Skip to content

Pipeline: Generate the readme and contents files instead of keeping them by hand - #155

Merged
ehennestad merged 2 commits into
follow-overlay-move-out-of-internalfrom
generate-readme-and-contents-files
Sep 6, 2026
Merged

ehennestad merged 2 commits into
follow-overlay-move-out-of-internalfrom
generate-readme-and-contents-files

Conversation

@ehennestad

@ehennestad ehennestad commented Sep 6, 2026 •

Copy link
Copy Markdown
Collaborator

Stacked on #154.

The readme of each generated folder, and the Contents.m listing a model version's modules, describe generated content. They were kept by hand in the toolbox and copied over the build afterwards, which is a shape that only works if someone remembers them. Nobody did:

  • v4.0 and v5.0 have no Contents.m at all
  • only latest has readme files
  • the Contents.m for latest has been missing neuroimaging since that module appeared

Both are derived from the schemas now and written with the classes they describe, so every version gets them and they cannot drift again. The Contents.m module list is the same one that fills the Modules enumeration.

Applying a build goes back to a single copy, with no second pass to restore anything, and the toolbox no longer needs to carry the sources.

What the text says

The generated Contents.m reproduces the hand-written one byte for byte for v1.0, v2.0 and v3.0. latest differs by exactly the one missing module line.

Two sentences in the readme files changed deliberately:

before after
types ...of the openMINDS metadata model. ...of the openMINDS metadata framework.
enumerations ...for the openMINDS models and metadata types ...for the modules and metadata types of the openMINDS metadata framework.

openMINDS uses "metadata model" one level down, for each of core, SANDS, controlledTerms and the rest, and calls the whole thing the openMINDS metadata framework. Using it for the model named the wrong scope. The second is a slip: the enumerations generated are Modules and Types, not models.

"Metadata type" stays. The documentation says "schema type", but this toolbox deliberately reserves "schema" for the schema file and says "type" for the openMINDS type, which is what openminds.enum.Types and getTypeName are named after. Qualifying it as "metadata type" matches the framework's own "metadata" family without reintroducing "schema".

Follow-up in the toolbox

code/resources/content_files and code/resources/readme_files become dead once this lands and can be deleted. That deletion is best done after #135, since it moves the same folders and a deletion against main would conflict with it.

🤖 Generated with Claude Code

ehennestad and others added 2 commits September 6, 2026 14:41
…by hand

The readme of each generated folder and the Contents.m listing a version's
modules describe generated content, but were maintained by hand in the
toolbox and copied over the build afterwards. Nobody remembered them when a
model version was added: v4.0 and v5.0 had no Contents.m at all, only latest
had readme files, and the Contents.m for latest had been missing neuroimaging
since that module appeared.

Both are derived from the schemas now and written with the classes they
describe, so every version has them and they cannot drift. Applying a build is
one copy again, with no second pass to restore anything.

The generated files reproduce the hand-written ones exactly for v1.0, v2.0 and
v3.0, and for the three readme files. latest differs by the module that was
missing.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The readme for the types folder called the model a "metadata model". openMINDS
uses that name one level down, for each of core, SANDS, controlledTerms and
the rest, and calls the whole thing the openMINDS metadata framework. Using it
for the model named the wrong scope.

The enumerations readme said "the openMINDS models and metadata types", where
the enumerations generated are Modules and Types. That looks like a slip for
"modules", so it now names both, and the framework they belong to.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@ehennestad
ehennestad merged commit 432f0c6 into pipeline Sep 6, 2026
@ehennestad
ehennestad deleted the generate-readme-and-contents-files branch September 6, 2026 18:23
@ehennestad ehennestad changed the title Generate the readme and contents files instead of keeping them by hand Pipeline: Generate the readme and contents files instead of keeping them by hand Sep 6, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant