Battam commited on
Commit
aba697c
·
verified ·
1 Parent(s): 37a0103

Release [run release-20261116T092102Z]

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 +8 -0
  2. README.md +18 -18
  3. README.zh.md +18 -18
  4. campaign_files/2026-10-g-solver-instances-2/asiekierka-free-game-levels/README.md +2 -2
  5. campaign_files/2026-10-gt-browser-tests-2/allenporter-ical-recurring-single-change/README.md +2 -2
  6. campaign_files/2026-10-gt-browser-tests-2/apache-poi-formula-eval-testdata/README.md +1 -1
  7. campaign_files/2026-10-gt-browser-tests-2/apache-poi-function-testdata/README.md +1 -1
  8. campaign_files/2026-10-m-issue-trackers-2/academy-bus-nj-gtfs/README.md +2 -2
  9. campaign_files/2026-10-m-issue-trackers-2/agl-gerrit-changes/README.md +1 -1
  10. campaign_files/2026-10-m-issue-trackers-2/altlinux-devel/README.md +1 -1
  11. campaign_files/2026-10-m-issue-trackers-2/aomedia-gerrit-changes/README.md +1 -1
  12. campaign_files/2026-10-m-issue-trackers-2/aosp-gerrit-changes/README.md +1 -1
  13. campaign_files/2026-10-m-issue-trackers-2/arin-arin-ppml/README.md +1 -1
  14. campaign_files/2026-10-m-issue-trackers-2/atlassian-public-jira/README.md +1 -1
  15. campaign_files/2026-10-m-issue-trackers-2/atoc-gtfs-2021-archive/README.md +2 -2
  16. campaign_files/2026-10-md-media-and-documents/alumik-nkthesis-testfiles/README.md +2 -2
  17. campaign_files/2026-10-md-media-and-documents/ancuber-cubst-tests/README.md +2 -2
  18. campaign_files/2026-10-md-media-and-documents/annabel-lee-wikisource-reading/README.md +2 -2
  19. campaign_files/2026-10-md-media-and-documents/apache-xmlgraphics-fop-fop-test-layoutengine-standard-testcases/README.md +2 -2
  20. campaign_files/2026-10-md-media-and-documents/apollo-13-houston-we-have-had-a-problem/README.md +2 -2
  21. campaign_files/2026-10-md-media-and-documents/apriltag-test-data/README.md +2 -2
  22. campaign_files/2026-10-md-media-and-documents/arnog-mathlive-test/README.md +2 -2
  23. campaign_files/2026-10-md-media-and-documents/asciidoc-py-tests-oldtables/README.md +2 -2
  24. campaign_files/2026-10-md-media-and-documents/asciidoc-tck/README.md +2 -2
  25. campaign_files/2026-10-md-media-and-documents/asciidoctor-html5s/README.md +2 -2
  26. campaign_files/2026-10-md-media-and-documents/audiomnist/README.md +2 -2
  27. campaign_files/2026-10-md-media-and-documents/bach-chorale-figured-bass/README.md +1 -1
  28. campaign_files/2026-10-ml-serving-configs-2/aiops2018-kpi/README.md +2 -2
  29. campaign_files/2026-10-ml-serving-configs-2/arm-optimized-routines-directed-tests/README.md +2 -2
  30. campaign_files/2026-10-op-rule-calculations-2/accordproject-cicero-template-library/README.md +10 -10
  31. campaign_files/2026-10-op-rule-calculations-2/adventure-works-oltp-sample/README.md +2 -2
  32. campaign_files/2026-10-op-rule-calculations-2/apra-adi-centralised-publication/README.md +2 -2
  33. campaign_files/2026-10-op-rule-calculations-2/apra-phi-medical-gap-2026q2/README.md +1 -1
  34. campaign_files/2026-10-op-rule-calculations-2/apra-phi-membership-and-benefits-2026q2/README.md +1 -1
  35. campaign_files/2026-10-op-rule-calculations-2/apra-phi-membership-coverage-2026q2/README.md +1 -1
  36. campaign_files/2026-10-op-rule-calculations-2/apra-phi-membership-trends-2026q2/README.md +1 -1
  37. campaign_files/2026-10-op-rule-calculations-2/atticus-acord/README.md +2 -2
  38. campaign_files/2026-10-s-security-detection-2/0xdea-semgrep-rules/README.md +2 -2
  39. campaign_files/2026-10-s-security-detection-2/aqua-trivy-checks/README.md +2 -2
  40. campaign_files/2026-10-s-security-detection-2/atlas-attack-investigation-logs/README.md +2 -2
  41. campaign_files/2026-10-s-security-detection/aquasecurity-trivy-checks/README.md +2 -2
  42. campaign_files/2026-10-sc-science-reference-data-2/abacus-integrate-tests/README.md +2 -2
  43. campaign_files/2026-10-sc-science-reference-data-2/acwf-verification/README.md +2 -2
  44. campaign_files/2026-10-sc-science-reference-data-2/agat-tests/README.md +2 -2
  45. campaign_files/2026-10-sc-science-reference-data-2/bagel-test-inputs-and-unit-tests/README.md +2 -2
  46. campaign_files/2026-10-sc-science-reference-data-2/bcftools-test-pl/README.md +2 -2
  47. campaign_files/2026-10-sc-science-reference-data-2/bertrand/README.md +2 -2
  48. campaign_files/2026-10-sc-science-reference-data/bedtools2-test-suite/README.md +2 -2
  49. campaign_files/2026-10-sw-tests-and-exercises-2/1brc-test-samples/README.md +2 -2
  50. campaign_files/2026-10-sw-tests-and-exercises-2/activity-schema-aggregate-in-between-1/README.md +2 -2
CHANGELOG.md CHANGED
@@ -4639,3 +4639,11 @@ Changes to the content of this index, oldest first. There is one entry for each
4639
  - mitre-cwe-249de311: one source, one upstream: mitre-cwe-249de311: no declared upstream (upstream_kind dataset or catalog, upstream_agent and upstream_entry: a corpus's row of catalog/sources.tsv, a material set's row of catalog/upstreams.tsv)
4640
  - software-packages: one source, one upstream: https://pypi.org/simple/: one upstream_entry split by 2 sources of the layer corpus whose licenses do not tell them apart (pypi, software-packages)
4641
  - zenodo-workbooks-cc-by: 2 new name hit(s)
 
 
 
 
 
 
 
 
 
4639
  - mitre-cwe-249de311: one source, one upstream: mitre-cwe-249de311: no declared upstream (upstream_kind dataset or catalog, upstream_agent and upstream_entry: a corpus's row of catalog/sources.tsv, a material set's row of catalog/upstreams.tsv)
4640
  - software-packages: one source, one upstream: https://pypi.org/simple/: one upstream_entry split by 2 sources of the layer corpus whose licenses do not tell them apart (pypi, software-packages)
4641
  - zenodo-workbooks-cc-by: 2 new name hit(s)
4642
+
4643
+ ## 2026-11-16: sources held back
4644
+
4645
+ - Each source below keeps its newest version that passed every check; its newer batch waits until it passes them:
4646
+ - crypto-library-trees-2856d04b: one source, one upstream: crypto-library-trees-2856d04b: several upstreams, 5 agents and 5 entries (https://github.com/google/boringssl, https://github.com/openssl/openssl, https://github.com/pyca/cryptography, https://github.com/jedisct1/libsodium, https://github.com/C2SP/wycheproof)
4647
+ - mitre-cwe-249de311: one source, one upstream: mitre-cwe-249de311: no declared upstream (upstream_kind dataset or catalog, upstream_agent and upstream_entry: a corpus's row of catalog/sources.tsv, a material set's row of catalog/upstreams.tsv)
4648
+ - software-packages: one source, one upstream: https://pypi.org/simple/: one upstream_entry split by 2 sources of the layer corpus whose licenses do not tell them apart (pypi, software-packages)
4649
+ - zenodo-workbooks-cc-by: 2 new name hit(s)
README.md CHANGED
@@ -35,18 +35,18 @@ this dataset is the foundation's public index.
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-2026-11-16.3/docs/figures/stages-narrow-light.svg">
39
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/stages-wide-light.svg" alt="What each training stage can use in this release. Pre-training (provisional): 30.7 TB deduplicated size, 151,161,846 rows, 30.0 TB allowed for training. Mid-training (provisional): 3.87 TB deduplicated size, 34,242,743 rows, 1.42 TB allowed for training. Post-training: 1,769 material sets, from 1,942 upstream sources, 1,415 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-2026-11-16.3/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-2026-11-16.3/docs/figures/stages-wide-dark.svg" alt="What each training stage can use in this release. Pre-training (provisional): 30.7 TB deduplicated size, 151,161,846 rows, 30.0 TB allowed for training. Mid-training (provisional): 3.87 TB deduplicated size, 34,242,743 rows, 1.42 TB allowed for training. Post-training: 1,769 material sets, from 1,942 upstream sources, 1,415 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-2026-11-16.3/tools/fetch.py --kind material_set --form spreadsheets --out ./data
50
  ```
51
 
52
  This fetches every spreadsheet material set whole into `./data`, one folder per unit, and checks each file against the size and hashes its row records.
@@ -57,7 +57,7 @@ the same command takes the same units of this release.
57
  <details>
58
  <summary>What the command takes, and in which order</summary>
59
 
60
- The command and the queries below are pinned to the release tag `release-2026-11-16.3`.
61
  [uv](https://docs.astral.sh/uv/) reads the script's dependencies, at the versions published before the day of this release.
62
 
63
  The command returns units: every unit that passes the default profile of each material set whose form is `spreadsheets`, each in a folder of its own named after its `unit_id`.
@@ -85,7 +85,7 @@ Choose material sets from the catalog with [DuckDB](https://duckdb.org/):
85
 
86
  ```sql
87
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
88
- FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-16.3/meta/catalog.parquet'
89
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
90
  ORDER BY collection, units DESC;
91
  ```
@@ -96,7 +96,7 @@ with `kind = 'corpus'` it returns corpora instead, counted in rows (`files`).
96
  Choose files in one corpus:
97
 
98
  ```sql
99
- SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-16.3/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-16.3/meta/catalog.parquet' WHERE source = 'habs-haer-hals'));
100
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
101
  FROM read_parquet(getvariable('files'))
102
  WHERE source = 'habs-haer-hals' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
@@ -112,12 +112,12 @@ A data file may hold the rows of several sources, so the query keeps this one's
112
  ## What's inside
113
 
114
  <picture class="not-prose">
115
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/inside-narrow-light.svg">
116
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/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 29.9 TB; Science 531 GB; Machine learning 82.8 GB; Media 48.7 GB; Unclassified 47.0 GB; System administration 35.7 GB; Many domains 29.6 GB; Hardware 29.6 GB; Security 4.18 GB; Misc 376 MB; Software engineering 183 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 462 GB; Data science 429 GB; Hardware 177 GB; Security 13.9 GB; Unclassified 4.43 GB; Misc 3.45 GB; Software engineering 3.00 GB; Media 735 MB; Operations 13.9 MB. Post-training, material sets, by domain: Data science 348; Software engineering 286; Media 273; Operations 207; Machine learning 182; Science 175; Security 104; Hardware 87; System administration 71; Misc 36.">
117
  </picture>
118
  <picture class="not-prose">
119
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/inside-narrow-dark.svg">
120
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/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 29.9 TB; Science 531 GB; Machine learning 82.8 GB; Media 48.7 GB; Unclassified 47.0 GB; System administration 35.7 GB; Many domains 29.6 GB; Hardware 29.6 GB; Security 4.18 GB; Misc 376 MB; Software engineering 183 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 462 GB; Data science 429 GB; Hardware 177 GB; Security 13.9 GB; Unclassified 4.43 GB; Misc 3.45 GB; Software engineering 3.00 GB; Media 735 MB; Operations 13.9 MB. Post-training, material sets, by domain: Data science 348; Software engineering 286; Media 273; Operations 207; Machine learning 182; Science 175; Security 104; Hardware 87; System administration 71; Misc 36.">
121
  </picture>
122
 
123
  <details>
@@ -136,7 +136,7 @@ A data file may hold the rows of several sources, so the query keeps this one's
136
  | Media | 48.7 GB | 16,401 | 735 MB | 35,708 | 273 |
137
  | Misc | 376 MB | 2,403 | 3.45 GB | 102,989 | 36 |
138
  | Many domains | 29.6 GB | 8,274,002 | 1.00 TB | 16,195,869 | 0 |
139
- | Unclassified | 47.0 GB | 494,973 | 4.43 GB | 484,030 | 0 |
140
 
141
  Each source is counted under its main domain, the top-level domain of its first domain code.
142
  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/stage-counting)).
@@ -149,12 +149,12 @@ Every material set and every corpus, with its counts, size, license tier and log
149
  ## How it is made
150
 
151
  <picture class="not-prose">
152
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/steps-narrow-light.svg">
153
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/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-2026-11-16.3.">
154
  </picture>
155
  <picture class="not-prose">
156
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/steps-narrow-dark.svg">
157
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/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-2026-11-16.3.">
158
  </picture>
159
 
160
  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.
@@ -182,7 +182,7 @@ The full text, with the license details, personal information read from sample f
182
  @misc{training_data_foundation,
183
  title = {training-data-foundation: an index of public materials for building agent environments},
184
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
185
- note = {Release release-2026-11-16.3},
186
  year = {2026}
187
  }
188
  ```
 
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-2026-11-16.4/docs/figures/stages-narrow-light.svg">
39
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/stages-wide-light.svg" alt="What each training stage can use in this release. Pre-training (provisional): 30.7 TB deduplicated size, 151,161,846 rows, 30.0 TB allowed for training. Mid-training (provisional): 3.87 TB deduplicated size, 34,243,169 rows, 1.42 TB allowed for training. Post-training: 1,769 material sets, from 1,942 upstream sources, 1,414 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-2026-11-16.4/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-2026-11-16.4/docs/figures/stages-wide-dark.svg" alt="What each training stage can use in this release. Pre-training (provisional): 30.7 TB deduplicated size, 151,161,846 rows, 30.0 TB allowed for training. Mid-training (provisional): 3.87 TB deduplicated size, 34,243,169 rows, 1.42 TB allowed for training. Post-training: 1,769 material sets, from 1,942 upstream sources, 1,414 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-2026-11-16.4/tools/fetch.py --kind material_set --form spreadsheets --out ./data
50
  ```
51
 
52
  This fetches every spreadsheet material set whole into `./data`, one folder per unit, and checks each file against the size and hashes its row records.
 
