Recipe: Add Travsr to CI

Run Travsr in your CI pipeline to catch blast-radius regressions, enforce dependency rules, or generate change impact reports on every PR.

GitHub Actions example

# .github/workflows/travsr.yml
name: Travsr graph check

on:
  pull_request:

jobs:
  blast-radius:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 2  # need HEAD~1 for delta index

      - name: Install Travsr
        run: npm install -g @travsr.com/travsr

      - name: Index repository
        # --semantic runs Phase B (call edges) before returning so the graph is
        # query-ready immediately. The git hook is installed but never fires in CI.
        run: travsr init --semantic

      - name: Check blast radius of changed files
        run: |
          CHANGED=$(git diff --name-only HEAD~1 HEAD | grep -E '\.(ts|rs|py|go)$' || true)
          if [ -z "$CHANGED" ]; then
            echo "No source files changed."
            exit 0
          fi

          for FILE in $CHANGED; do
            echo "=== Blast radius: $FILE ==="
            # --direction callers follows incoming edges: everything that depends on this file.
            travsr graph "$FILE" --direction callers --format json \
              | jq -r '.nodes | length as $n | "\($n) affected nodes"'
          done

      - name: Fail if a change affects more than 50 nodes
        run: |
          for FILE in $(git diff --name-only HEAD~1 HEAD | grep -E '\.(ts|rs)$'); do
            COUNT=$(travsr graph "$FILE" --direction callers --format json | jq '.nodes | length')
            if [ "$COUNT" -gt 50 ]; then
              echo "HIGH BLAST RADIUS: $FILE affects $COUNT nodes (> 50)"
              exit 1
            fi
          done

Caching the graph between runs

For large repositories, cache .travsr/graph.db between CI runs to avoid re-indexing from scratch:

      - uses: actions/cache@v4
        with:
          path: .travsr/graph.db
          key: travsr-${{ runner.os }}-${{ hashFiles('**/*.ts', '**/*.rs') }}
          restore-keys: travsr-${{ runner.os }}-

      - name: Index repository
        # travsr init is idempotent: it refreshes an existing graph or builds one
        # from scratch. With a cached .travsr/graph.db only changed files are re-parsed.
        run: travsr init --semantic

Dependency rule enforcement

Enforce architectural rules (e.g. "controllers must not import db directly"):

# --direction deps follows outgoing edges: what this file imports / depends on.
travsr graph "src/controllers/UserController.ts" --direction deps --format json \
  | jq -e '.nodes[] | select(.path | startswith("src/db/"))' >/dev/null \
  && { echo "VIOLATION: Controller imports db directly"; exit 1; } \
  || echo "OK: no direct db import"