e-6; the test file is withheld when posed
33
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
34
  - `scilab/modules/cacsd/tests/unit_tests/lqe.tst`: Kalman filter gain with correlated process and measurement noise, 5 states and 2 outputs (Scilab lqe unit test).
35
  Format: Scilab test script .tst (3274 bytes); the posed case is lines 10 to 70; later cases (default S, discrete time) have only residual and stability checks
36
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
 
 
37
  Check: pose: the 5-state system and noise model at lines 10 to 50 (A, C, G, H, Q_e = I, R_v = diag(1,2), S_ev(1,1) = 0.8, Q = G Q_e G', R = H Q_e H' + S_ev'H' + H S_ev + R_v, S = G(Q_e H' + S_ev)); the agent computes the continuous-time steady-state filter gain K for the estimator with A + KC stable, sign convention K = -(X C' + S) R^-1 with X the stabilising solution of (A - S R^-1 C)X + X(A - S R^-1 C)' - X C' R^-1 C X + Q - S R^-1 S' = 0 (line 64); compared by norm(K - K_ref, 1) < 1e-10; the test file is withheld when posed (only the long-syntax case of lines 59 to 70 is posed)
38
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
39
  - `scilab/modules/cacsd/tests/unit_tests/margins.tst`: Gain and phase margins of continuous and discrete-time transfer functions, Maple reference values (Scilab margins unit test).
40
  Format: Scilab test script .tst (6051 bytes); asserts at lines 42 to 174
41
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
 
 
42
  Check: cases posed one by one: four discrete-time (sample time 0.1 or 1) and four continuous-time transfer functions for the gain margin (lines 39 to 108), and three discrete-time (lines 134, 142 and 148; line 147 is overwritten by line 148) and two continuous-time (lines 162 and 169) transfer functions for the phase margin; the agent reports the gain margin in dB with its phase-crossover frequency in Hz, and the phase margin in degrees with its gain-crossover frequency in Hz (the first crossing, as the reference values take it); each value is compared with the reference by relative error < 1e-10 (gain margin) or < 1e-9 (phase margin), and by absolute error when the reference is 0; the test file is withheld when posed
43
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
44
  - `scilab/modules/cacsd/tests/unit_tests/damp.tst`: Natural frequencies and damping ratios of the poles of a 7th-degree polynomial, continuous and discrete time, as roots, polynomial, transfer function, zero-pole-gain and state-space forms (Scilab damp unit test).
45
  Format: Scilab test script .tst (4625 bytes); asserts at lines 13 to 128; invalid-call checks at lines 132 to 135 are not part of the task
46
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
 
 
47
  Check: pose: P(s) = s(s-1)(s-2)(s^2-2s+2)(s^2-4s+13) (line 43, the product of its entries); the agent reports, for each pole r sorted by increasing natural frequency, wn = |r| and zeta = -Re(r)/|r| in continuous time, and wn = |log(r)/dt| and zeta = -Re(log(r)/dt)/wn in discrete time with dt = 1 (a pole at 0 gives wn = inf and zeta = nan; a pole with wn = 0 gives zeta = -1); expected continuous wn = [0; 1; sqrt(2); sqrt(2); 2; sqrt(13); sqrt(13)] and zeta = [-1; -1; -sqrt(2)/2; -sqrt(2)/2; -1; -2/sqrt(13); -2/sqrt(13)], discrete wn = [0; log 2; |log(1+i)|; |log(1+i)|; |log(2-3i)|; |log(2-3i)|; inf]; compared element-wise as in the test; the first damping ratio is checked only up to sign (abs(abs(zeta(1)) - 1) < 100 %eps); infinite and nan entries must sit at the same places; the test file is withheld when posed
48
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
49
  - `scilab/modules/cacsd/tests/unit_tests/csim.tst`: Step, impulse and user-input time responses of a first-order integrator model with zero and non-zero initial state (Scilab csim unit test).
50
  Format: Scilab test script .tst (3811 bytes); asserts at lines 22 to 158
51
  Origin: unit test of the Scilab CACSD (computer-aided control system design) module; the reference literals are the module's regression values
 