57
  <details>
58
  <summary>What the command takes, and in which order</summary>
59
 
60
+ The command and the queries below are pinned to the release tag `release-2026-11-16.4`.
61
  [uv](https://docs.astral.sh/uv/) reads the script's dependencies, at the versions published before the day of this release.
62
 
63
  The command returns units: every unit that passes the default profile of each material set whose form is `spreadsheets`, each in a folder of its own named after its `unit_id`.
 
85
 
86
  ```sql
87
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
88
+ FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-16.4/meta/catalog.parquet'
89
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
90
  ORDER BY collection, units DESC;
91
  ```
 
96
  Choose files in one corpus:
97
 
98
  ```sql
99
+ SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-16.4/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-16.4/meta/catalog.parquet' WHERE source = 'habs-haer-hals'));
100
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
101
  FROM read_parquet(getvariable('files'))
102
  WHERE source = 'habs-haer-hals' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
 
112
  ## What's inside
113
 
114
  <picture class="not-prose">
115
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/inside-narrow-light.svg">
116
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/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 29.9 TB; Science 531 GB; Machine learning 82.8 GB; Media 48.7 GB; Unclassified 47.0 GB; System administration 35.7 GB; Many domains 29.6 GB; Hardware 29.6 GB; Security 4.18 GB; Misc 376 MB; Software engineering 183 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 462 GB; Data science 429 GB; Hardware 177 GB; Security 13.9 GB; Unclassified 4.43 GB; Misc 3.45 GB; Software engineering 3.00 GB; Media 735 MB; Operations 13.9 MB. Post-training, material sets, by domain: Data science 348; Software engineering 286; Media 273; Operations 207; Machine learning 182; Science 175; Security 104; Hardware 87; System administration 71; Misc 36.">
117
  </picture>
118
  <picture class="not-prose">
119
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/inside-narrow-dark.svg">
120
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/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 29.9 TB; Science 531 GB; Machine learning 82.8 GB; Media 48.7 GB; Unclassified 47.0 GB; System administration 35.7 GB; Many domains 29.6 GB; Hardware 29.6 GB; Security 4.18 GB; Misc 376 MB; Software engineering 183 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 462 GB; Data science 429 GB; Hardware 177 GB; Security 13.9 GB; Unclassified 4.43 GB; Misc 3.45 GB; Software engineering 3.00 GB; Media 735 MB; Operations 13.9 MB. Post-training, material sets, by domain: Data science 348; Software engineering 286; Media 273; Operations 207; Machine learning 182; Science 175; Security 104; Hardware 87; System administration 71; Misc 36.">
121
  </picture>
122
 
123
  <details>
 
136
  | Media | 48.7 GB | 16,401 | 735 MB | 35,708 | 273 |
137
  | Misc | 376 MB | 2,403 | 3.45 GB | 102,989 | 36 |
138
  | Many domains | 29.6 GB | 8,274,002 | 1.00 TB | 16,195,869 | 0 |
139
+ | Unclassified | 47.0 GB | 494,973 | 4.43 GB | 484,456 | 0 |
140
 
141
  Each source is counted under its main domain, the top-level domain of its first domain code.
142
  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/stage-counting)).
 
149
  ## How it is made
150
 
151
  <picture class="not-prose">
152
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/steps-narrow-light.svg">
153
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/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-2026-11-16.4.">
154
  </picture>
155
  <picture class="not-prose">
156
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/steps-narrow-dark.svg">
157
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/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-2026-11-16.4.">
158
  </picture>
159
 
160
  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.
 
182
  @misc{training_data_foundation,
183
  title = {training-data-foundation: an index of public materials for building agent environments},
184
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
185
+ note = {Release release-2026-11-16.4},
186
  year = {2026}
187
  }
188
  ```
README.zh.md CHANGED
@@ -22,18 +22,18 @@ tags:
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-2026-11-16.3/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-2026-11-16.3/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 30.7 TB,151,161,846 行,其中 30.0 TB 允许用于训练。中训练(暂定):去重后大小 3.87 TB,34,242,743 行,其中 1.42 TB 允许用于训练。后训练:材料集 1,769,来自 1,942 个上游来源,其中 1,415 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-2026-11-16.3/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-2026-11-16.3/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 30.7 TB,151,161,846 行,其中 30.0 TB 允许用于训练。中训练(暂定):去重后大小 3.87 TB,34,242,743 行,其中 1.42 TB 允许用于训练。后训练:材料集 1,769,来自 1,942 个上游来源,其中 1,415 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-2026-11-16.3/tools/fetch.py --kind material_set --form spreadsheets --out ./data
37
  ```
38
 
39
  这条命令把每个电子表格材料集整个取进 `./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。
@@ -44,7 +44,7 @@ uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/r
44
  <details>
45
  <summary>命令取什么、按什么顺序取</summary>
46
 
47
- 这条命令和下面的查询都固定在发布标签 `release-2026-11-16.3` 上。
48
  [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。
49
 
50
  命令返回单元:形式为 `spreadsheets` 的每个材料集中通过默认配置的全部单元,每个单元放在以它的 `unit_id` 命名的单独文件夹里。
@@ -72,7 +72,7 @@ uv run https://huggingface.co/datasets/Battam/training-data-foundation/resolve/r
72
 
73
  ```sql
74
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
75
- FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-16.3/meta/catalog.parquet'
76
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
77
  ORDER BY collection, units DESC;
78
  ```
@@ -83,7 +83,7 @@ ORDER BY collection, units DESC;
83
  在一个语料库里挑选文件:
84
 
85
  ```sql
86
- SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-16.3/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-16.3/meta/catalog.parquet' WHERE source = 'habs-haer-hals'));
87
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
88
  FROM read_parquet(getvariable('files'))
89
  WHERE source = 'habs-haer-hals' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
@@ -99,12 +99,12 @@ ORDER BY size_bytes LIMIT 20;
99
  ## 内容概览
100
 
101
  <picture class="not-prose">
102
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/inside-narrow-light.zh.svg">
103
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/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 29.9 TB;科学 531 GB;机器学习 82.8 GB;媒体 48.7 GB;未分类 47.0 GB;系统管理 35.7 GB;多领域 29.6 GB;硬件 29.6 GB;安全 4.18 GB;杂项 376 MB;软件工程 183 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 462 GB;数据科学 429 GB;硬件 177 GB;安全 13.9 GB;未分类 4.43 GB;杂项 3.45 GB;软件工程 3.00 GB;媒体 735 MB;运营 13.9 MB。后训练,材料集,按领域:数据科学 348;软件工程 286;媒体 273;运营 207;机器学习 182;科学 175;安全 104;硬件 87;系统管理 71;杂项 36。">
104
  </picture>
105
  <picture class="not-prose">
106
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/inside-narrow-dark.zh.svg">
107
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/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:数据科学 29.9 TB;科学 531 GB;机器学习 82.8 GB;媒体 48.7 GB;未分类 47.0 GB;系统管理 35.7 GB;多领域 29.6 GB;硬件 29.6 GB;安全 4.18 GB;杂项 376 MB;软件工程 183 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 462 GB;数据科学 429 GB;硬件 177 GB;安全 13.9 GB;未分类 4.43 GB;杂项 3.45 GB;软件工程 3.00 GB;媒体 735 MB;运营 13.9 MB。后训练,材料集,按领域:数据科学 348;软件工程 286;媒体 273;运营 207;机器学习 182;科学 175;安全 104;硬件 87;系统管理 71;杂项 36。">
108
  </picture>
109
 
110
  <details>
@@ -123,7 +123,7 @@ ORDER BY size_bytes LIMIT 20;
123
  | 媒体 | 48.7 GB | 16,401 | 735 MB | 35,708 | 273 |
124
  | 杂项 | 376 MB | 2,403 | 3.45 GB | 102,989 | 36 |
125
  | 多领域 | 29.6 GB | 8,274,002 | 1.00 TB | 16,195,869 | 0 |
126
- | 未分类 | 47.0 GB | 494,973 | 4.43 GB | 484,030 | 0 |
127
 
128
  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。
129
  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/stage-counting))。
@@ -136,12 +136,12 @@ ORDER BY size_bytes LIMIT 20;
136
  ## 怎样做成
137
 
138
  <picture class="not-prose">
139
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/steps-narrow-light.zh.svg">
140
- <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/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-2026-11-16.3。">
141
  </picture>
142
  <picture class="not-prose">
143
- <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/docs/figures/steps-narrow-dark.zh.svg">
144
- <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.3/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-2026-11-16.3。">
145
  </picture>
146
 
147
  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。
@@ -169,7 +169,7 @@ ORDER BY size_bytes LIMIT 20;
169
  @misc{training_data_foundation,
170
  title = {training-data-foundation: an index of public materials for building agent environments},
171
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
172
- note = {Release release-2026-11-16.3},
173
  year = {2026}
174
  }
175
  ```
 
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-2026-11-16.4/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-2026-11-16.4/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 30.7 TB,151,161,846 行,其中 30.0 TB 允许用于训练。中训练(暂定):去重后大小 3.87 TB,34,243,169 行,其中 1.42 TB 允许用于训练。后训练:材料集 1,769,来自 1,942 个上游来源,其中 1,414 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-2026-11-16.4/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-2026-11-16.4/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 30.7 TB,151,161,846 行,其中 30.0 TB 允许用于训练。中训练(暂定):去重后大小 3.87 TB,34,243,169 行,其中 1.42 TB 允许用于训练。后训练:材料集 1,769,来自 1,942 个上游来源,其中 1,414 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-2026-11-16.4/tools/fetch.py --kind material_set --form spreadsheets --out ./data
37
  ```
38
 
39
  这条命令把每个电子表格材料集整个取进 `./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。
 
44
  <details>
45
  <summary>命令取什么、按什么顺序取</summary>
46
 
47
+ 这条命令和下面的查询都固定在发布标签 `release-2026-11-16.4` 上。
48
  [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。
49
 
50
  命令返回单元:形式为 `spreadsheets` 的每个材料集中通过默认配置的全部单元,每个单元放在以它的 `unit_id` 命名的单独文件夹里。
 
72
 
73
  ```sql
74
  SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
75
+ FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-16.4/meta/catalog.parquet'
76
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
77
  ORDER BY collection, units DESC;
78
  ```
 
83
  在一个语料库里挑选文件:
84
 
85
  ```sql
86
+ SET VARIABLE files = (SELECT list('hf://datasets/Battam/training-data-foundation@release-2026-11-16.4/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/training-data-foundation@release-2026-11-16.4/meta/catalog.parquet' WHERE source = 'habs-haer-hals'));
87
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
88
  FROM read_parquet(getvariable('files'))
89
  WHERE source = 'habs-haer-hals' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
 
99
  ## 内容概览
100
 
101
  <picture class="not-prose">
102
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/inside-narrow-light.zh.svg">
103
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/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 29.9 TB;科学 531 GB;机器学习 82.8 GB;媒体 48.7 GB;未分类 47.0 GB;系统管理 35.7 GB;多领域 29.6 GB;硬件 29.6 GB;安全 4.18 GB;杂项 376 MB;软件工程 183 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 462 GB;数据科学 429 GB;硬件 177 GB;安全 13.9 GB;未分类 4.43 GB;杂项 3.45 GB;软件工程 3.00 GB;媒体 735 MB;运营 13.9 MB。后训练,材料集,按领域:数据科学 348;软件工程 286;媒体 273;运营 207;机器学习 182;科学 175;安全 104;硬件 87;系统管理 71;杂项 36。">
104
  </picture>
105
  <picture class="not-prose">
106
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/inside-narrow-dark.zh.svg">
107
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/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:数据科学 29.9 TB;科学 531 GB;机器学习 82.8 GB;媒体 48.7 GB;未分类 47.0 GB;系统管理 35.7 GB;多领域 29.6 GB;硬件 29.6 GB;安全 4.18 GB;杂项 376 MB;软件工程 183 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 462 GB;数��科学 429 GB;硬件 177 GB;安全 13.9 GB;未分类 4.43 GB;杂项 3.45 GB;软件工程 3.00 GB;媒体 735 MB;运营 13.9 MB。后训练,材料集,按领域:数据科学 348;软件工程 286;媒体 273;运营 207;机器学习 182;科学 175;安全 104;硬件 87;系统管理 71;杂项 36。">
108
  </picture>
109
 
110
  <details>
 
123
  | 媒体 | 48.7 GB | 16,401 | 735 MB | 35,708 | 273 |
124
  | 杂项 | 376 MB | 2,403 | 3.45 GB | 102,989 | 36 |
125
  | 多领域 | 29.6 GB | 8,274,002 | 1.00 TB | 16,195,869 | 0 |
126
+ | 未分类 | 47.0 GB | 494,973 | 4.43 GB | 484,456 | 0 |
127
 
128
  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。
129
  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/stage-counting))。
 
136
  ## 怎样做成
137
 
138
  <picture class="not-prose">
139
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/steps-narrow-light.zh.svg">
140
+ <img class="not-prose dark:hidden" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/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-2026-11-16.4。">
141
  </picture>
142
  <picture class="not-prose">
143
+ <source media="(max-width: 1023px)" srcset="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/docs/figures/steps-narrow-dark.zh.svg">
144
+ <img class="not-prose hidden dark:block" src="https://huggingface.co/datasets/Battam/training-data-foundation/resolve/release-2026-11-16.4/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-2026-11-16.4。">
145
  </picture>
146
 
147
  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。
 
169
  @misc{training_data_foundation,
170
  title = {training-data-foundation: an index of public materials for building agent environments},
171
  howpublished = {https://huggingface.co/datasets/Battam/training-data-foundation},
172
+ note = {Release release-2026-11-16.4},
173
  year = {2026}
174
  }
175
  ```
campaign_files/2026-10-g-solver-instances-2/asiekierka-free-game-levels/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # asiekierka-free-game-levels
2
 
3
- **What this is:** 19 files published by David W. Skinner (author); mirrored by asiekierka (Adrian Siekierka) in free-game-levels; David Holland (author); mirrored by asiekierka (Adrian Siekierka) in free-game-levels; Evgeny Grigoriev (author, with some levels by his friend Mottled); mirrored by asiekierka (Adrian Siekierka) in free-game-levels; Rick Sladkey (author); mirrored by asiekierka (Adrian Siekierka) in free-game-levels. Collection theme: Public puzzle sets that ship with solutions (sudoku, nonogram, logic puzzle collections).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. The format of each file is in the notes.
6
 
7
  **Where it comes from:** Upstream: https://github.com/asiekierka/free-game-levels. Kind: real files published upstream. Details per file in the notes.
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 19 of 19 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: Correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
12
 
 
1
  # asiekierka-free-game-levels
2
 
3
+ **What this is:** Files published by David W. Skinner (author); mirrored by asiekierka (Adrian Siekierka) in free-game-levels; David Holland (author); mirrored by asiekierka (Adrian Siekierka) in free-game-levels; Evgeny Grigoriev (author, with some levels by his friend Mottled); mirrored by asiekierka (Adrian Siekierka) in free-game-levels; Rick Sladkey (author); mirrored by asiekierka (Adrian Siekierka) in free-game-levels. Collection theme: Public puzzle sets that ship with solutions (sudoku, nonogram, logic puzzle collections).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. The format of each file is in the notes.
6
 
7
  **Where it comes from:** Upstream: https://github.com/asiekierka/free-game-levels. Kind: real files published upstream. Details per file in the notes.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
12
 
campaign_files/2026-10-gt-browser-tests-2/allenporter-ical-recurring-single-change/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # allenporter-ical-recurring-single-change
2
 
3
- **What this is:** 1 file published by Allen Porter (allenporter/ical). Collection theme: Mail and calendar clients: test corpora of storage formats such as mbox, Maildir and iCalendar and of their clients.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: iCalendar (.ics) file; expected values in a pytest file.
6
 
7
  **Where it comes from:** Upstream: https://github.com/allenporter/ical. Kind: test or example files shipped with a tool. allenporter/ical example test fixture (tool_fixture).
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:** Python ical (IcsCalendarStream.calendar_from_ics, calendar.timeline) or icalendar plus recurring-ical-events; a script expands the series and writes the first occurrences (summary, recurrence id, start) to a CSV or JSON file
12
 
 
1
  # allenporter-ical-recurring-single-change
2
 
3
+ **What this is:** Files published by Allen Porter (allenporter/ical). Collection theme: Mail and calendar clients: test corpora of storage formats such as mbox, Maildir and iCalendar and of their clients.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: iCalendar (.ics) file; expected values in a pytest file.
6
 
7
  **Where it comes from:** Upstream: https://github.com/allenporter/ical. Kind: test or example files shipped with a tool. allenporter/ical example test fixture (tool_fixture).
8
 
9
+ **Self-contained:** The file is self-contained. Each file names 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:** Python ical (IcsCalendarStream.calendar_from_ics, calendar.timeline) or icalendar plus recurring-ical-events; a script expands the series and writes the first occurrences (summary, recurrence id, start) to a CSV or JSON file
12
 
campaign_files/2026-10-gt-browser-tests-2/apache-poi-formula-eval-testdata/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # apache-poi-formula-eval-testdata
2
 
3
- **What this is:** 1 file published by Apache Software Foundation (apache/poi). Collection theme: File format library test corpora (Apache POI, openpyxl, python-docx, odfpy and the like): tasks operate on documents usually opened in a GUI, answers come from the tests; cross-checked against line A real workbooks.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Excel 97-2003 workbook (.xls, BIFF8).
6
 
 
1
  # apache-poi-formula-eval-testdata
2
 
3
+ **What this is:** Files published by Apache Software Foundation (apache/poi). Collection theme: File format library test corpora (Apache POI, openpyxl, python-docx, odfpy and the like): tasks operate on documents usually opened in a GUI, answers come from the tests; cross-checked against line A real workbooks.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Excel 97-2003 workbook (.xls, BIFF8).
6
 
campaign_files/2026-10-gt-browser-tests-2/apache-poi-function-testdata/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # apache-poi-function-testdata
2
 
3
- **What this is:** 21 files published by Apache Software Foundation (apache/poi). Collection theme: File format library test corpora (Apache POI, openpyxl, python-docx, odfpy and the like): tasks operate on documents usually opened in a GUI, answers come from the tests; cross-checked against line A real workbooks.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Excel 97-2003 workbook (.xls, BIFF8).
6
 
 
1
  # apache-poi-function-testdata
2
 
3
+ **What this is:** Files published by Apache Software Foundation (apache/poi). Collection theme: File format library test corpora (Apache POI, openpyxl, python-docx, odfpy and the like): tasks operate on documents usually opened in a GUI, answers come from the tests; cross-checked against line A real workbooks.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Excel 97-2003 workbook (.xls, BIFF8).
6
 
campaign_files/2026-10-m-issue-trackers-2/academy-bus-nj-gtfs/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # academy-bus-nj-gtfs
2
 
3
- **What this is:** 1 unit (1 file) published by Academy Bus, published on NJ TRANSIT developer resources.
4
 
5
  **What one unit is:** Each unit is one file read out of an upstream archive (the archive's address is in the notes), kept at its upstream path. A file whose upstream path has no folder is kept in a folder named after its unit. Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt).
6
 
7
  **Where it comes from:** Upstream: https://www.njtransit.com/Academy_bus_data.zip. Kind: real files published upstream. Live static GTFS feed published by Academy Bus, published on NJ TRANSIT developer resources at https://www.njtransit.com/Academy_bus_data.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.
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
 
 
1
  # academy-bus-nj-gtfs
2
 
3
+ **What this is:** Units published by Academy Bus, published on NJ TRANSIT developer resources.
4
 
5
  **What one unit is:** Each unit is one file read out of an upstream archive (the archive's address is in the notes), kept at its upstream path. A file whose upstream path has no folder is kept in a folder named after its unit. Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt).
6
 
7
  **Where it comes from:** Upstream: https://www.njtransit.com/Academy_bus_data.zip. Kind: real files published upstream. Live static GTFS feed published by Academy Bus, published on NJ TRANSIT developer resources at https://www.njtransit.com/Academy_bus_data.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.
8
 
9
+ **Self-contained:** The unit is self-contained. Each unit names 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
 
campaign_files/2026-10-m-issue-trackers-2/agl-gerrit-changes/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # agl-gerrit-changes
2
 
3
- **What this is:** 1 file published by Automotive Grade Linux (Linux Foundation), gerrit.automotivelinux.org. Collection theme: Code hosting platform issue and merge request timelines (GitHub, GitLab, Gitea/Forgejo and similar, as fixed exports or datasets); Gerrit code review changes count as this family.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
6
 
 
1
  # agl-gerrit-changes
2
 
3
+ **What this is:** Files published by Automotive Grade Linux (Linux Foundation), gerrit.automotivelinux.org. Collection theme: Code hosting platform issue and merge request timelines (GitHub, GitLab, Gitea/Forgejo and similar, as fixed exports or datasets); Gerrit code review changes count as this family.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
6
 
campaign_files/2026-10-m-issue-trackers-2/altlinux-devel/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # altlinux-devel
2
 
3
- **What this is:** 1 file published by ALT Linux Team (lists.altlinux.org Mailman, devel list). Collection theme: Public mailing-list archives (Mailman, Pipermail, HyperKitty, public-inbox; taken per thread).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. A file whose upstream path has no folder is kept in a folder named after its unit. 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).
6
 
 
1
  # altlinux-devel
2
 
3
+ **What this is:** Files published by ALT Linux Team (lists.altlinux.org Mailman, devel list). Collection theme: Public mailing-list archives (Mailman, Pipermail, HyperKitty, public-inbox; taken per thread).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. A file whose upstream path has no folder is kept in a folder named after its unit. 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).
6
 
campaign_files/2026-10-m-issue-trackers-2/aomedia-gerrit-changes/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # aomedia-gerrit-changes
2
 
3
- **What this is:** 1 file published by Alliance for Open Media, aomedia-review.googlesource.com. Collection theme: Code hosting platform issue and merge request timelines (GitHub, GitLab, Gitea/Forgejo and similar, as fixed exports or datasets); Gerrit code review changes count as this family.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote, messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
6
 
 
1
  # aomedia-gerrit-changes
2
 
3
+ **What this is:** Files published by Alliance for Open Media, aomedia-review.googlesource.com. Collection theme: Code hosting platform issue and merge request timelines (GitHub, GitLab, Gitea/Forgejo and similar, as fixed exports or datasets); Gerrit code review changes count as this family.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote, messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
6
 
campaign_files/2026-10-m-issue-trackers-2/aosp-gerrit-changes/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # aosp-gerrit-changes
2
 
3
- **What this is:** 1 file published by Android Open Source Project (Google), android-review.googlesource.com. Collection theme: Code hosting platform issue and merge request timelines (GitHub, GitLab, Gitea/Forgejo and similar, as fixed exports or datasets); Gerrit code review changes count as this family.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
6
 
 
1
  # aosp-gerrit-changes
2
 
3
+ **What this is:** Files published by Android Open Source Project (Google), android-review.googlesource.com. Collection theme: Code hosting platform issue and merge request timelines (GitHub, GitLab, Gitea/Forgejo and similar, as fixed exports or datasets); Gerrit code review changes count as this family.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
6
 
campaign_files/2026-10-m-issue-trackers-2/arin-arin-ppml/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # arin-arin-ppml
2
 
3
- **What this is:** 1 file published by American Registry for Internet Numbers (lists.arin.net Mailman, ARIN-PPML public policy list). Collection theme: Public mailing-list archives (Mailman, Pipermail, HyperKitty, public-inbox; taken per thread).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. A file whose upstream path has no folder is kept in a folder named after its unit. 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).
6
 
 
1
  # arin-arin-ppml
2
 
3
+ **What this is:** Files published by American Registry for Internet Numbers (lists.arin.net Mailman, ARIN-PPML public policy list). Collection theme: Public mailing-list archives (Mailman, Pipermail, HyperKitty, public-inbox; taken per thread).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. A file whose upstream path has no folder is kept in a folder named after its unit. 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).
6
 
campaign_files/2026-10-m-issue-trackers-2/atlassian-public-jira/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # atlassian-public-jira
2
 
3
- **What this is:** 3 files published by Atlassian's public Jira Data Center site jira.atlassian.com, project Confluence Cloud (CONFCLOUD); Atlassian's public Jira Data Center site jira.atlassian.com, project Confluence Data Center (CONFSERVER); Atlassian's public Jira Data Center site jira.atlassian.com, project Jira Data Center (JRASERVER). Collection theme: Self-hosted issue trackers (Bugzilla, Jira, Trac, Redmine, Launchpad and similar, with status transitions, comments and assignment history).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Jira REST API v2 issue JSON with expand=changelog: fields (status, resolution, resolutiondate, assignee, reporter, comment.comments[]) and changelog.histories[] (author, created, items[] with field, fromString, toString).
6
 
 
1
  # atlassian-public-jira
2
 
3
+ **What this is:** Files published by Atlassian's public Jira Data Center site jira.atlassian.com, project Confluence Cloud (CONFCLOUD); Atlassian's public Jira Data Center site jira.atlassian.com, project Confluence Data Center (CONFSERVER); Atlassian's public Jira Data Center site jira.atlassian.com, project Jira Data Center (JRASERVER). Collection theme: Self-hosted issue trackers (Bugzilla, Jira, Trac, Redmine, Launchpad and similar, with status transitions, comments and assignment history).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Jira REST API v2 issue JSON with expand=changelog: fields (status, resolution, resolutiondate, assignee, reporter, comment.comments[]) and changelog.histories[] (author, created, items[] with field, fromString, toString).
6
 
campaign_files/2026-10-m-issue-trackers-2/atoc-gtfs-2021-archive/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # atoc-gtfs-2021-archive
2
 
3
- **What this is:** 1 unit (1 file) published by Association of Train Operating Companies timetable data, converted to GTFS and archived by OpenMobilityData (Transitfeeds), now held by MobilityData.
4
 
5
  **What one unit is:** Each unit is one file read out of an upstream archive (the archive's address is in the notes), kept at its upstream path. A file whose upstream path has no folder is kept in a folder named after its unit. Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt).
6
 
7
  **Where it comes from:** Upstream: https://openmobilitydata-data.s3.us-west-1.amazonaws.com/public/feeds/association-of-train-operating-companies/284/20210423/gtfs.zip. Kind: real files published upstream. Live static GTFS feed published by Association of Train Operating Companies timetable data, converted to GTFS and archived by OpenMobilityData (Transitfeeds), now held by MobilityData at https://openmobilitydata-data.s3.us-west-1.amazonaws.com/public/feeds/association-of-train-operating-companies/284/20210423/gtfs.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
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
 
