Repository navigation

Standardize dataset metadata in DATASETS.toml - #70

Draft

user27182 wants to merge 112 commits into

masterfrom

maint/dataset-metadata

Draft

Standardize dataset metadata in DATASETS.toml#70
user27182 wants to merge 112 commits into
masterfrom
maint/dataset-metadata

Commits on Sep 5, 2026
- Add LICENSES/ with the full text of every licence used
  
  user27182
  and
  claude
  committed
  e52fe77
- Add DATASETS.toml describing every file under Data/
  
  user27182
  and
  claude
  committed
  c0bb136
- Migrate per-dataset licence files into DATASETS.toml
  
  user27182
  and
  claude
  committed
  2716b14
- Migrate per-dataset README and CITATION files into DATASETS.toml
  
  user27182
  and
  claude
  committed
  b91ceb8
- Add tools/validate_datasets.py
  
  user27182
  and
  claude
  committed
  e56db20
- Add CI workflow enforcing the dataset metadata format
  
  user27182
  and
  claude
  committed
  554f596
- Document the dataset metadata format
  
  user27182
  and
  claude
  committed
  b5661bd
Commits on Sep 6, 2026
- Close two ways past the validator's forbidden-file and coverage rules
  
  user27182
  and
  claude
  committed
  bc632a5
- Report a malformed table instead of raising
  
  user27182
  and
  claude
  committed
  8ee2bbf
- Point the Victorian goblet at the Thingiverse page it actually came from
  
  user27182
  and
  claude
  committed
  0f5fc17
- Record Stanford's cultural restriction on the dragon and Lucy
  
  user27182
  and
  claude
  committed
  03ae0fb
- Reclassify the top-level mars and stars textures
  
  user27182
  and
  claude
  committed
  280045b
- Describe ExportBunny as what it is
  
  user27182
  and
  claude
  committed
  23d0acb
- Record both halves of the UNISIM-II licence
  
  user27182
  and
  claude
  committed
  e86b247
- Restore the CPTAC acknowledgement statement
  
  user27182
  and
  claude
  committed
  995f21a
- Resolve four datasets that were recorded as undetermined
  
  user27182
  and
  claude
  committed
  24e930f
- Correct three claims that the data contradicts
  
  user27182
  and
  claude
  committed
  0bc5de8
- Repoint five source URLs that no longer support their claims
  
  user27182
  and
  claude
  committed
  404b7ba
- Record what verification found for four uncertain entries
  
  user27182
  and
  claude
  committed
  a8f3faa
- Quote the Stanford conditions of use verbatim
  
  user27182
  and
  claude
  committed
  ac3df4f
- Fix the Dataset Gallery link
  
  user27182
  and
  claude
  committed
  eb10302
- Document what `provenance` covers and how `path` patterns match
  
  user27182
  and
  claude
  committed
  89341d4
- Say what the root LICENSE covers, and narrow the REUSE claim
  
  user27182
  and
  claude
  committed
  2456f57
- Leave Data/ out of the CI working tree
  
  user27182
  and
  claude
  committed
  9c2c0fe
- Downgrade two Laser Design attributions to inferred
  
  user27182
  and
  claude
  committed
  95c3784
- Stop treating an SPDX exception as a licence identifier
  
  user27182
  and
  claude
  committed
  4a4047e
- Correct the Nefertiti licence: CC BY-SA 4.0, not CC BY-NC-SA 4.0
  
  user27182
  and
  claude
  committed
  438fdeb
- Correct the DICOM stack: MR forearm, CC BY 4.0
  
  user27182
  and
  claude
  committed
  3dff4b8
- Record the terms of use the Laser Design pages do carry
  
  user27182
  and
  claude
  committed
  6009df4
- Describe the Ivan Nikolov meshes as the files actually are
  
  user27182
  and
  claude
  committed
  ed96120
- Correct four descriptions the files or sources contradict
  
  user27182
  and
  claude
  committed
  294da85
- Resolve the sextant, and record two disclaimed origins
  
  user27182
  and
  claude
  committed
  5a92c3b
- Catch the two metadata filenames this branch itself deletes
  
  user27182
  and
  claude
  committed
  23e6626
- Correct the Unknown licence text and three stale sources
  
  user27182
  and
  claude
  committed
  ff8d555
- Stop asserting a copyright notice nobody published
  
  user27182
  and
  claude
  committed
  de6bd4f
- Give byte-identical files one licence
  
  user27182
  and
  claude
  committed
  b9a4e53
- Resolve skybox2 to the VTK Examples import
  