 
52
  Check: pose: the continuous-time model x' = u, y = 2x (a = 0, b = 1, c = 2, lines 13 and 15) sampled at t = 0, 0.1, ..., 1 (line 14); the agent reports y and x for a unit step from x(0) = 0, for a unit impulse, and for a unit step from x(0) = 1, in state-space, transfer-function and zero-pole-gain form and with the input given as a function; expected step y = 0:0.2:2 and x = 0:0.1:1, impulse y = 2 and x = 1 at every sample, step from x(0) = 1 y = 2:0.2:4 and x = 1:0.1:2; compared element-wise with the default relative tolerance sqrt(%eps); the test file is withheld when posed
53
  Dependencies: Scilab at 2614b17ec9199ee90788d69cd4bc8e9bb036fc25 (pinned commit), needed only to re-run the upstream test; not packed
campaign_files/2026-10-sc-science-reference-data-2/scine-sparrow-python-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. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: reference_numbers_with_tolerance. 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: Give the agent the formaldehyde geometry and the method (PM3 reference, CIS-NDDO, 10 roots) and ask for the singlet and triplet excitation energies in eV; each passes when round(expected - computed, 2) == 0, the unittest assertAlmostEqual(places=2) rule of lines 64 and 74 (about |diff| <= 0.005 eV).
14
 
15
  **Licence (SPDX):** BSD-3-Clause
@@ -21,7 +19,5 @@ Files:
21
  - `src/Sparrow/Python/Tests/CalculatorWithReferenceTest.py`: SCINE Sparrow: formaldehyde (4 atoms, inline Angstrom geometry) with PM3 reference and CIS-NDDO; compute the 10 lowest singlet and triplet vertical excitation energies in eV.
22
  Format: Python unittest file with inline Cartesian geometry (Angstrom) and reference excitation energies (eV)
23
  Origin: Python regression test of SCINE Sparrow (tool_fixture); reference values written by the developers
24
- Reference result: reference_numbers_with_tolerance
25
- Reference source: CalculatorWithReferenceTest.py line 62: singlet transition energies [2.716, 5.476, 6.506, 7.986, 8.526, 9.111, 9.114, 9.606, 10.719, 11.041] eV; line 72: triplet [2.341, 4.316, 4.868, 5.706, 7.094, 7.908, 8.533, 9.037, 10.271, 10.369] eV; line 92: TD-DFTB singlets of a distorted methane (geometry lines 22-29); formaldehyde geometry at lines 15-20, PM3 with self_consistence_criterion 1e-9 at lines 36-37
26
  Check: give the agent the formaldehyde geometry and the method (PM3 reference, CIS-NDDO, 10 roots) and ask for the singlet and triplet excitation energies in eV; each passes when round(expected - computed, 2) == 0, the unittest assertAlmostEqual(places=2) rule of lines 64 and 74 (about |diff| <= 0.005 eV)
27
  Dependencies: SCINE Sparrow and scine_utilities Python bindings
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Give the agent the formaldehyde geometry and the method (PM3 reference, CIS-NDDO, 10 roots) and ask for the singlet and triplet excitation energies in eV; each passes when round(expected - computed, 2) == 0, the unittest assertAlmostEqual(places=2) rule of lines 64 and 74 (about |diff| <= 0.005 eV).
12
 
13
  **Licence (SPDX):** BSD-3-Clause
 
19
  - `src/Sparrow/Python/Tests/CalculatorWithReferenceTest.py`: SCINE Sparrow: formaldehyde (4 atoms, inline Angstrom geometry) with PM3 reference and CIS-NDDO; compute the 10 lowest singlet and triplet vertical excitation energies in eV.
20
  Format: Python unittest file with inline Cartesian geometry (Angstrom) and reference excitation energies (eV)
21
  Origin: Python regression test of SCINE Sparrow (tool_fixture); reference values written by the developers
 