 
1
  # atoc-gtfs-2021-archive
2
 
3
+ **What this is:** Units published by Association of Train Operating Companies timetable data, converted to GTFS and archived by OpenMobilityData (Transitfeeds), now held by MobilityData.
4
 
5
  **What one unit is:** Each unit is one file read out of an upstream archive (the archive's address is in the notes), kept at its upstream path. A file whose upstream path has no folder is kept in a folder named after its unit. Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt).
6
 
7
  **Where it comes from:** Upstream: https://openmobilitydata-data.s3.us-west-1.amazonaws.com/public/feeds/association-of-train-operating-companies/284/20210423/gtfs.zip. Kind: real files published upstream. Live static GTFS feed published by Association of Train Operating Companies timetable data, converted to GTFS and archived by OpenMobilityData (Transitfeeds), now held by MobilityData at https://openmobilitydata-data.s3.us-west-1.amazonaws.com/public/feeds/association-of-train-operating-companies/284/20210423/gtfs.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
8
 
9
+ **Self-contained:** The unit is self-contained. Each unit names 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
 
campaign_files/2026-10-md-media-and-documents/alumik-nkthesis-testfiles/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # alumik-nkthesis-testfiles
2
 
3
- **What this is:** 1 file published by Zhenyu Zhong (alumik), github.com/alumik/nkthesis. Collection theme: Typesetting engine sources with reference renderings (LaTeX, Typst and similar regression tests).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: LaTeX PDF test file (.pvt) plus expected normalised PDF (.tpf).
6
 
7
  **Where it comes from:** Upstream name: alumik/nkthesis testfiles/ (l3build PDF-based regression tests of a thesis class). Upstream: https://github.com/alumik/nkthesis. Kind: test or example files shipped with a tool. Regression test of the class author; headings-en.tpf is the accepted PDF with binary streams normalised by l3build.
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:** l3build check headings-en (headless XeLaTeX run, two passes, PDF normalised and diffed against headings-en.tpf)
12
 
 
1
  # alumik-nkthesis-testfiles
2
 
3
+ **What this is:** Files published by Zhenyu Zhong (alumik), github.com/alumik/nkthesis. Collection theme: Typesetting engine sources with reference renderings (LaTeX, Typst and similar regression tests).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: LaTeX PDF test file (.pvt) plus expected normalised PDF (.tpf).
6
 
7
  **Where it comes from:** Upstream name: alumik/nkthesis testfiles/ (l3build PDF-based regression tests of a thesis class). Upstream: https://github.com/alumik/nkthesis. Kind: test or example files shipped with a tool. Regression test of the class author; headings-en.tpf is the accepted PDF with binary streams normalised by l3build.
8
 
9
+ **Self-contained:** Whether the file needs other upstream files is not stated. Each file names 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:** l3build check headings-en (headless XeLaTeX run, two passes, PDF normalised and diffed against headings-en.tpf)
12
 
campaign_files/2026-10-md-media-and-documents/ancuber-cubst-tests/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # ancuber-cubst-tests
2
 
3
- **What this is:** 1 file published by Anson Lai (ANCuber), github.com/ANCuber/cubst. Collection theme: Typesetting engine sources with reference renderings (LaTeX, Typst and similar regression tests).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Typst source (.typ) plus reference PNG.
6
 
7
  **Where it comes from:** Upstream name: ANCuber/cubst tests/ (tytanic regression tests of a Typst package for twisty-puzzle diagrams). Upstream: https://github.com/ANCuber/cubst. Kind: test or example files shipped with a tool. Regression test of the package author; ref/1.png is the accepted rendering, made with Typst 0.15 and tytanic 0.4 as the CI file says.
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:** just test, that is tt run --no-fail-fast --max-deviations 10 (tytanic, headless Typst compile and pixel compare)
12
 
 
1
  # ancuber-cubst-tests
2
 
3
+ **What this is:** Files published by Anson Lai (ANCuber), github.com/ANCuber/cubst. Collection theme: Typesetting engine sources with reference renderings (LaTeX, Typst and similar regression tests).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Typst source (.typ) plus reference PNG.
6
 
7
  **Where it comes from:** Upstream name: ANCuber/cubst tests/ (tytanic regression tests of a Typst package for twisty-puzzle diagrams). Upstream: https://github.com/ANCuber/cubst. Kind: test or example files shipped with a tool. Regression test of the package author; ref/1.png is the accepted rendering, made with Typst 0.15 and tytanic 0.4 as the CI file says.
8
 
9
+ **Self-contained:** Whether the file needs other upstream files is not stated. Each file names 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:** just test, that is tt run --no-fail-fast --max-deviations 10 (tytanic, headless Typst compile and pixel compare)
12
 
campaign_files/2026-10-md-media-and-documents/annabel-lee-wikisource-reading/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # annabel-lee-wikisource-reading
2
 
3
- **What this is:** 1 file published by Wikimedia Commons (volunteer reading uploaded by PopularOutcast); text on English Wikisource. Collection theme: Human-transcribed English audio (read speech, lectures, broadcasts with human transcripts).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Ogg Vorbis, 1,361,520 bytes, one reading of the poem.
6
 
7
  **Where it comes from:** Upstream: https://commons.wikimedia.org/wiki/File:Annabel_Lee.ogg. Kind: real files published upstream. Modern volunteer reading of a public domain poem.
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: Give the agent the audio file only; it writes the transcript; score word error rate (WER) against the reference words, reported as a number with no pass mark; the Wikisource page furniture (header, notes, licence box, footer) is removed and only the work's text is the reference; case and punctuation folded on both sides; the versions page lists several printings that differ in a few words; it is not recorded which one the reader used, so the score is a soft reference score.
12
 
 
1
  # annabel-lee-wikisource-reading
2
 
3
+ **What this is:** Files published by Wikimedia Commons (volunteer reading uploaded by PopularOutcast); text on English Wikisource. Collection theme: Human-transcribed English audio (read speech, lectures, broadcasts with human transcripts).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Ogg Vorbis, 1,361,520 bytes, one reading of the poem.
6
 
7
  **Where it comes from:** Upstream: https://commons.wikimedia.org/wiki/File:Annabel_Lee.ogg. Kind: real files published upstream. Modern volunteer reading of a public domain poem.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Give the agent the audio file only; it writes the transcript; score word error rate (WER) against the reference words, reported as a number with no pass mark; the Wikisource page furniture (header, notes, licence box, footer) is removed and only the work's text is the reference; case and punctuation folded on both sides; the versions page lists several printings that differ in a few words; it is not recorded which one the reader used, so the score is a soft reference score.
12
 
campaign_files/2026-10-md-media-and-documents/apache-xmlgraphics-fop-fop-test-layoutengine-standard-testcases/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # apache-xmlgraphics-fop-fop-test-layoutengine-standard-testcases
2
 
3
- **What this is:** 1 file published by The Apache Software Foundation (Apache XML Graphics FOP). Collection theme: Typesetting engine sources with reference renderings (LaTeX, Typst and similar regression tests).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: FOP testcase XML (XSL-FO plus checks).
6
 
7
  **Where it comes from:** Upstream name: apache/xmlgraphics-fop fop/test/layoutengine/standard-testcases. Upstream: https://github.com/apache/xmlgraphics-fop. Kind: test or example files shipped with a tool. Layout engine testcase of Apache FOP (XSL-FO to PDF formatter), written by the FOP developers; each file holds the FO input and the expected area tree checks.
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:** fop -fo block.fo -at area.xml (FOP command line, area tree output), then evaluate the XPath checks; in the source tree the LayoutEngineTestCase JUnit suite does both
12
 
 
1
  # apache-xmlgraphics-fop-fop-test-layoutengine-standard-testcases
2
 
3
+ **What this is:** Files published by The Apache Software Foundation (Apache XML Graphics FOP). Collection theme: Typesetting engine sources with reference renderings (LaTeX, Typst and similar regression tests).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: FOP testcase XML (XSL-FO plus checks).
6
 
7
  **Where it comes from:** Upstream name: apache/xmlgraphics-fop fop/test/layoutengine/standard-testcases. Upstream: https://github.com/apache/xmlgraphics-fop. Kind: test or example files shipped with a tool. Layout engine testcase of Apache FOP (XSL-FO to PDF formatter), written by the FOP developers; each file holds the FO input and the expected area tree checks.
8
 
9
+ **Self-contained:** Whether the file needs other upstream files is not stated. Each file names 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:** fop -fo block.fo -at area.xml (FOP command line, area tree output), then evaluate the XPath checks; in the source tree the LayoutEngineTestCase JUnit suite does both
12
 
campaign_files/2026-10-md-media-and-documents/apollo-13-houston-we-have-had-a-problem/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # apollo-13-houston-we-have-had-a-problem
2
 
3
- **What this is:** 1 file published by Wikimedia Commons (NASA air-to-ground recording); English timed captions on Commons TimedText. Collection theme: Human-transcribed English audio (read speech, lectures, broadcasts with human transcripts).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Ogg Vorbis, 3,644,189 bytes, air-to-ground radio exchange between Apollo 13 and Mission Control.
6
 
7
  **Where it comes from:** Upstream: https://commons.wikimedia.org/wiki/File:Apollo13-wehaveaproblem.ogg. Kind: real files published upstream. NASA mission recording; captions written by Commons volunteers.
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: Give the agent the ogg only; it writes the transcript; score word error rate (WER) against the caption words, reported as a number with no pass mark; cue numbers and timestamps are removed; case and punctuation folded on both sides; the radio channel is noisy and clipped, so this is a hard far-field item; the captions are a volunteer transcription, so the score is a soft reference score.
12
 
 
1
  # apollo-13-houston-we-have-had-a-problem
2
 
3
+ **What this is:** Files published by Wikimedia Commons (NASA air-to-ground recording); English timed captions on Commons TimedText. Collection theme: Human-transcribed English audio (read speech, lectures, broadcasts with human transcripts).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Ogg Vorbis, 3,644,189 bytes, air-to-ground radio exchange between Apollo 13 and Mission Control.
6
 
7
  **Where it comes from:** Upstream: https://commons.wikimedia.org/wiki/File:Apollo13-wehaveaproblem.ogg. Kind: real files published upstream. NASA mission recording; captions written by Commons volunteers.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Give the agent the ogg only; it writes the transcript; score word error rate (WER) against the caption words, reported as a number with no pass mark; cue numbers and timestamps are removed; case and punctuation folded on both sides; the radio channel is noisy and clipped, so this is a hard far-field item; the captions are a volunteer transcription, so the score is a soft reference score.
12
 
campaign_files/2026-10-md-media-and-documents/apriltag-test-data/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # apriltag-test-data
2
 
3
- **What this is:** 3 files published by AprilRobotics (APRIL Robotics Laboratory, University of Michigan); photo by NASA Kennedy Space Center. Collection theme: Annotated or reference-output images and video (test material of image and video libraries, annotated public image sets).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: JPEG photo (799 by 533) plus the expected detections as text (12 lines); JPEG photo (799 by 533) plus the expected detections as text (25 lines); JPEG photo (799 by 533) plus the expected detections as text (10 lines).
6
 
7
  **Where it comes from:** Upstream: https://github.com/AprilRobotics/apriltag. Kind: test or example files shipped with a tool. Regression test image of the AprilTag 3 detector with its expected detections; test/data/README.md (AprilRobotics/apriltag @b7c0ebe9aa20f82ec7a828579004f9e706bfecd9) line 1: "Test image from NASA Swarmathon competition at KSC. Licenced under "CC BY-SA 2.0 DEED"."; test/test_detection.c (AprilRobotics/apriltag @b7c0ebe9aa20f82ec7a828579004f9e706bfecd9) line 82: "char* const path_det_true = format("%s.txt", argv[1]);".
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 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:** Tools are named in the check written per file in the notes.
12
 
 
1
  # apriltag-test-data
2
 
3
+ **What this is:** Files published by AprilRobotics (APRIL Robotics Laboratory, University of Michigan); photo by NASA Kennedy Space Center. Collection theme: Annotated or reference-output images and video (test material of image and video libraries, annotated public image sets).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: JPEG photo (799 by 533) plus the expected detections as text (12 lines); JPEG photo (799 by 533) plus the expected detections as text (25 lines); JPEG photo (799 by 533) plus the expected detections as text (10 lines).
6
 
7
  **Where it comes from:** Upstream: https://github.com/AprilRobotics/apriltag. Kind: test or example files shipped with a tool. Regression test image of the AprilTag 3 detector with its expected detections; test/data/README.md (AprilRobotics/apriltag @b7c0ebe9aa20f82ec7a828579004f9e706bfecd9) line 1: "Test image from NASA Swarmathon competition at KSC. Licenced under "CC BY-SA 2.0 DEED"."; test/test_detection.c (AprilRobotics/apriltag @b7c0ebe9aa20f82ec7a828579004f9e706bfecd9) line 82: "char* const path_det_true = format("%s.txt", argv[1]);".
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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:** Tools are named in the check written per file in the notes.
12
 
campaign_files/2026-10-md-media-and-documents/arnog-mathlive-test/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # arnog-mathlive-test
2
 
3
- **What this is:** 1 file published by Arno Gourdol (arnog/mathlive). Collection theme: Typesetting engine sources with reference renderings (LaTeX, Typst and similar regression tests).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: TypeScript Jest test plus snapshot file.
6
 
7
  **Where it comes from:** Upstream name: arnog/mathlive test/. Upstream: https://github.com/arnog/mathlive. Kind: test or example files shipped with a tool. Unit test of MathLive's LaTeX to HTML markup renderer, written by its author; expected markup stored as a Jest snapshot.
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:** npx jest test/latex.test.ts in a mathlive checkout (Node, headless)
12
 
