Background
PR #46 fixed broken vocabulary links by regenerating docs from TTL sources using vocab2md.py. This works, but commits ~6,400 lines of generated markdown files.
Suggested Improvements
1. Generate at build time, not commit time
Current state:
- Generated files committed to
models/generated/vocabularies/ and models/generated/extensions/
- Any TTL source changes require manual regeneration and re-commit
Proposed:
- Add
models/generated/ to .gitignore
- Run
scripts/generate_vocab_docs.sh as part of Quarto pre-render or CI build
- Generated files only exist in build artifacts, not in repo
Benefits:
- Smaller repo size
- Always fresh from source TTL files
- No manual regeneration step
2. Clean up script comments
scripts/generate_vocab_docs.sh has dead code that should be removed:
# Remove these lines (the `vocab` CLI is unavailable):
# vocab markdown "${SOURCE_BASE}${src}" > "${DEST_FOLDER}${fname}"
Related
🤖 Created by Claude Code