 
22
  Check: give the agent the formaldehyde geometry and the method (PM3 reference, CIS-NDDO, 10 roots) and ask for the singlet and triplet excitation energies in eV; each passes when round(expected - computed, 2) == 0, the unittest assertAlmostEqual(places=2) rule of lines 64 and 74 (about |diff| <= 0.005 eV)
23
  Dependencies: SCINE Sparrow and scine_utilities Python bindings
campaign_files/2026-10-sc-science-reference-data/samtools-test-suite/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: expected_output_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:** Check: Run samtools stats -t 11.stats.targets 11_target.sam, drop the first 3 header lines, diff with 11.stats.expected; an agent answer can be checked field by field on the SN summary block (raw total sequences 26, reads mapped 26, and so on).
14
 
15
  **Licence (SPDX):** MIT
@@ -21,7 +19,5 @@ Files:
21
  - `test/stat/11_target.sam`: Samtools stats with a target-region file on a small multi-read-group SAM (stat test 11, expected 11.stats.expected).
22
  Format: SAM text (28 alignment records, 26 of them inside the target regions; 4 read groups) + target list + expected samtools stats report
23
  Origin: test data in the tool's own test suite (was tool_test_fixture)
24
- Reference result: expected_output_file
25
- Reference source: test/stat/11.stats.expected; test.pl runs samtools stats -t 11.stats.targets 11_target.sam | tail -n +4 and compares
26
  Check: run samtools stats -t 11.stats.targets 11_target.sam, drop the first 3 header lines, diff with 11.stats.expected; an agent answer can be checked field by field on the SN summary block (raw total sequences 26, reads mapped 26, and so on)
27
  Companions (not included): test/stat/11.stats.targets; test/stat/11.stats.expected (expected output); test/test.pl line 3442 (command)
 
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: Run samtools stats -t 11.stats.targets 11_target.sam, drop the first 3 header lines, diff with 11.stats.expected; an agent answer can be checked field by field on the SN summary block (raw total sequences 26, reads mapped 26, and so on).
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/stat/11_target.sam`: Samtools stats with a target-region file on a small multi-read-group SAM (stat test 11, expected 11.stats.expected).
20
  Format: SAM text (28 alignment records, 26 of them inside the target regions; 4 read groups) + target list + expected samtools stats report
21
  Origin: test data in the tool's own test suite (was tool_test_fixture)
 
 
22
  Check: run samtools stats -t 11.stats.targets 11_target.sam, drop the first 3 header lines, diff with 11.stats.expected; an agent answer can be checked field by field on the SN summary block (raw total sequences 26, reads mapped 26, and so on)
23
  Companions (not included): test/stat/11.stats.targets; test/stat/11.stats.expected (expected output); test/test.pl line 3442 (command)
campaign_files/2026-10-sw-tests-and-exercises-2/rsbench/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: builtin_checksum. 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: main.c line 72 reduces the hash with `vhash % 999983`; validate_and_print_results() in io.c lines 292 to 340 compares it exactly with the hard-coded value and main returns is_invalid; pass = exit code 0 and "Verification checksum: ... (Valid)".
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `openmp-threading/io.c`: RSBench OpenMP threading: multipole cross-section lookup mini-app whose verification hash must equal the hard-coded value for the method and size.
22
  Format: C source (OpenMP) + Makefile
23
  Origin: HPC proxy app with built-in verification hash (tool_fixture)
24
- Reference result: builtin_checksum
25
- Reference source: openmp-threading/io.c lines 308 to 316: hard-coded hashes (history-based: large 351485, small 879693; event-based: large 358389, small 880018)
26
  Check: main.c line 72 reduces the hash with `vhash % 999983`; validate_and_print_results() in io.c lines 292 to 340 compares it exactly with the hard-coded value and main returns is_invalid; pass = exit code 0 and "Verification checksum: ... (Valid)"
27
  Companions (not included): LICENSE; README.md; openmp-threading/Makefile; openmp-threading/init.c; openmp-threading/main.c; openmp-threading/material.c; openmp-threading/rsbench.h; openmp-threading/simulation.c; openmp-threading/utils.c
28
  Dependencies: C compiler with OpenMP; GNU make
 