 
1
  # arnog-mathlive-test
2
 
3
+ **What this is:** Files published by Arno Gourdol (arnog/mathlive). Collection theme: Typesetting engine sources with reference renderings (LaTeX, Typst and similar regression tests).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: TypeScript Jest test plus snapshot file.
6
 
7
  **Where it comes from:** Upstream name: arnog/mathlive test/. Upstream: https://github.com/arnog/mathlive. Kind: test or example files shipped with a tool. Unit test of MathLive's LaTeX to HTML markup renderer, written by its author; expected markup stored as a Jest snapshot.
8
 
9
+ **Self-contained:** Whether the file needs other upstream files is not stated. Each file names 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:** npx jest test/latex.test.ts in a mathlive checkout (Node, headless)
12
 
campaign_files/2026-10-md-media-and-documents/asciidoc-py-tests-oldtables/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # asciidoc-py-tests-oldtables
2
 
3
- **What this is:** 1 file published by Stuart Rackham and asciidoc-py contributors (asciidoc-py/asciidoc-py). Collection theme: Document conversion tool test cases (source document plus expected output, pandoc and similar).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: AsciiDoc source oldtables.txt (1784 bytes) with the expected DocBook output oldtables-docbook.xml (5950 bytes).
6
 
7
  **Where it comes from:** Upstream: https://github.com/asciidoc-py/asciidoc-py. Kind: test or example files shipped with a tool. Test document of the asciidoc-py suite for the deprecated "old tables" syntax; the sister file newtables.txt was not taken (it needs source-highlight and includes Northwind sample customer data).
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:** asciidoc -b docbook -o out.xml oldtables.txt (the asciidoc command line converter); asciidoctor -b docbook is comparable
12
 
 
1
  # asciidoc-py-tests-oldtables
2
 
3
+ **What this is:** Files published by Stuart Rackham and asciidoc-py contributors (asciidoc-py/asciidoc-py). Collection theme: Document conversion tool test cases (source document plus expected output, pandoc and similar).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: AsciiDoc source oldtables.txt (1784 bytes) with the expected DocBook output oldtables-docbook.xml (5950 bytes).
6
 
7
  **Where it comes from:** Upstream: https://github.com/asciidoc-py/asciidoc-py. Kind: test or example files shipped with a tool. Test document of the asciidoc-py suite for the deprecated "old tables" syntax; the sister file newtables.txt was not taken (it needs source-highlight and includes Northwind sample customer data).
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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:** asciidoc -b docbook -o out.xml oldtables.txt (the asciidoc command line converter); asciidoctor -b docbook is comparable
12
 
campaign_files/2026-10-md-media-and-documents/asciidoc-tck/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # asciidoc-tck
2
 
3
- **What this is:** 1 file published by Eclipse Foundation AsciiDoc Working Group (gitlab.eclipse.org/eclipse/asciidoc-lang/asciidoc-tck). Collection theme: Document conversion tool test cases (source document plus expected output, pandoc and similar).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: AsciiDoc input, expected ASG JSON, test config JSON.
6
 
7
  **Where it comes from:** Upstream: https://gitlab.eclipse.org/eclipse/asciidoc-lang/asciidoc-tck. Kind: test or example files shipped with a tool. Eclipse AsciiDoc Language technology compatibility kit (TCK): input document, expected ASG JSON and config written by the AsciiDoc Working Group; harness/lib/suites/suite.js is the 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:** an AsciiDoc parser adapter (for example asciidoctor with an ASG adapter) run by the TCK harness: node harness/test/index.js, the parser's JSON output compared with title-body-output.json
12
 
 
1
  # asciidoc-tck
2
 
3
+ **What this is:** Files published by Eclipse Foundation AsciiDoc Working Group (gitlab.eclipse.org/eclipse/asciidoc-lang/asciidoc-tck). Collection theme: Document conversion tool test cases (source document plus expected output, pandoc and similar).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: AsciiDoc input, expected ASG JSON, test config JSON.
6
 
7
  **Where it comes from:** Upstream: https://gitlab.eclipse.org/eclipse/asciidoc-lang/asciidoc-tck. Kind: test or example files shipped with a tool. Eclipse AsciiDoc Language technology compatibility kit (TCK): input document, expected ASG JSON and config written by the AsciiDoc Working Group; harness/lib/suites/suite.js is the judge.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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:** an AsciiDoc parser adapter (for example asciidoctor with an ASG adapter) run by the TCK harness: node harness/test/index.js, the parser's JSON output compared with title-body-output.json
12
 
campaign_files/2026-10-md-media-and-documents/asciidoctor-html5s/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # asciidoctor-html5s
2
 
3
- **What this is:** 1 file published by Jakub Jirutka (jirutka/asciidoctor-html5s). Collection theme: Document conversion tool test cases (source document plus expected output, pandoc and similar).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: AsciiDoc file of named examples (// .name comments) and HTML file of named expected examples (<!-- .name --> comments).
6
 
7
  **Where it comes from:** Upstream: https://github.com/jirutka/asciidoctor-html5s. Kind: test or example files shipped with a tool. Named AsciiDoc examples and expected HTML5 examples written by the converter author; the Rakefile doctest task (asciidoctor-doctest) is the judge. Only the examples named in the .adoc file (source-highlighter-highlightjs, source-highlighter-html-pipeline) have their input in this repository; the other examples in listing.html take their input from the asciidoctor-doctest gem's example set and are not part of this unit.
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:** asciidoctor -r asciidoctor-html5s -b html5s on each named example, output compared with the matching named example in listing.html by asciidoctor-doctest
12
 
 
1
  # asciidoctor-html5s
2
 
3
+ **What this is:** Files published by Jakub Jirutka (jirutka/asciidoctor-html5s). Collection theme: Document conversion tool test cases (source document plus expected output, pandoc and similar).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: AsciiDoc file of named examples (// .name comments) and HTML file of named expected examples (<!-- .name --> comments).
6
 
7
  **Where it comes from:** Upstream: https://github.com/jirutka/asciidoctor-html5s. Kind: test or example files shipped with a tool. Named AsciiDoc examples and expected HTML5 examples written by the converter author; the Rakefile doctest task (asciidoctor-doctest) is the judge. Only the examples named in the .adoc file (source-highlighter-highlightjs, source-highlighter-html-pipeline) have their input in this repository; the other examples in listing.html take their input from the asciidoctor-doctest gem's example set and are not part of this unit.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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:** asciidoctor -r asciidoctor-html5s -b html5s on each named example, output compared with the matching named example in listing.html by asciidoctor-doctest
12
 
campaign_files/2026-10-md-media-and-documents/audiomnist/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # audiomnist
2
 
3
- **What this is:** 1 file published by Soeren Becker et al. (soerenab), AudioMNIST, GitHub. Collection theme: Human-transcribed English audio (read speech, lectures, broadcasts with human transcripts).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: WAV, 48 kHz mono 16-bit, one spoken English digit.
6
 
7
  **Where it comes from:** Upstream: https://github.com/soerenab/AudioMNIST. Kind: real files published upstream. Read single digits recorded by the authors.
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: Give the agent the wav under a neutral name (hide the file name); it answers which digit is spoken; exact match with the digit from the original file name.
12
 
 
1
  # audiomnist
2
 
3
+ **What this is:** Files published by Soeren Becker et al. (soerenab), AudioMNIST, GitHub. Collection theme: Human-transcribed English audio (read speech, lectures, broadcasts with human transcripts).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: WAV, 48 kHz mono 16-bit, one spoken English digit.
6
 
7
  **Where it comes from:** Upstream: https://github.com/soerenab/AudioMNIST. Kind: real files published upstream. Read single digits recorded by the authors.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Give the agent the wav under a neutral name (hide the file name); it answers which digit is spoken; exact match with the digit from the original file name.
12
 
campaign_files/2026-10-md-media-and-documents/bach-chorale-figured-bass/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # bach-chorale-figured-bass
2
 
3
- **What this is:** 2 files published by Yaolong Ju et al., Bach Chorales Figured Bass dataset (GitHub juyaolongpaul/Bach_chorale_FB; Zenodo 5084914); Yaolong Ju et al., Bach Chorales Figured Bass dataset (GitHub juyaolongpaul/Bach_chorale_FB). Collection theme: Annotated score corpora (MusicXML, MIDI, ABC with harmony or form annotations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. A file whose upstream path has no folder is kept in a folder named after its unit. Format: MusicXML, 100,913 bytes, voices plus continuo staff carrying <figured-bass> elements; MusicXML, 104,521 bytes, voices plus continuo staff carrying 42 <figured-bass> elements.
6
 
 
1
  # bach-chorale-figured-bass
2
 
3
+ **What this is:** Files published by Yaolong Ju et al., Bach Chorales Figured Bass dataset (GitHub juyaolongpaul/Bach_chorale_FB; Zenodo 5084914); Yaolong Ju et al., Bach Chorales Figured Bass dataset (GitHub juyaolongpaul/Bach_chorale_FB). Collection theme: Annotated score corpora (MusicXML, MIDI, ABC with harmony or form annotations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. A file whose upstream path has no folder is kept in a folder named after its unit. Format: MusicXML, 100,913 bytes, voices plus continuo staff carrying <figured-bass> elements; MusicXML, 104,521 bytes, voices plus continuo staff carrying 42 <figured-bass> elements.
6
 
campaign_files/2026-10-ml-serving-configs-2/aiops2018-kpi/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # aiops2018-kpi
2
 
3
- **What this is:** 1 file published by NetManAIOps (Tsinghua University) for the AIOps Challenge 2018. Collection theme: Labelled monitoring and drift data (time-series anomaly detection, changepoint detection and model monitoring data with labelled anomaly, change or drift points).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: ZIP holding one CSV (timestamp,value,label,KPI ID).
6
 
7
  **Where it comes from:** Upstream: NetManAIOps/KPI-Anomaly-Detection. Kind: real files published upstream. Real web-service KPI curves of large internet companies released for the AIOps Challenge 2018, republished by NetManAIOps.
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: Hold out the label column. The agent emits 0 or 1 per point for a chosen KPI ID (or for all 29); score by point-wise F1 against the label column, per KPI and averaged. The challenge also used a detection-delay rule that the repository does not state, so it is not applied.
12
 
 
1
  # aiops2018-kpi
2
 
3
+ **What this is:** Files published by NetManAIOps (Tsinghua University) for the AIOps Challenge 2018. Collection theme: Labelled monitoring and drift data (time-series anomaly detection, changepoint detection and model monitoring data with labelled anomaly, change or drift points).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: ZIP holding one CSV (timestamp,value,label,KPI ID).
6
 
7
  **Where it comes from:** Upstream: NetManAIOps/KPI-Anomaly-Detection. Kind: real files published upstream. Real web-service KPI curves of large internet companies released for the AIOps Challenge 2018, republished by NetManAIOps.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Hold out the label column. The agent emits 0 or 1 per point for a chosen KPI ID (or for all 29); score by point-wise F1 against the label column, per KPI and averaged. The challenge also used a detection-delay rule that the repository does not state, so it is not applied.
12
 
campaign_files/2026-10-ml-serving-configs-2/arm-optimized-routines-directed-tests/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # arm-optimized-routines-directed-tests
2
 
3
- **What this is:** 8 files published by Arm (ARM-software/optimized-routines, math/test/testcases/directed). Collection theme: Operator reference implementations and test vectors (backend tests of operator specifications, reference implementations and test data of operator libraries).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Text file, one case per line: func=<name> op1=<hex bits> [op2=<hex bits>] result=<hex bits>[.<3 hex digits of extra bits>] errno=<E...> status=<flags>; ';' comment lines.
6
 
7
  **Where it comes from:** Upstream: ARM-software/optimized-routines. Kind: test or example files shipped with a tool. Test data of the Arm optimized math routines: hand-picked hard and special cases with high precision reference results.
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 8 of 8 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:** Tools are named in the check written per file in the notes.
12
 
 
1
  # arm-optimized-routines-directed-tests
2
 
3
+ **What this is:** Files published by Arm (ARM-software/optimized-routines, math/test/testcases/directed). Collection theme: Operator reference implementations and test vectors (backend tests of operator specifications, reference implementations and test data of operator libraries).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Text file, one case per line: func=<name> op1=<hex bits> [op2=<hex bits>] result=<hex bits>[.<3 hex digits of extra bits>] errno=<E...> status=<flags>; ';' comment lines.
6
 
7
  **Where it comes from:** Upstream: ARM-software/optimized-routines. Kind: test or example files shipped with a tool. Test data of the Arm optimized math routines: hand-picked hard and special cases with high precision reference results.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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:** Tools are named in the check written per file in the notes.
12
 
campaign_files/2026-10-op-rule-calculations-2/accordproject-cicero-template-library/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # accordproject-cicero-template-library
2
 
3
- **What this is:** 9 units (53 files) published by Accord Project (Linux Foundation project). Collection theme: Contract templates (open-licence standard contracts, government model contracts and clauses, contract sets with clause annotations).
4
 
5
  **What one unit is:** Each unit is one upstream folder, delivered as all the files fetched from it, kept at its upstream path. Format: Directory of 13 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 20544 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 16449 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 34520 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10340 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 11481 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 9577 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10382 bytes; directory of 5 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 4818 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10456 bytes.
6
 
