diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 34eedb29..0e926386 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -135,15 +135,15 @@ jobs: fi pip install ../../dist/*.whl - - if [ "${GITHUB_EVENT_NAME}" == "schedule" ] || [ "${GITHUB_EVENT_NAME}" == "workflow_dispatch" ]; then - curl -s https://raw.githubusercontent.com/Helicone/helicone/main/mitmproxy.sh | bash -s start - agbenchmark start --maintain - else + + if [ "${GITHUB_EVENT_NAME}" == "pull_request" ]; then agbenchmark start --maintain --mock agbenchmark start --improve --mock agbenchmark start --mock agbenchmark start --mock --category=retrieval + else + curl -s https://raw.githubusercontent.com/Helicone/helicone/main/mitmproxy.sh | bash -s start + agbenchmark start --maintain fi env: GITHUB_EVENT_NAME: ${{ github.event_name }} @@ -155,9 +155,10 @@ jobs: HELICONE_CACHE_ENABLED: true HELICONE_PROPERTY_AGENT: ${{ matrix.agent-name }} - - name: Upload logs as artifact + + - name: Upload reports if: always() uses: actions/upload-artifact@v3 with: - name: gpt-engineer-projects - path: agent/gpt-engineer/projects + name: ${{ matrix.agent-name }} + path: agent/${{ matrix.agent-name }}/agbenchmark