mirror of
https://github.com/google-gemini/gemini-cli.git
synced 2026-03-25 05:21:03 -07:00
96 lines
3.0 KiB
YAML
96 lines
3.0 KiB
YAML
name: 'Evals: PR Impact'
|
|
|
|
on:
|
|
pull_request:
|
|
paths:
|
|
- 'packages/core/src/prompts/**'
|
|
- 'packages/core/src/tools/**'
|
|
- 'packages/core/src/agents/**'
|
|
- 'evals/**'
|
|
|
|
permissions:
|
|
contents: 'read'
|
|
checks: 'write'
|
|
pull-requests: 'write'
|
|
|
|
jobs:
|
|
eval-impact:
|
|
name: 'Eval Impact Analysis'
|
|
runs-on: 'gemini-cli-ubuntu-16-core'
|
|
if: "github.repository == 'google-gemini/gemini-cli'"
|
|
steps:
|
|
- name: 'Checkout'
|
|
uses: 'actions/checkout@v4'
|
|
|
|
- name: 'Set up Node.js'
|
|
uses: 'actions/setup-node@v4'
|
|
with:
|
|
node-version-file: '.nvmrc'
|
|
cache: 'npm'
|
|
|
|
- name: 'Install dependencies'
|
|
run: 'npm ci'
|
|
|
|
- name: 'Build project'
|
|
run: 'npm run build'
|
|
|
|
- name: 'Create logs directory'
|
|
run: 'mkdir -p evals/logs'
|
|
|
|
- name: 'Run Evals (Lite)'
|
|
env:
|
|
GEMINI_API_KEY: '${{ secrets.GEMINI_API_KEY }}'
|
|
RUN_EVALS: 'true'
|
|
run: |
|
|
# Run 1 attempt on critical models for speed in PR
|
|
MODELS=("gemini-3.1-pro-preview-customtools" "gemini-3-flash-preview")
|
|
for model in "${MODELS[@]}"; do
|
|
echo "Running evals for $model..."
|
|
mkdir -p "evals/logs/eval-logs-$model-1"
|
|
# Use || true to ensure the loop continues even if tests fail
|
|
GEMINI_MODEL=$model npm run test:all_evals -- --outputFile.json="evals/logs/eval-logs-$model-1/report.json" || true
|
|
done
|
|
|
|
- name: 'Generate Impact Report'
|
|
id: 'generate-report'
|
|
if: 'always()'
|
|
env:
|
|
GH_TOKEN: '${{ secrets.GITHUB_TOKEN }}'
|
|
run: |
|
|
if [ ! -d "evals/logs" ] || [ -z "$(ls -A evals/logs)" ]; then
|
|
echo "No logs found, skipping report generation."
|
|
exit 0
|
|
fi
|
|
echo "<!-- eval-impact-report -->" > report.md
|
|
node scripts/aggregate_evals.js evals/logs --compare-main --pr-comment >> report.md
|
|
cat report.md >> "$GITHUB_STEP_SUMMARY"
|
|
|
|
# Check if there are any regressions (🔴)
|
|
if grep -q "🔴" report.md; then
|
|
echo "REGRESSION_DETECTED=true" >> "$GITHUB_ENV"
|
|
fi
|
|
|
|
- name: 'Comment on PR'
|
|
if: 'always()'
|
|
env:
|
|
GH_TOKEN: '${{ secrets.GITHUB_TOKEN }}'
|
|
run: |
|
|
if [ ! -f "report.md" ]; then
|
|
echo "No report found, skipping comment."
|
|
exit 0
|
|
fi
|
|
PR_NUMBER=${{ github.event.pull_request.number }}
|
|
EXISTING_COMMENT=$(gh pr view $PR_NUMBER --json comments --jq '.comments[] | select(.body | contains("eval-impact-report")) | .id' | head -n 1)
|
|
|
|
if [ -n "$EXISTING_COMMENT" ]; then
|
|
gh pr comment $PR_NUMBER --body-file report.md
|
|
else
|
|
gh pr comment $PR_NUMBER --body-file report.md
|
|
fi
|
|
|
|
- name: 'Fail if regression detected'
|
|
if: "env.REGRESSION_DETECTED == 'true'"
|
|
run: |
|
|
echo "Behavioral evaluation regressions detected. Please check the impact report."
|
|
exit 1
|