@@ -16,47 +16,47 @@
16
 
17
  Units:
18
 
19
- - `src/fragile-goods/` (6 files): Fragile goods delivery clause: payout after accelerometer-breach and late-delivery penalties.
20
  Format: directory of 13 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 20544 bytes
21
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
22
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
23
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
24
- - `src/perishable-goods/` (6 files): Perishable goods (avocado shipment) clause: price after temperature and humidity penalties and late-arrival forfeiture.
25
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 16449 bytes
26
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
27
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
28
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
29
- - `src/supplyagreement-perishable-goods/` (6 files): Supply agreement for perishable goods: price, temperature penalty and late forfeiture.
30
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 34520 bytes
31
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
32
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
33
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
34
- - `src/installment-sale/` (6 files): Installment sale clause: balance and total paid across monthly installments and the closing payment.
35
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10340 bytes
36
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
37
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
38
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
39
- - `src/rental-deposit/` (6 files): Rental security deposit clause: deposit balance returned after inspection penalties.
40
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 11481 bytes
41
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
42
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
43
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
44
- - `src/volumediscount/` (6 files): Volume-based card acceptance agreement: discount rate tier for a given annual volume.
45
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 9577 bytes
46
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
47
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
48
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
49
- - `src/ip-payment/` (6 files): IP licence payment clause: royalty total and due date.
50
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10382 bytes
51
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
52
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
53
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
54
- - `src/lateinvoicewithpayment/` (5 files): Late invoice clause: whether payment is still required and the amount due.
55
  Format: directory of 5 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 4818 bytes
56
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
57
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
58
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
59
- - `src/acceptance-of-delivery/` (6 files): Acceptance of delivery clause: inspection outcome status.
60
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10456 bytes
61
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
62
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
 
1
  # accordproject-cicero-template-library
2
 
3
+ **What this is:** Units published by Accord Project (Linux Foundation project). Collection theme: Contract templates (open-licence standard contracts, government model contracts and clauses, contract sets with clause annotations).
4
 
5
  **What one unit is:** Each unit is one upstream folder, delivered as all the files fetched from it, kept at its upstream path. Format: Directory of 13 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 20544 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 16449 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 34520 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10340 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 11481 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 9577 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10382 bytes; directory of 5 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 4818 bytes; directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10456 bytes.
6
 
 
16
 
17
  Units:
18
 
19
+ - `src/fragile-goods/`: Fragile goods delivery clause: payout after accelerometer-breach and late-delivery penalties.
20
  Format: directory of 13 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 20544 bytes
21
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
22
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
23
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
24
+ - `src/perishable-goods/`: Perishable goods (avocado shipment) clause: price after temperature and humidity penalties and late-arrival forfeiture.
25
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 16449 bytes
26
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
27
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
28
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
29
+ - `src/supplyagreement-perishable-goods/`: Supply agreement for perishable goods: price, temperature penalty and late forfeiture.
30
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 34520 bytes
31
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
32
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
33
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
34
+ - `src/installment-sale/`: Installment sale clause: balance and total paid across monthly installments and the closing payment.
35
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10340 bytes
36
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
37
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
38
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
39
+ - `src/rental-deposit/`: Rental security deposit clause: deposit balance returned after inspection penalties.
40
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 11481 bytes
41
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
42
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
43
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
44
+ - `src/volumediscount/`: Volume-based card acceptance agreement: discount rate tier for a given annual volume.
45
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 9577 bytes
46
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
47
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
48
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
49
+ - `src/ip-payment/`: IP licence payment clause: royalty total and due date.
50
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10382 bytes
51
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
52
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
53
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
54
+ - `src/lateinvoicewithpayment/`: Late invoice clause: whether payment is still required and the amount due.
55
  Format: directory of 5 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 4818 bytes
56
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
57
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
58
  Tools: the clause text (text/sample.md, text/grammar.tem.md) is plain Markdown read with cat or pandoc; the agent reads the template model and the request data from the test file, computes the result in a Python or Node script and writes it as JSON
59
+ - `src/acceptance-of-delivery/`: Acceptance of delivery clause: inspection outcome status.
60
  Format: directory of 6 small text files (TypeScript test, Markdown clause text and grammar, Concerto model, JSON), 10456 bytes
61
  Origin: Accord Project template library: a sample clause text with its grammar and the template's own unit tests (test inputs and expected outputs written by the template authors)
62
  Check: give the agent the clause text and grammar plus, per test case, the template data and the request written in the test file; it reports the result fields; pass when every field equals the value the test expects
campaign_files/2026-10-op-rule-calculations-2/adventure-works-oltp-sample/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # adventure-works-oltp-sample
2
 
3
- **What this is:** 4 files published by Microsoft (sql-server-samples repository). Collection theme: Inventory and order records: stock valuation and stock-move tests of open-source ERPs, order and inventory tables of sample databases.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Tab-separated text without header row, columns in the order of CREATE TABLE [Production].[ProductInventory] (instawdb.sql lines 1344-1351): ProductID, LocationID, Shelf, Bin, Quantity, rowguid, ModifiedDate; 81,385 bytes; tab-separated text without header row, columns in the order of CREATE TABLE [Production].[TransactionHistory] (instawdb.sql lines 1721-1730): TransactionID, ProductID, ReferenceOrderID, ReferenceOrderLineID, TransactionDate, TransactionType, Quantity, ActualCost, ModifiedDate; 8,937,794 bytes; tab-separated text without header row: WorkOrderID, ProductID, OrderQty, StockedQty, ScrappedQty, StartDate, EndDate, DueDate, ScrapReasonID, ModifiedDate; 8,239,894 bytes; tab-separated text without header row: SalesOrderID, SalesOrderDetailID, CarrierTrackingNumber, OrderQty, ProductID, SpecialOfferID, UnitPrice, UnitPriceDiscount, LineTotal, rowguid, ModifiedDate; 13,727,911 bytes.
6
 
7
  **Where it comes from:** Upstream: https://github.com/microsoft/sql-server-samples. Kind: test or example files shipped with a tool. Sample database shipped by the vendor for product demos and tutorials; made-up business data.
8
 
9
- **Self-contained:** Whether each file needs other upstream files is not stated. 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:** the CSV is tab-separated plain text: Python csv module or pandas, sqlite3 after import, or csvkit (csvsql, csvstat); no spreadsheet program needed
12
 
 
1
  # adventure-works-oltp-sample
2
 
3
+ **What this is:** Files published by Microsoft (sql-server-samples repository). Collection theme: Inventory and order records: stock valuation and stock-move tests of open-source ERPs, order and inventory tables of sample databases.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Tab-separated text without header row, columns in the order of CREATE TABLE [Production].[ProductInventory] (instawdb.sql lines 1344-1351): ProductID, LocationID, Shelf, Bin, Quantity, rowguid, ModifiedDate; 81,385 bytes; tab-separated text without header row, columns in the order of CREATE TABLE [Production].[TransactionHistory] (instawdb.sql lines 1721-1730): TransactionID, ProductID, ReferenceOrderID, ReferenceOrderLineID, TransactionDate, TransactionType, Quantity, ActualCost, ModifiedDate; 8,937,794 bytes; tab-separated text without header row: WorkOrderID, ProductID, OrderQty, StockedQty, ScrappedQty, StartDate, EndDate, DueDate, ScrapReasonID, ModifiedDate; 8,239,894 bytes; tab-separated text without header row: SalesOrderID, SalesOrderDetailID, CarrierTrackingNumber, OrderQty, ProductID, SpecialOfferID, UnitPrice, UnitPriceDiscount, LineTotal, rowguid, ModifiedDate; 13,727,911 bytes.
6
 
7
  **Where it comes from:** Upstream: https://github.com/microsoft/sql-server-samples. Kind: test or example files shipped with a tool. Sample database shipped by the vendor for product demos and tutorials; made-up business data.
8
 
9
+ **Self-contained:** Whether each file needs other upstream files is not stated. Each file names 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:** the CSV is tab-separated plain text: Python csv module or pandas, sqlite3 after import, or csvkit (csvsql, csvstat); no spreadsheet program needed
12
 
campaign_files/2026-10-op-rule-calculations-2/apra-adi-centralised-publication/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # apra-adi-centralised-publication
2
 
3
- **What this is:** 1 file published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filing data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. A file whose upstream path has no folder is kept in a folder named after its unit. Format: Xlsx workbook (values only, no formulas).
6
 
7
  **Where it comes from:** Upstream: https://www.apra.gov.au/system/files/2026-09/Authorised%20deposit-taking%20institution%20centralised%20publication%20-%20March%202013%20to%20June%202026%20%281%29.xlsx. Kind: real files published upstream. Official prudential statistics compiled from ADI regulatory returns.
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:** read the xlsx with openpyxl or pandas.read_excel, or convert sheets to CSV with libreoffice --headless --convert-to csv; write answers to CSV
12
 
 
1
  # apra-adi-centralised-publication
2
 
3
+ **What this is:** Files published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filing data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. A file whose upstream path has no folder is kept in a folder named after its unit. Format: Xlsx workbook (values only, no formulas).
6
 
7
  **Where it comes from:** Upstream: https://www.apra.gov.au/system/files/2026-09/Authorised%20deposit-taking%20institution%20centralised%20publication%20-%20March%202013%20to%20June%202026%20%281%29.xlsx. Kind: real files published upstream. Official prudential statistics compiled from ADI regulatory returns.
8
 
9
+ **Self-contained:** The file is self-contained. Each file names 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:** read the xlsx with openpyxl or pandas.read_excel, or convert sheets to CSV with libreoffice --headless --convert-to csv; write answers to CSV
12
 
campaign_files/2026-10-op-rule-calculations-2/apra-phi-medical-gap-2026q2/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # apra-phi-medical-gap-2026q2
2
 
3
- **What this is:** 1 file published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filed data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: XLSX workbook with formulas stored in the cells and cached values; Selection sheet drives the T tables (cached for Year 2026, MonthEnd Jun).
6
 
 
1
  # apra-phi-medical-gap-2026q2
2
 
3
+ **What this is:** Files published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filed data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: XLSX workbook with formulas stored in the cells and cached values; Selection sheet drives the T tables (cached for Year 2026, MonthEnd Jun).
6
 
campaign_files/2026-10-op-rule-calculations-2/apra-phi-membership-and-benefits-2026q2/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # apra-phi-membership-and-benefits-2026q2
2
 
3
- **What this is:** 1 file published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filed data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: XLSX workbook with formulas stored in the cells and cached values; Selection sheet drives the T tables (cached for Year 2026, MonthEnd Jun).
6
 
 
1
  # apra-phi-membership-and-benefits-2026q2
2
 
3
+ **What this is:** Files published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filed data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: XLSX workbook with formulas stored in the cells and cached values; Selection sheet drives the T tables (cached for Year 2026, MonthEnd Jun).
6
 
campaign_files/2026-10-op-rule-calculations-2/apra-phi-membership-coverage-2026q2/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # apra-phi-membership-coverage-2026q2
2
 
3
- **What this is:** 1 file published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filed data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: XLSX workbook with formulas stored in the cells and cached values; Selection sheet drives the T tables (cached for Year 2026, MonthEnd Jun).
6
 
 
1
  # apra-phi-membership-coverage-2026q2
2
 
3
+ **What this is:** Files published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filed data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: XLSX workbook with formulas stored in the cells and cached values; Selection sheet drives the T tables (cached for Year 2026, MonthEnd Jun).
6
 
campaign_files/2026-10-op-rule-calculations-2/apra-phi-membership-trends-2026q2/README.md CHANGED
@@ -1,6 +1,6 @@
1
  # apra-phi-membership-trends-2026q2
2
 
3
- **What this is:** 1 file published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filed data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: XLSX workbook with formulas stored in the cells and cached values; Selection sheet drives the T tables (cached for Year 2026, MonthEnd Jun).
6
 
 
1
  # apra-phi-membership-trends-2026q2
2
 
3
+ **What this is:** Files published by Australian Prudential Regulation Authority (APRA). Collection theme: Regulatory reporting and regulatory calculation (filed data with officially computed ratios, XBRL validation rules, official worked examples of regulatory capital calculations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: XLSX workbook with formulas stored in the cells and cached values; Selection sheet drives the T tables (cached for Year 2026, MonthEnd Jun).
6
 
campaign_files/2026-10-op-rule-calculations-2/atticus-acord/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # atticus-acord
2
 
3
- **What this is:** 1 file published by The Atticus Project (Wang et al., ACORD, NAACL 2025). Collection theme: Contract templates (open-licence standard contracts, government model contracts and clauses, contract sets with clause annotations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: TSV 3,885,909 bytes with companions corpus.jsonl 4,508,873 bytes and queries.jsonl 25,482 bytes.
6
 
7
  **Where it comes from:** Upstream: https://www.atticusprojectai.org/acord/. Kind: real files published upstream. Clauses extracted from CUAD contracts (real SEC filings); 114 drafting queries written and every query-clause pair rated by lawyers of The Atticus Project.
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:** the clause corpus and queries are JSON lines read with jq or Python; the agent ranks clauses per drafting query and writes a run file (query-id, corpus-id, score) scored with a plain Python NDCG script
12
 
 
1
  # atticus-acord
2
 
3
+ **What this is:** Files published by The Atticus Project (Wang et al., ACORD, NAACL 2025). Collection theme: Contract templates (open-licence standard contracts, government model contracts and clauses, contract sets with clause annotations).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: TSV 3,885,909 bytes with companions corpus.jsonl 4,508,873 bytes and queries.jsonl 25,482 bytes.
6
 
7
  **Where it comes from:** Upstream: https://www.atticusprojectai.org/acord/. Kind: real files published upstream. Clauses extracted from CUAD contracts (real SEC filings); 114 drafting queries written and every query-clause pair rated by lawyers of The Atticus Project.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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:** the clause corpus and queries are JSON lines read with jq or Python; the agent ranks clauses per drafting query and writes a run file (query-id, corpus-id, score) scored with a plain Python NDCG script
12
 
campaign_files/2026-10-s-security-detection-2/0xdea-semgrep-rules/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # 0xdea-semgrep-rules
2
 
3
- **What this is:** 1 file published by Marco Ivaldi (raptor, 0xdea). Collection theme: Code and configuration security rules with their own tests and expected findings (Semgrep, CodeQL, KICS, trivy-checks class).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: C source (1,837 B) annotated with Semgrep ruleid and ok comments; rule in YAML.
6
 
