Add github actions for QC checks

[cmungall]

We are lacking standard linkml github actions

• qc checks (may need customized for this repo)
• deploy_docs
• pypi

Use the standard https://github.com/linkml/linkml-project-copier -- but this may need tweaking as this repo doesn't fully conform (e.g still uses Makefiles not justfile for some things)

[github-actions]

Issue Summary

This issue requests adding standard LinkML GitHub Actions workflows to the aidocs repository, specifically:

• QC checks (customized for this repo)
• Documentation deployment
• PyPI publishing

Current State

Existing Workflows:
The repository currently has 6 GitHub Actions workflows:

• deploy_documentation.yml - Already handles MkDocs documentation deployment to GitHub Pages
• dragon-ai.yml - AI agent automation
• claude.yml, claude-issue-triage.yml, claude-issue-summarize.yml, claude-code-review.yml - Claude AI integrations

Missing Workflows:

• ✗ QC checks (linting, testing, validation)
• ✓ Documentation deployment (already exists, though not using LinkML template)
• ✗ PyPI publishing workflow

LinkML Project Copier Template

The linkml-project-copier template provides 4 standard workflows:

1. main.yaml - Primary CI/CD with QC checks:

   • Multi-version Python testing (3.9-3.13)
   • Runs just test for test suite execution
   • Uses uv for dependency management

2. deploy-docs.yaml - Documentation deployment to GitHub Pages

3. test_pages_build.yaml (conditional) - Pre-deployment documentation build testing

4. pypi-publish.yaml (conditional) - Automated PyPI package publishing

Implementation Considerations

Customization Needed:

1. No justfile present - The current repo uses neither Makefile nor justfile. The LinkML template assumes just commands (e.g., just test, just lint). This would need adaptation.

2. Different project structure - This is a documentation-focused MkDocs project, not a typical LinkML schema project. Standard LinkML QC checks may need significant customization:

   • No schema validation needed
   • Focus on documentation quality (link checking, markdown linting)
   • Python package is minimal (currently just aidocs module)

3. PyPI relevance - Consider whether PyPI publishing is needed. The current pyproject.toml is minimal with only MkDocs dependencies.

Recommended Approach

Rather than directly copying LinkML template workflows, consider:

1. QC Workflow - Create custom checks appropriate for documentation:

   • Markdown linting (markdownlint)
   • Link validation (using mkdocs-htmlproofer or similar)
   • Python code quality checks if needed (ruff, mypy)
   • MkDocs build validation

2. Keep existing deploy_documentation.yml - It's already working and uses modern uv tooling

3. PyPI workflow - Only add if there's a clear need to publish the aidocs package to PyPI

4. Build tooling choice - Decide between:

   • Adding a justfile to align with LinkML ecosystem conventions
   • Using GitHub Actions run steps directly
   • Creating a Makefile for compatibility

Would you like me to propose specific QC workflow implementations tailored for this documentation-focused repository?