Evaluate redundancy between pre-commit hooks and GitHub CI/CD workflows

[jonpspri]

Problem Statement

There appears to be significant overlap between our pre-commit hooks and GitHub CI/CD workflows, particularly in the quality checks. This redundancy may be creating unnecessary maintenance overhead.

Current Redundancies

Quality Checks in Both Systems

Pre-commit hooks (.pre-commit-config.yaml):

• Ruff linting (--fix --exit-non-zero-on-fix)
• Ruff formatting
• MyPy type checking
• MCP documentation compliance checks

GitHub CI/CD (.github/workflows/test.yml):

• Ruff linting (ruff check src/ tests/)
• Ruff format check (ruff format --check src/ tests/)
• MyPy type checking (mypy src/databeak/)
• MCP documentation compliance checks (separate job)

Questions to Address

1. Should we simplify CI/CD? Could we replace the individual quality check steps with a single pre-commit run --all-files command?

2. What are the trade-offs?

   • Pro: Single source of truth for quality checks, easier maintenance
   • Con: Less granular GitHub Actions output, potentially slower feedback
   • Con: Pre-commit uses --fix mode, CI/CD should be read-only

3. What should remain separate?

   • Test execution (unit/integration) - clearly separate concern
   • Multi-version testing (Python 3.12, 3.13) - matrix strategy
   • Coverage reporting to Codecov - CI/CD specific

4. Hybrid approach? Could we:

   • Use pre-commit run --all-files in CI/CD for most checks
   • Override specific hooks to use read-only mode (e.g., --check for ruff)
   • Keep tests as separate jobs

Proposed Investigation

• Research best practices for pre-commit + GitHub Actions integration
• Test pre-commit run --all-files in CI environment
• Evaluate impact on GitHub Actions UI/output clarity
• Measure performance differences
• Consider using pre-commit.ci service as alternative

Additional Context

Pre-commit strengths:

• Runs locally before commits (shift-left testing)
• Auto-fix capabilities
• Comprehensive hook ecosystem
• pre-commit.ci for automated PR fixes

GitHub Actions strengths:

• Matrix testing across Python versions
• Integration with GitHub UI
• Codecov integration
• Can run on different triggers (push, PR, schedule)

Expected Outcome

A clear strategy for:

1. Which checks should run where
2. How to minimize duplication while maintaining effectiveness
3. Whether to adopt pre-commit run --all-files in CI/CD or keep current approach
4. Documentation of the rationale for our chosen approach

---

Labels: enhancement, ci/cd, maintenance, discussion