fix: switch default test model to llama-3.1-8b-instant and add delay … #60
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: Eval | |
| on: | |
| push: | |
| branches: [main] | |
| pull_request: | |
| branches: [main] | |
| workflow_dispatch: | |
| jobs: | |
| eval: | |
| runs-on: ubuntu-latest | |
| services: | |
| postgres: | |
| image: pgvector/pgvector:pg16 | |
| env: | |
| POSTGRES_USER: postgres | |
| POSTGRES_PASSWORD: postgres | |
| POSTGRES_DB: codex_db | |
| ports: | |
| - 5432:5432 | |
| options: >- | |
| --health-cmd pg_isready | |
| --health-interval 10s | |
| --health-timeout 5s | |
| --health-retries 5 | |
| steps: | |
| - uses: actions/checkout@v4 | |
| - uses: actions/setup-python@v5 | |
| with: | |
| python-version: "3.12" | |
| - name: Install dependencies | |
| working-directory: codex-backend | |
| run: | | |
| python -m pip install --upgrade pip | |
| pip install -r requirements.txt | |
| - name: Run upload pipeline unit tests | |
| working-directory: codex-backend | |
| run: python -m unittest tests/test_upload_pipeline.py | |
| env: | |
| SUPABASE_URL: https://dummy-project.supabase.co | |
| SUPABASE_ANON_KEY: dummy-anon-key | |
| GROQ_API_KEY: dummy-key | |
| DB_URL: postgresql+asyncpg://u:p@localhost/db | |
| - name: Seed database schema | |
| working-directory: codex-backend | |
| run: | | |
| python -c " | |
| from sqlalchemy import create_engine, text | |
| engine = create_engine('postgresql+psycopg://postgres:postgres@localhost:5432/codex_db') | |
| with open('supabase/seed.sql') as f: | |
| sql = f.read() | |
| with engine.connect() as conn: | |
| conn.execute(text(sql)) | |
| conn.commit() | |
| print('Database seeded successfully') | |
| " | |
| - name: Run golden query test | |
| working-directory: codex-backend | |
| run: python tests/test_golden.py | |
| env: | |
| DB_URL: postgresql+psycopg://postgres:postgres@localhost:5432/codex_db | |
| GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }} | |
| GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }} | |
| SUPABASE_URL: https://dummy-project.supabase.co | |
| SUPABASE_ANON_KEY: dummy-anon-key | |
| - name: Run rigorous sweep | |
| working-directory: codex-backend | |
| run: python tests/test_rigorous.py | |
| env: | |
| DB_URL: postgresql+psycopg://postgres:postgres@localhost:5432/codex_db | |
| GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }} | |
| GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }} | |
| SUPABASE_URL: https://dummy-project.supabase.co | |
| SUPABASE_ANON_KEY: dummy-anon-key | |
| - name: Run RAGAS eval | |
| working-directory: codex-backend | |
| run: python eval/ragas_eval.py | |
| continue-on-error: true | |
| env: | |
| DB_URL: postgresql+psycopg://postgres:postgres@localhost:5432/codex_db | |
| GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }} | |
| GOOGLE_API_KEY: ${{ secrets.GOOGLE_API_KEY }} | |
| SUPABASE_URL: https://dummy-project.supabase.co | |
| SUPABASE_ANON_KEY: dummy-anon-key |