Periodic Evaluations #700
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: Periodic Evaluations | |
| on: | |
| schedule: | |
| - cron: "0 */4 * * *" # Run every 4 hours | |
| workflow_dispatch: # Allow manual triggering | |
| jobs: | |
| setup: | |
| runs-on: ubuntu-latest | |
| outputs: | |
| models: ${{ steps.set-models.outputs.models }} | |
| has_models: ${{ steps.set-models.outputs.has_models }} | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Set up Node.js | |
| uses: actions/setup-node@v4 | |
| with: | |
| node-version: "24" | |
| - name: Install Bun | |
| run: npm install -g bun@1.3.8 | |
| - name: Install dependencies | |
| run: bun install | |
| - id: set-models | |
| env: | |
| CONVEX_EVAL_URL: ${{ secrets.CONVEX_EVAL_URL }} | |
| OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }} | |
| CURSOR_API_KEY: ${{ secrets.CURSOR_API_KEY }} | |
| run: | | |
| bun run scripts/listPeriodicModels.ts --format json --output-file /tmp/periodic-models.json | |
| models=$(< /tmp/periodic-models.json) | |
| echo "models=$models" >> $GITHUB_OUTPUT | |
| if [ "$models" = "[]" ]; then | |
| echo "has_models=false" >> $GITHUB_OUTPUT | |
| else | |
| echo "has_models=true" >> $GITHUB_OUTPUT | |
| fi | |
| run-evals: | |
| needs: setup | |
| if: needs.setup.outputs.has_models == 'true' | |
| runs-on: ubuntu-latest | |
| timeout-minutes: 120 | |
| strategy: | |
| fail-fast: false | |
| max-parallel: 4 | |
| matrix: | |
| model: ${{ fromJson(needs.setup.outputs.models) }} | |
| env: | |
| ENVIRONMENT: ci | |
| OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }} | |
| CURSOR_API_KEY: ${{ secrets.CURSOR_API_KEY }} | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - name: Set up Node.js | |
| uses: actions/setup-node@v4 | |
| with: | |
| node-version: "24" | |
| - name: Install Bun | |
| run: npm install -g bun@1.3.8 | |
| - name: Install dependencies | |
| run: bun install | |
| - name: Create temp directory | |
| run: mkdir -p /tmp/convex-evals | |
| - name: Set main-branch-only environment variables | |
| if: github.ref == 'refs/heads/main' | |
| run: | | |
| echo "CONVEX_EVAL_URL=${{ secrets.CONVEX_EVAL_URL }}" >> $GITHUB_ENV | |
| echo "CONVEX_AUTH_TOKEN=${{ secrets.CONVEX_AUTH_TOKEN }}" >> $GITHUB_ENV | |
| - name: Set model | |
| run: | | |
| echo "MODELS=${{ matrix.model }}" >> $GITHUB_ENV | |
| - name: Run evaluations | |
| env: | |
| OUTPUT_TEMPDIR: /tmp/convex-evals | |
| run: bun run runner/index.ts | |
| - name: Run evaluations (no_guidelines experiment) | |
| env: | |
| OUTPUT_TEMPDIR: /tmp/convex-evals-no-guidelines | |
| EVALS_EXPERIMENT: no_guidelines | |
| run: bun run runner/index.ts |