  user27182
  and
  claude
  committed
  ee996d3
- Read the Thingiverse licence version instead of inferring it
  
  user27182
  and
  claude
  committed
  964d104
- Correct four records against the files and archives themselves
  
  user27182
  and
  claude
  committed
  ac6f498
- Correct the acceptance test and the licence-reading guidance
  
  user27182
  and
  claude
  committed
  a1d3f23
- Check element types, wildcard-only paths and duplicate blobs
  
  user27182
  and
  claude
  committed
  1d4622d
- Raise the shark to verified, and record three smaller leads
  
  user27182
  and
  claude
  committed
  9f927d0
- Record the Viewpoint terms the cow files carry in their headers
  
  user27182
  and
  claude
  committed
  1d399b4
- Record 3GQP under the licence the Protein Data Bank uses
  
  user27182
  and
  claude
  committed
  929a22c
- Move two byte-identity notes to the entries they describe
  
  user27182
  and
  claude
  committed
  cb19774
- Correct five dataset types, the import month and the notch pair
  
  user27182
  and
  claude
  committed
  8931a3b
- Point the VTK Data source at a URL that resolves
  
  user27182
  and
  claude
  committed
  fc1769d
- Stop the worked examples demonstrating what was just corrected
  
  user27182
  and
  claude
  committed
  dd9efad
- Document how the provenance was established, and how to check it
  
  user27182
  and
  claude
  committed
  9d35bd9
- Credit the sky textures to Solar System Scope, not the MATLAB repository
  
  user27182
  and
  claude
  committed
  cf2afdb
- Pull five records out of the inherited buckets
  
  user27182
  and
  claude
  committed
  cd37e46
- Identify the multiblock test data, and correct three provenance sentences
  
  user27182
  and
  claude
  committed
  5774663
- Close the `?` wildcard hole and report scalar type errors
  
  user27182
  and
  claude
  committed
  6a3d195
- Fix the worked example, and name the Tango author the file records
  
  user27182
  and
  claude
  committed
  a4473f2
- Separate Pluto from the Solar System Scope textures
  
  user27182
  and
  claude
  committed
  0088e15
- Correct four descriptions and a citation I quoted wrong
  
  user27182
  and
  claude
  committed
  65e3889
- Record the Viewpoint terms on cowHead.vtp
  
  user27182
  and
  claude
  committed
  078cf02
- Repoint four datasets at lorensen/VTKExamples
  
  user27182
  and
  claude
  committed
  6e888c6
- Separate the generated wavy animation from the ParaView data
  
  user27182
  and
  claude
  committed
  b8a55ce
- Correct four more datasets that claimed to predate the fork
  
  user27182
  and
  claude
  committed
  c925938
- Report a malformed `license` table instead of crashing on it
  
  user27182
  and
  claude
  committed
  195a52f
- Drop a modification the Hertzian archive never had
  
  user27182
  and
  claude
  committed
  ebc9603
- Name the 3D Studio flamingo for what it is
  
  user27182
  and
  claude
  committed
  c6ae639
- Cite the archived Thingiverse pages, not the live ones
  
  user27182
  and
  claude
  committed
  760bb99
- Say where the Exodus Nemesis metadata actually lives
  
  user27182
  and
  claude
  committed
  f169ec0
- Date the Lincoln mask to the cast that was scanned
  
  user27182
  and
  claude
  committed
  c3d43b0
- Document the three techniques that found this round's errors
  
  user27182
  and
  claude
  committed
  34cbb16
- Credit the skybox cube map to Emil Persson under CC BY 3.0
  
  user27182
  and
  claude
  committed
  f912f4c
- Stop teaching that a hash match settles the licence
  
  user27182
  and
  claude
  committed
  70830bc
- Separate two hand-written PVD files from the ParaView ones
  
  user27182
  and
  claude
  committed
  652fa59
- Record the public-domain files as public domain
  
  user27182
  and
  claude
  committed
  908d268
- Record the St Helens DEM as the federal work it is
  
  user27182
  and
  claude
  committed
  3e20a66
- Separate the Cyberware horse from the VTK Examples bulk entry
  
  user27182
  and
  claude
  committed
  dfb4e3b
- Identify the OSM street graph and the OMF project file
  
  user27182
  and
  claude
  committed
  1466b09
- Rename `source_url` to `origin_url`
  
  user27182
  and
  claude
  committed
  80e7c9e
- Identify the Pluto texture as the New Horizons colour map
  
