n8nmermaid-update-check

name	n8nmermaid-update-check
description	Check for Python runtime and package updates in n8nmermaid. Handles FastAPI/Uvicorn updates and creates update PRs. Use when user asks to check for updates or upgrade dependencies.

n8nmermaid Update Check

Check for Python runtime and package updates in the FastAPI n8nmermaid service.

Workflow

1. Run version check script
2. Analyze results by category
3. Decide update strategy
4. Apply updates and create PR (or report blockers)

Step 1: Check Versions

uv run .agents/skills/n8nmermaid-update-check/scripts/fetch_versions.py .

Output includes:

• Python: current version, latest available
• Packages: categorized as patch/minor/major updates (via uv tree --outdated)

Step 2: Analyze Results

Python Updates

Same minor (patch): e.g., 3.13.0 → 3.13.1

• Safe to update, change the Dockerfile PYTHON_IMAGE default

New minor version: e.g., 3.13 → 3.14

• Check Python release notes for deprecations
• Verify package compatibility (especially native extensions)
• Test build locally before updating

Package Updates

Patch updates: Auto-update safe Minor updates: Usually safe, quick changelog review Major updates: Review changelogs carefully, especially:

• fastapi - Web framework, check migration guides
• uvicorn - ASGI server, usually backward compatible
• pydantic - Data validation, v2 had significant changes
• typer - CLI framework
• ruff - Linter, may add new rules

FastAPI + Starlette + Pydantic

These packages are tightly coupled. When updating FastAPI:

• Check compatible Starlette version
• Check compatible Pydantic version
• FastAPI release notes usually specify compatible versions

Step 3: Decision Tree

No updates available?
  → Report "All up to date" and stop

Only patch updates?
  → Auto-update: uv lock --upgrade, test, create PR

Minor updates present?
  → Update all, test build and API
  → If issues: revert problematic packages, report

Major updates present?
  → Fetch changelogs for major packages
  → Check for breaking changes
  → Safe: include in update
  → Breaking: report to user with migration notes

Python update available?
  → Check package compatibility
  → Test locally before updating Dockerfile

Step 4a: Apply Updates

# Create branch
git checkout -b update/deps-$(date +%Y%m%d)

# Update all packages
uv lock --upgrade

# For specific package updates:
uv add fastapi@latest uvicorn@latest

# For Python update, edit Dockerfile:
# Change: ARG PYTHON_IMAGE=python:3.13-slim → ARG PYTHON_IMAGE=python:3.14-slim

# Sync dependencies
uv sync

# Test the API
uv run uvicorn src.n8nmermaid.api.main:app --host 0.0.0.0 --port 8000

# Run linting
uv run ruff check .

# If everything passes, commit
git add -A
git commit -m "Update dependencies

Python: [version change if applicable]
Major: [list major updates]
Minor: [count] packages
Patch: [count] packages"

# Push branch for review
git push -u origin update/deps-$(date +%Y%m%d)

Step 4b: Report Blockers

If updates are blocked, report to user:

1. What major updates have breaking changes
2. Specific migration steps needed
3. Recommended order of updates

FastAPI Special Handling

Check for breaking changes before updating:

# Check FastAPI releases
gh api repos/tiangolo/fastapi/releases --jq '.[0:5] | .[].body' | head -100

Common breaking patterns:

• Dependency injection changes
• Request/Response model changes
• OpenAPI schema changes
• Middleware API changes

After FastAPI updates, verify:

• API starts: uv run uvicorn n8nmermaid.api.main:app
• Endpoints respond correctly
• OpenAPI docs generate (/docs)

Pydantic v2 Notes

If still on Pydantic v1, migration to v2 requires:

• Model syntax changes (.dict() → .model_dump())
• Validator decorator changes
• Config class changes

This project uses Pydantic v2, so minor updates should be safe.

Testing

After any updates, verify:

uv sync                    # Sync dependencies
uv run ruff check .        # Linting
uv run uvicorn n8nmermaid.api.main:app  # Start API
# Test endpoints manually or with curl

Release Notes

Production deploys are signed Forgejo releases from jwa91/vps-n8nmermaid-api. After dependency or Python runtime updates, the release package must build a Linux/amd64 Docker image from uv.lock, include the image archive, and avoid edge-side docker compose up --build.