7
  **Where it comes from:** Upstream name: 0xdea semgrep-rules (raptor). Upstream: 0xdea/semgrep-rules. Kind: test or example files shipped with a tool. Test fixture written by the rule maintainers to pin the expected findings of their own rule (synthetic, not recorded production code or telemetry).
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: report the lines with a format string bug. Score precision and recall of the reported lines against the 10 ruleid lines; an ok line reported counts as a false positive. todoruleid would count as a missed real bug and todook as a false positive (none here).
12
 
 
1
  # 0xdea-semgrep-rules
2
 
3
+ **What this is:** Files published by Marco Ivaldi (raptor, 0xdea). Collection theme: Code and configuration security rules with their own tests and expected findings (Semgrep, CodeQL, KICS, trivy-checks class).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: C source (1,837 B) annotated with Semgrep ruleid and ok comments; rule in YAML.
6
 
7
  **Where it comes from:** Upstream name: 0xdea semgrep-rules (raptor). Upstream: 0xdea/semgrep-rules. Kind: test or example files shipped with a tool. Test fixture written by the rule maintainers to pin the expected findings of their own rule (synthetic, not recorded production code or telemetry).
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: report the lines with a format string bug. Score precision and recall of the reported lines against the 10 ruleid lines; an ok line reported counts as a false positive. todoruleid would count as a missed real bug and todook as a false positive (none here).
12
 
campaign_files/2026-10-s-security-detection-2/aqua-trivy-checks/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # aqua-trivy-checks
2
 
3
- **What this is:** 5 files published by Aqua Security. Collection theme: Code and configuration security rules with their own tests and expected findings (Semgrep, CodeQL, KICS, trivy-checks class).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Rego check with METADATA header, its _test.rego (3 opa unit tests), the examples YAML (good and bad CloudFormation and Terraform snippets), and the lib.test helper package; Rego check with METADATA header, its _test.rego (opa unit tests), the examples YAML (good and bad Terraform snippets) and the lib helper packages it imports; Rego check with METADATA header, its _test.rego (opa unit tests), the examples YAML (good and bad Terraform snippets).
6
 
7
  **Where it comes from:** Upstream name: Aqua trivy-checks (Rego misconfiguration checks with opa unit tests and good and bad examples). Upstream: aquasecurity/trivy-checks. Kind: test or example files shipped with a tool. Test fixture written by the rule maintainers to pin the expected findings of their own rule (synthetic, not recorded production code or telemetry).
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 5 of 5 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
  **Tools to open and check:** Check: Task: given the check and each input (3 test inputs, 4 example snippets, labels removed), say whether the check denies it. Score: exact match against the test expectations and the good or bad label of each example; accuracy over 7; Task: given the check and each input (3 test inputs, 3 example snippets, labels removed), say whether the check denies it. Score: exact match against the test expectations and the good or bad label of each example; accuracy over 6; Task: given the check and each input (4 test inputs, 5 example snippets, labels removed), say whether the check denies it. Score: exact match against the test expectations and the good or bad label of each example; accuracy over 9; Task: given the check and each input (3 test inputs, 2 example snippets, labels removed), say whether the check denies it. Score: exact match against the test expectations and the good or bad label of each example; accuracy over 5.
12
 
 
1
  # aqua-trivy-checks
2
 
3
+ **What this is:** Files published by Aqua Security. Collection theme: Code and configuration security rules with their own tests and expected findings (Semgrep, CodeQL, KICS, trivy-checks class).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Rego check with METADATA header, its _test.rego (3 opa unit tests), the examples YAML (good and bad CloudFormation and Terraform snippets), and the lib.test helper package; Rego check with METADATA header, its _test.rego (opa unit tests), the examples YAML (good and bad Terraform snippets) and the lib helper packages it imports; Rego check with METADATA header, its _test.rego (opa unit tests), the examples YAML (good and bad Terraform snippets).
6
 
7
  **Where it comes from:** Upstream name: Aqua trivy-checks (Rego misconfiguration checks with opa unit tests and good and bad examples). Upstream: aquasecurity/trivy-checks. Kind: test or example files shipped with a tool. Test fixture written by the rule maintainers to pin the expected findings of their own rule (synthetic, not recorded production code or telemetry).
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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 check and each input (3 test inputs, 4 example snippets, labels removed), say whether the check denies it. Score: exact match against the test expectations and the good or bad label of each example; accuracy over 7; Task: given the check and each input (3 test inputs, 3 example snippets, labels removed), say whether the check denies it. Score: exact match against the test expectations and the good or bad label of each example; accuracy over 6; Task: given the check and each input (4 test inputs, 5 example snippets, labels removed), say whether the check denies it. Score: exact match against the test expectations and the good or bad label of each example; accuracy over 9; Task: given the check and each input (3 test inputs, 2 example snippets, labels removed), say whether the check denies it. Score: exact match against the test expectations and the good or bad label of each example; accuracy over 5.
12
 
campaign_files/2026-10-s-security-detection-2/atlas-attack-investigation-logs/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # atlas-attack-investigation-logs
2
 
3
- **What this is:** 10 files published by Alsaheel et al., Purdue University (purseclab/ATLAS). Collection theme: Host and cloud telemetry (event logs, EDR exports, cloud audit logs).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. The format of each file is in the notes.
6
 
7
  **Where it comes from:** Upstream name: ATLAS (USENIX Security 2021) attack investigation logs. Upstream: purseclab/ATLAS. Kind: real files published upstream. Logs recorded by the ATLAS authors on their own Windows lab hosts while they replayed real exploits against them, released as the paper's artifacts (raw_logs folder).
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 10 of 10 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes.
10
 
11
  **Tools to open and check:** Check: Investigation task: given the raw logs and one starting point (e.g. the attacker IP or the payload name), list the attack entities (domains, IP addresses, files, processes) involved; compare the answer set to malicious_labels.txt of the host after normalising case and path separators and merging an entity with its process instances (payload.exe and payload.exe_892 count as one) as README.md section (D), testing phase output, describes; score precision and recall on the entity set; multi-host scenarios are scored over both hosts together as the README Notes describe. Classification variant: name the exploited CVE or the initial access document from the logs.
12
 
 
1
  # atlas-attack-investigation-logs
2
 
3
+ **What this is:** Files published by Alsaheel et al., Purdue University (purseclab/ATLAS). Collection theme: Host and cloud telemetry (event logs, EDR exports, cloud audit logs).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. The format of each file is in the notes.
6
 
7
  **Where it comes from:** Upstream name: ATLAS (USENIX Security 2021) attack investigation logs. Upstream: purseclab/ATLAS. Kind: real files published upstream. Logs recorded by the ATLAS authors on their own Windows lab hosts while they replayed real exploits against them, released as the paper's artifacts (raw_logs folder).
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Investigation task: given the raw logs and one starting point (e.g. the attacker IP or the payload name), list the attack entities (domains, IP addresses, files, processes) involved; compare the answer set to malicious_labels.txt of the host after normalising case and path separators and merging an entity with its process instances (payload.exe and payload.exe_892 count as one) as README.md section (D), testing phase output, describes; score precision and recall on the entity set; multi-host scenarios are scored over both hosts together as the README Notes describe. Classification variant: name the exploited CVE or the initial access document from the logs.
12
 
campaign_files/2026-10-s-security-detection/aquasecurity-trivy-checks/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # aquasecurity-trivy-checks
2
 
3
- **What this is:** 1 file published by Aqua Security. Collection theme: Detection rule repositories with own tests (Sigma, YARA, Snort class).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Rego check + YAML good/bad examples (CloudFormation 1 good, 1 bad; Terraform 2 good, 1 bad) + Rego library.
6
 
7
  **Where it comes from:** Upstream: aquasecurity/trivy-checks. Kind: test or example files shipped with a tool. Test cases written by hand for the detection project's test suite (was constructed).
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: Write each YAML example to its own file and run trivy config with the check and lib/cloud/net.rego loaded as custom checks; every bad example must yield AWS-0107 and no good example may. For a check-writing task, the agent's check must give the same results on the same examples.
12
 
 
1
  # aquasecurity-trivy-checks
2
 
3
+ **What this is:** Files published by Aqua Security. Collection theme: Detection rule repositories with own tests (Sigma, YARA, Snort class).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Rego check + YAML good/bad examples (CloudFormation 1 good, 1 bad; Terraform 2 good, 1 bad) + Rego library.
6
 
7
  **Where it comes from:** Upstream: aquasecurity/trivy-checks. Kind: test or example files shipped with a tool. Test cases written by hand for the detection project's test suite (was constructed).
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Write each YAML example to its own file and run trivy config with the check and lib/cloud/net.rego loaded as custom checks; every bad example must yield AWS-0107 and no good example may. For a check-writing task, the agent's check must give the same results on the same examples.
12
 
campaign_files/2026-10-sc-science-reference-data-2/abacus-integrate-tests/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # abacus-integrate-tests
2
 
3
- **What this is:** 1 file published by ABACUS developers (DeepModeling community, AISI Beijing and Peking University). Collection theme: Chemistry and materials computational reference sets (molecular or crystal structures with computed properties).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: ABACUS reference result file (result.ref, text, one 'key value' per line); inputs INPUT, STRU, KPT (text); pseudopotential UPF (XML-like text).
6
 
7
  **Where it comes from:** Upstream name: ABACUS-integrate-tests. Upstream: deepmodeling/abacus-develop. Kind: test or example files shipped with a tool. Integration test of the ABACUS test suite (tool_fixture): inputs written by the developers, reference result produced by ABACUS and committed.
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: Give the agent INPUT, STRU, KPT and the pseudopotential and ask for the total energy and energy per atom of the cell (eV), its point group and space group and the number of irreducible k-points; energies pass when the absolute deviation from the reference is below the threshold; symmetry labels and the k-point count must match exactly.
12
 
 
1
  # abacus-integrate-tests
2
 
3
+ **What this is:** Files published by ABACUS developers (DeepModeling community, AISI Beijing and Peking University). Collection theme: Chemistry and materials computational reference sets (molecular or crystal structures with computed properties).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: ABACUS reference result file (result.ref, text, one 'key value' per line); inputs INPUT, STRU, KPT (text); pseudopotential UPF (XML-like text).
6
 
7
  **Where it comes from:** Upstream name: ABACUS-integrate-tests. Upstream: deepmodeling/abacus-develop. Kind: test or example files shipped with a tool. Integration test of the ABACUS test suite (tool_fixture): inputs written by the developers, reference result produced by ABACUS and committed.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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 INPUT, STRU, KPT and the pseudopotential and ask for the total energy and energy per atom of the cell (eV), its point group and space group and the number of irreducible k-points; energies pass when the absolute deviation from the reference is below the threshold; symmetry labels and the k-point count must match exactly.
12
 
campaign_files/2026-10-sc-science-reference-data-2/acwf-verification/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # acwf-verification
2
 
3
- **What this is:** 1 file published by AiiDA team (common workflows verification, Bosoni et al. 2023). Collection theme: Chemistry and materials computational reference sets (molecular or crystal structures with computed properties).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: JSON (BM_fit_data per 'Element-Prototype': min_volume in A^3, bulk_modulus_ev_ang3, bulk_deriv, E0, residuals; num_atoms_in_sim_cell; script_version 0.0.3), 384 entries.
6
 
7
  **Where it comes from:** Upstream: aiidateam/acwf-verification-scripts. Kind: test or example files shipped with a tool. Published verification data of the ACWF project (tool_fixture): EOS fits produced by WIEN2k and FLEUR through AiiDA workflows, averaged and committed as JSON.
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
  **Tools to open and check:** Check: Give the agent the element and prototype and ask for the PBE equation of state (V0, B0, B1 per formula unit) of that crystal, or give a code's E(V) curve and ask for the fit; score the result against the AE reference with nu (quantities_for_comparison.py line 301, prefactor 100, weights 1/20 and 1/400) and epsilon (line 227); excellent agreement when nu <= 0.10 and epsilon <= 0.06, good when nu <= 0.33 and epsilon <= 0.20, outlier when epsilon > 1.0 or nu > 1.65.
12
 
 
1
  # acwf-verification
2
 
3
+ **What this is:** Files published by AiiDA team (common workflows verification, Bosoni et al. 2023). Collection theme: Chemistry and materials computational reference sets (molecular or crystal structures with computed properties).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: JSON (BM_fit_data per 'Element-Prototype': min_volume in A^3, bulk_modulus_ev_ang3, bulk_deriv, E0, residuals; num_atoms_in_sim_cell; script_version 0.0.3), 384 entries.
6
 
7
  **Where it comes from:** Upstream: aiidateam/acwf-verification-scripts. Kind: test or example files shipped with a tool. Published verification data of the ACWF project (tool_fixture): EOS fits produced by WIEN2k and FLEUR through AiiDA workflows, averaged and committed as JSON.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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 element and prototype and ask for the PBE equation of state (V0, B0, B1 per formula unit) of that crystal, or give a code's E(V) curve and ask for the fit; score the result against the AE reference with nu (quantities_for_comparison.py line 301, prefactor 100, weights 1/20 and 1/400) and epsilon (line 227); excellent agreement when nu <= 0.10 and epsilon <= 0.06, good when nu <= 0.33 and epsilon <= 0.20, outlier when epsilon > 1.0 or nu > 1.65.
12
 
campaign_files/2026-10-sc-science-reference-data-2/agat-tests/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # agat-tests
2
 