  user27182
  and
  claude
  committed
  0a1509d
- Trace the SGeMS grids to the Training Images Library
  
  user27182
  and
  claude
  committed
  1e3321c
- Stop describing the PyVista side as done
  
  user27182
  and
  claude
  committed
  6f79f13
- Match two descriptions to what their entries now hold
  
  user27182
  and
  claude
  committed
  632730d
- Identify the crater orthophoto as a LINZ Topo50 clip
  
  user27182
  and
  claude
  committed
  2ec7011
- Record two medical datasets from what they actually are
  
  user27182
  and
  claude
  committed
  9786771
- Identify the NIfTI samples and the Auto MPG exports
  
  user27182
  and
  claude
  committed
  070c440
- Read the terms at the head of the shark's chain
  
  user27182
  and
  claude
  committed
  7114a1b
- Check provenance across byte-identical files too
  
  user27182
  and
  claude
  committed
  23b86f6
- Repair claims that could not be re-checked as written
  
  user27182
  and
  claude
  committed
  16063eb
- Say where each licence text came from
  
  user27182
  and
  claude
  committed
  be7993a
- Flag an identifiable patient study, and a second hospital scan
  
  user27182
  and
  claude
  committed
  3c5cce9
- Record the NASA CFD cases as the federal work they are
  
  user27182
  and
  claude
  committed
  afe2fac
- Fold the duplicate combustor block into the NASA entry
  
  user27182
  and
  claude
  committed
  ce3b91f
- Attribute the frog to LBNL, and narrow the grant to what was given
  
  user27182
  and
  claude
  committed
  c9431ba
- Apply three more of the VTK textbook's data credits
  
  user27182
  and
  claude
  committed
  1fc6106
- Give the validator its first tests
  
  user27182
  and
  claude
  committed
  16d1b6c
- Stop three ways the validator could pass a file it never checked
  
  user27182
  and
  claude
  committed
  30f5170
- Fix the licence example that the new required key invalidated
  
  user27182
  and
  claude
  committed
  ace2abd
- Capture stderr in the job summary
  
  user27182
  and
  claude
  committed
  bc41397
Commits on Sep 7, 2026
- Merge remote-tracking branch 'origin/master' into maint/dataset-metadata
  user27182
  committed
  42d7d2a
- Record the procedural cow, and migrate its sidecars
  
  user27182
  and
  claude
  committed
  adaeaae
Commits on Sep 21, 2026
- Record the origins established for the held-back entries
  
  user27182
  and
  claude
  committed
  88e4149
Commits on Sep 29, 2026
- Harden the validator and test it against known-bad tables
  
  user27182
  and
  claude
  committed
  4476070
- Apply the final audit to the dataset records
  
  user27182
  and
  claude
  committed
  73f0e67
- Quote the terms each LicenseRef text rests on
  
  user27182
  and
  claude
  committed
  3832c0d
- Bring the docs in line with the audited table and validator
  
  user27182
  and
  claude
  committed
  869d9fb
- Reject embedded ** and guard a non-string name in the identical-files check
  
  user27182
  and
  claude
  committed
  ce2bd01
- Point licence urls at master, align head_2 and sparse_points with their sources, describe rather than narrate
  
  user27182
  and
  claude
  committed
  96d89af
- Link the woman scan's origin through the Wayback Machine and archive the SOLPE reference
  
  user27182
  and
  claude
  committed
  91bd1be
Commits on Sep 30, 2026
- Link every commit, pull request, issue and discussion the notes cite
  
  user27182
  and
  claude
  committed
  0ef4c8c
Commits on Oct 9, 2026
- Apply the verified research findings on the undetermined records
  user27182
  committed
  7c08bd2
Commits on Oct 10, 2026
- Add the dataset-provenance agent skill
  user27182
  committed
  3c54705
- Move the provenance skill to skills/ and add AGENTS.md as the agent entry point
  user27182
  committed
  339532e
- Link .claude/skills and .agents/skills to skills/
  user27182
  committed
  204ab8e
- Describe CLAUDE.md as the include it is
  user27182
  committed
  9b5540e
- Drop CLAUDE.md; Claude Code reads AGENTS.md when no CLAUDE.md exists
  user27182
  committed
  39743c4

Commits

Commits on Sep 5, 2026

Commits on Sep 6, 2026

Commits on Sep 7, 2026

Commits on Sep 21, 2026

Commits on Sep 29, 2026

Commits on Sep 30, 2026

Commits on Oct 9, 2026

Commits on Oct 10, 2026