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: main.c line 72 reduces the hash with `vhash % 999983`; validate_and_print_results() in io.c lines 292 to 340 compares it exactly with the hard-coded value and main returns is_invalid; pass = exit code 0 and "Verification checksum: ... (Valid)".
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `openmp-threading/io.c`: RSBench OpenMP threading: multipole cross-section lookup mini-app whose verification hash must equal the hard-coded value for the method and size.
20
  Format: C source (OpenMP) + Makefile
21
  Origin: HPC proxy app with built-in verification hash (tool_fixture)
 
 
22
  Check: main.c line 72 reduces the hash with `vhash % 999983`; validate_and_print_results() in io.c lines 292 to 340 compares it exactly with the hard-coded value and main returns is_invalid; pass = exit code 0 and "Verification checksum: ... (Valid)"
23
  Companions (not included): LICENSE; README.md; openmp-threading/Makefile; openmp-threading/init.c; openmp-threading/main.c; openmp-threading/material.c; openmp-threading/rsbench.h; openmp-threading/simulation.c; openmp-threading/utils.c
24
  Dependencies: C compiler with OpenMP; GNU make
campaign_files/2026-10-sw-tests-and-exercises-2/rust-by-practice/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained. 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: compile_and_assert. 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: Extract each ```rust block, apply the fix, compile with rustc and run; pass = it compiles and runs without panic (the assertions hold); the solution file shows one accepted fix per snippet.
14
 
15
  **Licence (SPDX):** CC-BY-4.0
@@ -21,7 +19,5 @@ Files:
21
  - `en/src/ownership/borrowing.md`: Rust by Practice, ownership/borrowing: fix the 11 Rust snippets so that each compiles and its assertions hold.
22
  Format: Markdown with 11 Rust code blocks + Markdown solutions
23
  Origin: course exercise: one page of the Rust by Practice exercise book with its solutions
24
- Reference result: compile_and_assert
25
- Reference source: solutions/ownership/borrowing.md (the published solutions); each snippet in the exercise file carries assert! lines or a must-compile requirement
26
  Check: extract each ```rust block, apply the fix, compile with rustc and run; pass = it compiles and runs without panic (the assertions hold); the solution file shows one accepted fix per snippet
27
  Companions (not included): LICENSE; solutions/ownership/borrowing.md
 
8
 
9
  **Self-contained:** The file is self-contained. 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: Extract each ```rust block, apply the fix, compile with rustc and run; pass = it compiles and runs without panic (the assertions hold); the solution file shows one accepted fix per snippet.
12
 
13
  **Licence (SPDX):** CC-BY-4.0
 
19
  - `en/src/ownership/borrowing.md`: Rust by Practice, ownership/borrowing: fix the 11 Rust snippets so that each compiles and its assertions hold.
20
  Format: Markdown with 11 Rust code blocks + Markdown solutions
21
  Origin: course exercise: one page of the Rust by Practice exercise book with its solutions
 
 
22
  Check: extract each ```rust block, apply the fix, compile with rustc and run; pass = it compiles and runs without panic (the assertions hold); the solution file shows one accepted fix per snippet
23
  Companions (not included): LICENSE; solutions/ownership/borrowing.md
campaign_files/2026-10-sw-tests-and-exercises-2/rust-latam-procedural-macros-workshop/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether the unit needs other upstream files is not stated. 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. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: test_pass. 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: Uncomment all nine lines of tests/progress.rs and run cargo test in builder/; pass = trybuild reports all nine as passing: eight compile and run, and 08 fails to compile with stderr equal to the .stderr file after trybuild normalisation. The stub panics with unimplemented!, so even 01 fails, a built-in negative check. The tests and the .stderr file must be the pinned ones (restore them before the run).
14
 
