[Perf] Enable zero-copy in to_torch() and to_numpy() #11
Workflow file for this run
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: Check test coverage for changes | |
| on: | |
| workflow_dispatch: | |
| pull_request: | |
| types: | |
| - opened | |
| - reopened | |
| - synchronize | |
| concurrency: | |
| group: ${{ github.workflow }}-${{ github.head_ref || github.ref }} | |
| cancel-in-progress: true | |
| jobs: | |
| check-test-coverage: | |
| name: Check test coverage for changes | |
| runs-on: ubuntu-latest | |
| steps: | |
| - uses: actions/checkout@v4 | |
| with: | |
| fetch-depth: 0 | |
| - name: Wait for builds to finish (dead-reckoning), to save AI cost if someone pushes many commits in a row | |
| run: sleep 1800 | |
| - name: Collect diffs and file lists | |
| id: changed | |
| env: | |
| BASE_REF: ${{ github.base_ref }} | |
| run: | | |
| MERGE_BASE=$(git merge-base "origin/$BASE_REF" HEAD) | |
| # Diff of source-only files (excluding tests/, benchmarks/, docs/, scripts/, misc/) | |
| git diff "$MERGE_BASE...HEAD" \ | |
| -- '*.py' '*.cpp' '*.h' '*.hpp' '*.c' '*.cc' '*.cu' \ | |
| ':!tests/' ':!benchmarks/' ':!docs/' ':!scripts/' ':!misc/' \ | |
| > /tmp/source_diff.patch | |
| # List of ALL changed files in the PR (including tests) | |
| git diff --name-only "$MERGE_BASE...HEAD" > /tmp/all_changed_files.txt | |
| # List of changed test files only | |
| git diff --name-only "$MERGE_BASE...HEAD" -- 'tests/' > /tmp/changed_test_files.txt | |
| if [ ! -s /tmp/source_diff.patch ]; then | |
| echo "skip=true" >> "$GITHUB_OUTPUT" | |
| echo "No source files changed (only tests/docs/scripts/etc)." | |
| else | |
| echo "skip=false" >> "$GITHUB_OUTPUT" | |
| echo "Source diff: $(wc -l < /tmp/source_diff.patch) lines" | |
| echo "All changed files: $(wc -l < /tmp/all_changed_files.txt)" | |
| echo "Changed test files: $(wc -l < /tmp/changed_test_files.txt)" | |
| fi | |
| - name: Install Cursor CLI | |
| if: steps.changed.outputs.skip != 'true' | |
| run: | | |
| curl https://cursor.com/install -fsS | bash | |
| echo "$HOME/.cursor/bin" >> $GITHUB_PATH | |
| - name: Check test coverage with Cursor agent | |
| if: steps.changed.outputs.skip != 'true' | |
| env: | |
| CURSOR_API_KEY: ${{ secrets.CURSOR_KEY_HUGH }} | |
| run: | | |
| RESULT=$(agent -p "$(cat <<'PROMPT' | |
| You are checking whether code changes in a PR have adequate test coverage. | |
| Inputs: | |
| - /tmp/source_diff.patch — unified diff of NON-TEST source files changed in this PR (.py, .cpp, .h, etc., excluding tests/) | |
| - /tmp/all_changed_files.txt — list of ALL files changed in this PR | |
| - /tmp/changed_test_files.txt — list of test files changed/added in this PR | |
| Your task: | |
| 1. Read the source diff to understand what functional code was added or modified. | |
| 2. Read the changed test files list to see what tests were added or modified in this PR. | |
| 3. For each meaningful source change (new function, new class, new method, changed behavior, new module), determine whether there is a corresponding test — either: | |
| a. An existing test file in the repo (under tests/) that already covers that code, OR | |
| b. A new/modified test file in this PR that covers that code. | |
| 4. Use the repo structure to find existing tests. The repo layout is: | |
| - Python source: python/quadrants/ | |
| - Python tests: tests/python/ | |
| - C++ source: quadrants/ (subdirs: codegen, ir, runtime, transforms, etc.) | |
| - C++ tests: tests/cpp/ (mirrors the source subdirs) | |
| Test files are typically named test_<feature>.py or test_<feature>.cpp. | |
| What to flag: | |
| - New public functions/methods/classes with no test coverage at all | |
| - New modules/files with no corresponding test file | |
| - Significant behavior changes (new branches, new error handling) with no test for the new behavior | |
| What NOT to flag: | |
| - Pure refactors that don't change behavior (renames, moves, reformatting) | |
| - Changes to internal/private helpers that are indirectly tested through public API tests | |
| - Trivial changes (imports, type hints, comments, docstrings, logging) | |
| - Config files, build files, CI files | |
| - Changes to test infrastructure itself (conftest.py, test utilities) | |
| - Bug fixes where the fix is obviously correct and narrow | |
| Be pragmatic. Not every line needs a test. Focus on substantial untested additions. | |
| Do NOT modify any files. | |
| Stop after finding 5 violations. | |
| If there are NO violations, your final output must start with the word PASS. | |
| If there ARE violations, your final output must start with the word FAIL, followed by a list of violations (up to 5) in the format: <filepath>:<function_or_class>: <what is untested> | |
| PROMPT | |
| )" --model claude-4.6-opus-high-thinking --mode ask --output-format text --trust) | |
| echo "$RESULT" | |
| if echo "$RESULT" | grep -qE "^FAIL([[:space:]]|$)"; then | |
| exit 1 | |
| fi |