Skip to content

feat(M92): Test Suite Generation from Divergent Samples#199

Open
hlin99 wants to merge 2 commits into
mainfrom
feat/m92-test-suite-gen
Open

feat(M92): Test Suite Generation from Divergent Samples#199
hlin99 wants to merge 2 commits into
mainfrom
feat/m92-test-suite-gen

Conversation

@hlin99
Copy link
Copy Markdown
Member

@hlin99 hlin99 commented Apr 6, 2026

Closes #198

M92: Test Suite Generation from Divergent Samples

Generate reusable test datasets from divergent samples in batch reports.

Changes

  • test_suite_gen.py: generate_suite(), write_suite(), format_suite_summary(), SuiteEntry, GenerateSuiteConfig
  • CLI: xpyd-acc generate-suite --report <path> --output <path>
  • Filters: --classification, --min-logprob-gap, --max-samples, --include-expected
  • Output: JSONL compatible with batch-compare --dataset
  • 18 tests covering generation, filtering, round-trip, file I/O, formatting
  • ROADMAP: M91 marked ✅, M92 added

Testing

  • All 1442 existing tests pass
  • 18 new tests pass
  • Lint clean

Copy link
Copy Markdown

@hlin99-Review-Bot hlin99-Review-Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

CI lint job is failing — isort found incorrectly sorted imports in two files:

  • src/xpyd_acc/cli/__init__.py
  • tests/test_test_suite_gen.py

Please run isort src tests and push the fix. All other checks (tests on 3.10/3.11/3.12) pass ✅

Lint job log

@hlin99
Copy link
Copy Markdown
Member Author

hlin99 commented May 3, 2026

Fixed in 00d4ed8. Ran isort to correct import ordering.

hlin99 and others added 2 commits May 3, 2026 10:31
- generate_suite() extracts divergent samples from batch reports into reusable JSONL datasets
- SuiteEntry and GenerateSuiteConfig dataclasses for structured suite generation
- Filters: --classification, --min-logprob-gap, --max-samples, --include-expected
- write_suite() outputs JSONL compatible with batch-compare --dataset
- format_suite_summary() for human-readable generation summary
- CLI: xpyd-acc generate-suite --report <path> --output <path>
- 18 tests covering generation, filtering, round-trip, file I/O, formatting
- Updated ROADMAP.md: M91 ✅, added M92
@hlin99 hlin99 force-pushed the feat/m92-test-suite-gen branch from 00d4ed8 to a50102a Compare May 3, 2026 02:31
@hlin99
Copy link
Copy Markdown
Member Author

hlin99 commented May 3, 2026

Fixed: Rebased onto latest main (was 2 commits behind). isort check passes. Ready for review.

@hlin99 hlin99 dismissed hlin99-Review-Bot’s stale review May 3, 2026 02:44

isort issues fixed in commit 00d4ed8

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

feat(M92): Test Suite Generation from Divergent Samples

2 participants