3
- **What this is:** 4 files published by NBIS (National Bioinformatics Infrastructure Sweden), AGAT authors. Collection theme: Bioinformatics tool test and tutorial data (sequence, alignment, variant calling) with expected outputs.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Expected GTF 3 file, 9 lines (1614 bytes); input GFF3 2003 bytes, one human gene (AIG1, GENCODE ids) with a stop codon split across an exon boundary; expected GFF3 file, 13 lines (1198 bytes); input GFF3 1637 bytes with one gene and two mRNA isoforms on a MAKER scaffold; expected text table, 650 bytes: sensitivity and specificity for gene, mRNA, CDS, exon, five_prime_UTR and three_prime_UTR (values 1.00, 0.29, 0.29, 0.47, 0.33 and 0); inputs ref2.gff3 395 bytes and query2.gff3 12131 bytes; expected GFF3 file, 1062 bytes (gene, transcript and exon of MT-TL1 with merged_ attributes); inputs 1004 and 359 bytes.
6
 
7
  **Where it comes from:** Upstream: NBISweden/AGAT. Kind: test or example files shipped with a tool. Test fixture of the AGAT test suite (tool_fixture): input written for a stop-codon edge case, expected output produced by AGAT and committed as the regression reference; test fixture of the AGAT test suite (tool_fixture): small input with two isoforms, expected output produced by AGAT and committed; test fixture of the AGAT test suite (tool_fixture): reference and query annotations, expected table produced by AGAT and committed; test fixture of the AGAT test suite (tool_fixture): regression input from AGAT issue 457, expected output produced by AGAT and committed.
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 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:** Tools are named in the check written per file in the notes.
12
 
 
1
  # agat-tests
2
 
3
+ **What this is:** Files published by NBIS (National Bioinformatics Infrastructure Sweden), AGAT authors. Collection theme: Bioinformatics tool test and tutorial data (sequence, alignment, variant calling) with expected outputs.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Expected GTF 3 file, 9 lines (1614 bytes); input GFF3 2003 bytes, one human gene (AIG1, GENCODE ids) with a stop codon split across an exon boundary; expected GFF3 file, 13 lines (1198 bytes); input GFF3 1637 bytes with one gene and two mRNA isoforms on a MAKER scaffold; expected text table, 650 bytes: sensitivity and specificity for gene, mRNA, CDS, exon, five_prime_UTR and three_prime_UTR (values 1.00, 0.29, 0.29, 0.47, 0.33 and 0); inputs ref2.gff3 395 bytes and query2.gff3 12131 bytes; expected GFF3 file, 1062 bytes (gene, transcript and exon of MT-TL1 with merged_ attributes); inputs 1004 and 359 bytes.
6
 
7
  **Where it comes from:** Upstream: NBISweden/AGAT. Kind: test or example files shipped with a tool. Test fixture of the AGAT test suite (tool_fixture): input written for a stop-codon edge case, expected output produced by AGAT and committed as the regression reference; test fixture of the AGAT test suite (tool_fixture): small input with two isoforms, expected output produced by AGAT and committed; test fixture of the AGAT test suite (tool_fixture): reference and query annotations, expected table produced by AGAT and committed; test fixture of the AGAT test suite (tool_fixture): regression input from AGAT issue 457, expected output produced by AGAT and committed.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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:** Tools are named in the check written per file in the notes.
12
 
campaign_files/2026-10-sc-science-reference-data-2/bagel-test-inputs-and-unit-tests/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # bagel-test-inputs-and-unit-tests
2
 
3
- **What this is:** 1 file published by Toru Shiozaki group, QSimulate, BAGEL developers. Collection theme: Chemistry and materials computational reference sets (molecular or crystal structures with computed properties).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: BAGEL input JSON.
6
 
7
  **Where it comes from:** Upstream name: BAGEL test inputs (test/) and unit tests (src/testimpl). Upstream: qsimulate-open/bagel. Kind: test or example files shipped with a tool. Regression test of BAGEL (tool_fixture): input and reference committed by the developers.
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: Give the agent hf_svp_hf.json and the SVP basis and ask for the RHF total energy; pass when |E - (-99.84779026)| < 1.0e-8 hartree (test_main.cc lines 48 and 51).
12
 
 
1
  # bagel-test-inputs-and-unit-tests
2
 
3
+ **What this is:** Files published by Toru Shiozaki group, QSimulate, BAGEL developers. Collection theme: Chemistry and materials computational reference sets (molecular or crystal structures with computed properties).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: BAGEL input JSON.
6
 
7
  **Where it comes from:** Upstream name: BAGEL test inputs (test/) and unit tests (src/testimpl). Upstream: qsimulate-open/bagel. Kind: test or example files shipped with a tool. Regression test of BAGEL (tool_fixture): input and reference committed by the developers.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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 hf_svp_hf.json and the SVP basis and ask for the RHF total energy; pass when |E - (-99.84779026)| < 1.0e-8 hartree (test_main.cc lines 48 and 51).
12
 
campaign_files/2026-10-sc-science-reference-data-2/bcftools-test-pl/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # bcftools-test-pl
2
 
3
- **What this is:** 4 files published by Genome Research Ltd. (samtools/bcftools authors). Collection theme: Bioinformatics tool test and tutorial data (sequence, alignment, variant calling) with expected outputs.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Expected VCF, 31 lines; input TSV of 24 lines (ID, CHROM, POS, REF, ALT with fake ids rs001 to rs025) and a toy reference 23andme.fa of 287 bytes (contigs 1, 2 and Y of 150, 77 and 40 bases); expected VCF, 20 lines (701 bytes); input gVCF of 9 lines (404 bytes) with three blocks on contig 22, and reference gvcf.fa of 472 bytes (one contig 22 of 460 bases, header length 450); expected HAPS matrix, 5 lines (60 bytes); input VCF 4.1 of 9 lines (390 bytes), 5 chromosome X sites with samples NA00001 to NA00003; expected annotated VCF body, 1 line (85 bytes, header removed by grep -v ^#).
6
 
7
  **Where it comes from:** Upstream: samtools/bcftools. Kind: test or example files shipped with a tool. Test fixture of the upstream test suite (tool_fixture): input written for the case, expected output produced by the tool and committed as the regression reference.
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 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:** Tools are named in the check written per file in the notes.
12
 
 
1
  # bcftools-test-pl
2
 
3
+ **What this is:** Files published by Genome Research Ltd. (samtools/bcftools authors). Collection theme: Bioinformatics tool test and tutorial data (sequence, alignment, variant calling) with expected outputs.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Expected VCF, 31 lines; input TSV of 24 lines (ID, CHROM, POS, REF, ALT with fake ids rs001 to rs025) and a toy reference 23andme.fa of 287 bytes (contigs 1, 2 and Y of 150, 77 and 40 bases); expected VCF, 20 lines (701 bytes); input gVCF of 9 lines (404 bytes) with three blocks on contig 22, and reference gvcf.fa of 472 bytes (one contig 22 of 460 bases, header length 450); expected HAPS matrix, 5 lines (60 bytes); input VCF 4.1 of 9 lines (390 bytes), 5 chromosome X sites with samples NA00001 to NA00003; expected annotated VCF body, 1 line (85 bytes, header removed by grep -v ^#).
6
 
7
  **Where it comes from:** Upstream: samtools/bcftools. Kind: test or example files shipped with a tool. Test fixture of the upstream test suite (tool_fixture): input written for the case, expected output produced by the tool and committed as the regression reference.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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:** Tools are named in the check written per file in the notes.
12
 
campaign_files/2026-10-sc-science-reference-data-2/bertrand/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # bertrand
2
 
3
- **What this is:** 1 file published by coq-community. Collection theme: Formal mathematics proof libraries (Lean and Coq theorems with proofs that a pinned checker accepts).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Coq source files, unit 18083 bytes, 159108 bytes with companions; 221 Theorem, Lemma, Remark and Corollary declarations, no Admitted.
6
 
7
  **Where it comes from:** Upstream: coq-community/bertrand. Kind: real files published upstream. Proof development written by people (Laurent Thery and coq-community), maintained in coq-community with an opam package.
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: Replace the proof of the main theorem Bertrand (line 406) (and, for a harder variant, of chosen lemmas) with Admitted and hand the agent the files; the agent writes the proofs; pass when the pinned Coq compiles all files of the row with the _CoqProject mapping without error, no Admitted or admit remains, the theorem statements are byte-identical to the originals, and Print Assumptions on the main theorem lists only what it lists for the reference proof.
12
 
 
1
  # bertrand
2
 
3
+ **What this is:** Files published by coq-community. Collection theme: Formal mathematics proof libraries (Lean and Coq theorems with proofs that a pinned checker accepts).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Coq source files, unit 18083 bytes, 159108 bytes with companions; 221 Theorem, Lemma, Remark and Corollary declarations, no Admitted.
6
 
7
  **Where it comes from:** Upstream: coq-community/bertrand. Kind: real files published upstream. Proof development written by people (Laurent Thery and coq-community), maintained in coq-community with an opam package.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Replace the proof of the main theorem Bertrand (line 406) (and, for a harder variant, of chosen lemmas) with Admitted and hand the agent the files; the agent writes the proofs; pass when the pinned Coq compiles all files of the row with the _CoqProject mapping without error, no Admitted or admit remains, the theorem statements are byte-identical to the originals, and Print Assumptions on the main theorem lists only what it lists for the reference proof.
12
 
campaign_files/2026-10-sc-science-reference-data/bedtools2-test-suite/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # bedtools2-test-suite
2
 
3
- **What this is:** 1 file published by Aaron Quinlan and bedtools contributors. Collection theme: Bioinformatics tool test and tutorial data (sequence, alignment, variant calling) with expected outputs.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: BED6 (12 intervals) + shell test with expected BED inline.
6
 
7
  **Where it comes from:** Upstream: arq5x/bedtools2. Kind: test or example files shipped with a tool. Test data in the tool's own test suite (was tool_test_fixture).
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 bedtools merge with each option and compare the interval lists exactly (4, 2 and 2 lines); intervals with strand . must not merge into stranded runs; an independent interval-union script gives the same check.
12
 
 
1
  # bedtools2-test-suite
2
 
3
+ **What this is:** Files published by Aaron Quinlan and bedtools contributors. Collection theme: Bioinformatics tool test and tutorial data (sequence, alignment, variant calling) with expected outputs.
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: BED6 (12 intervals) + shell test with expected BED inline.
6
 
7
  **Where it comes from:** Upstream: arq5x/bedtools2. Kind: test or example files shipped with a tool. Test data in the tool's own test suite (was tool_test_fixture).
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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 bedtools merge with each option and compare the interval lists exactly (4, 2 and 2 lines); intervals with strand . must not merge into stranded runs; an independent interval-union script gives the same check.
12
 
campaign_files/2026-10-sw-tests-and-exercises-2/1brc-test-samples/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # 1brc-test-samples
2
 
3
- **What this is:** 3 files published by Gunnar Morling and contributors (The One Billion Row Challenge, 1brc). Collection theme: Performance benchmark suites (benchmark programs, inputs and reference outputs).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: One measurements sample (lines station;temperature) with its expected .out (one line {station=min/mean/max, ...}, stations sorted), the test.sh driver, the tocsv.sh normaliser and the baseline Java implementation.
6
 
7
  **Where it comes from:** Upstream: gunnarmorling/1brc. Kind: test or example files shipped with a tool. Test sample kept in the 1brc repository with an expected output produced by the reference implementation and checked in by the maintainers (tool_fixture).
8
 
9
- **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 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: The comparison is exact, any differing line fails.
12
 
 
1
  # 1brc-test-samples
2
 
3
+ **What this is:** Files published by Gunnar Morling and contributors (The One Billion Row Challenge, 1brc). Collection theme: Performance benchmark suites (benchmark programs, inputs and reference outputs).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: One measurements sample (lines station;temperature) with its expected .out (one line {station=min/mean/max, ...}, stations sorted), the test.sh driver, the tocsv.sh normaliser and the baseline Java implementation.
6
 
7
  **Where it comes from:** Upstream: gunnarmorling/1brc. Kind: test or example files shipped with a tool. Test sample kept in the 1brc repository with an expected output produced by the reference implementation and checked in by the maintainers (tool_fixture).
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: The comparison is exact, any differing line fails.
12
 
campaign_files/2026-10-sw-tests-and-exercises-2/activity-schema-aggregate-in-between-1/README.md CHANGED
@@ -1,12 +1,12 @@
1
  # activity-schema-aggregate-in-between-1
2
 
3
- **What this is:** 1 file published by Teghan Nightengale (tnightengale). Collection theme: Data-pipeline example projects with test data (small ETL and dbt-style projects with expected outputs).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Dbt test project with input and expected output seeds and dbt_utils.equality tests.
6
 
7
  **Where it comes from:** Upstream: tnightengale/dbt-activity-schema. Kind: test or example files shipped with a tool. tool_fixture: the project's own tests and expected outputs.
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: Dbt deps, dbt seed, dbt run, dbt test in integration_tests/ on the adapter chosen at pin time; pass = the dbt_utils.equality test on dataset__aggregate_in_between_1 passes (exact). As a task: hide the aggregate_in_between macro and ask for one that passes.
12
 
 
1
  # activity-schema-aggregate-in-between-1
2
 
3
+ **What this is:** Files published by Teghan Nightengale (tnightengale). Collection theme: Data-pipeline example projects with test data (small ETL and dbt-style projects with expected outputs).
4
 
5
  **What one unit is:** One upstream file per unit, kept at its path in the upstream repository or site. Format: Dbt test project with input and expected output seeds and dbt_utils.equality tests.
6
 
7
  **Where it comes from:** Upstream: tnightengale/dbt-activity-schema. Kind: test or example files shipped with a tool. tool_fixture: the project's own tests and expected outputs.
8
 
9
+ **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. Each file names 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: Dbt deps, dbt seed, dbt run, dbt test in integration_tests/ on the adapter chosen at pin time; pass = the dbt_utils.equality test on dataset__aggregate_in_between_1 passes (exact). As a task: hide the aggregate_in_between macro and ask for one that passes.
12