15
  **Licence (SPDX):** MIT OR Apache-2.0
@@ -21,8 +19,6 @@ Units:
21
  - `builder/` (13 files): Proc-macro-workshop builder: write derive(Builder) so that nine staged trybuild tests pass, including one expected compile error.
22
  Format: Rust proc-macro crate stub plus staged trybuild tests and an expected stderr file
23
  Origin: workshop project from the procedural macros workshop by David Tolnay; the staged tests are the answer
24
- Reference result: test_pass
25
- Reference source: the nine test files are the specification: 01 to 07 and 09 must compile and run their asserts (t.pass), 08 must fail to compile with exactly the error in 08-unrecognized-attribute.stderr (builder/tests/08-unrecognized-attribute.stderr line 1: `error: expected `builder(each = "...")``); tests/progress.rs lists them (lines 4 to 12); no reference solution in the repository
26
  Check: Uncomment all nine lines of tests/progress.rs and run cargo test in builder/; pass = trybuild reports all nine as passing: eight compile and run, and 08 fails to compile with stderr equal to the .stderr file after trybuild normalisation. The stub panics with unimplemented!, so even 01 fails, a built-in negative check. The tests and the .stderr file must be the pinned ones (restore them before the run)
27
  Companions (not included): builder/Cargo.toml; builder/tests/01-parse.rs; builder/tests/02-create-builder.rs; builder/tests/03-call-setters.rs; builder/tests/04-call-build.rs; builder/tests/05-method-chaining.rs; builder/tests/06-optional-field.rs; builder/tests/07-repeated-field.rs; builder/tests/08-unrecognized-attribute.rs; builder/tests/08-unrecognized-attribute.stderr; builder/tests/09-redefined-prelude-types.rs; builder/tests/progress.rs; README.md; Cargo.toml; LICENSE-MIT; LICENSE-APACHE
28
  Dependencies: Rust toolchain, edition 2021: builder/Cargo.toml line 5: `edition = "2021"`; trybuild >= 1.0.108 with feature diff: builder/Cargo.toml line 16: `trybuild = { version = "1.0.108", features = ["diff"] }`; syn, quote and proc-macro2 are expected by the README but not declared (the stub leaves `# TODO` under [dependencies], builder/Cargo.toml lines 18 to 19).
 
8
 
9
  **Self-contained:** Whether the unit needs other upstream files is not stated. 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. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Uncomment all nine lines of tests/progress.rs and run cargo test in builder/; pass = trybuild reports all nine as passing: eight compile and run, and 08 fails to compile with stderr equal to the .stderr file after trybuild normalisation. The stub panics with unimplemented!, so even 01 fails, a built-in negative check. The tests and the .stderr file must be the pinned ones (restore them before the run).
12
 
13
  **Licence (SPDX):** MIT OR Apache-2.0
 
19
  - `builder/` (13 files): Proc-macro-workshop builder: write derive(Builder) so that nine staged trybuild tests pass, including one expected compile error.
20
  Format: Rust proc-macro crate stub plus staged trybuild tests and an expected stderr file
21
  Origin: workshop project from the procedural macros workshop by David Tolnay; the staged tests are the answer
 
 
22
  Check: Uncomment all nine lines of tests/progress.rs and run cargo test in builder/; pass = trybuild reports all nine as passing: eight compile and run, and 08 fails to compile with stderr equal to the .stderr file after trybuild normalisation. The stub panics with unimplemented!, so even 01 fails, a built-in negative check. The tests and the .stderr file must be the pinned ones (restore them before the run)
23
  Companions (not included): builder/Cargo.toml; builder/tests/01-parse.rs; builder/tests/02-create-builder.rs; builder/tests/03-call-setters.rs; builder/tests/04-call-build.rs; builder/tests/05-method-chaining.rs; builder/tests/06-optional-field.rs; builder/tests/07-repeated-field.rs; builder/tests/08-unrecognized-attribute.rs; builder/tests/08-unrecognized-attribute.stderr; builder/tests/09-redefined-prelude-types.rs; builder/tests/progress.rs; README.md; Cargo.toml; LICENSE-MIT; LICENSE-APACHE
24
  Dependencies: Rust toolchain, edition 2021: builder/Cargo.toml line 5: `edition = "2021"`; trybuild >= 1.0.108 with feature diff: builder/Cargo.toml line 16: `trybuild = { version = "1.0.108", features = ["diff"] }`; syn, quote and proc-macro2 are expected by the README but not declared (the stub leaves `# TODO` under [dependencies], builder/Cargo.toml lines 18 to 19).