fix(diagnosis): validate contexts before chunking analysis

[Nithin00614]

Summary

Fixes #62.

This PR ensures that ChunkingQualityAnalyzer receives validated string contexts instead of raw evaluation output.

Changes

• Added context validation before invoking ChunkingQualityAnalyzer
• Filtered non-string context entries before analysis
• Added a regression test covering mixed valid and invalid context inputs

Validation

• Added regression test
• Diagnosis test suite passes
• Ruff passes
• Mypy passes

[himanshu231204]

/oc review pr

[opencode-agent]

PR Review: fix/validated-contexts

Summary

Fixes #62 — ensures ChunkingQualityAnalyzer receives validated list[str] contexts instead of raw evaluation output that may contain non-string items.

Changes

File	Δ	What
analyzer.py	+43/-12	New _extract_validated_contexts() helper; metadata validation; context filtering in _extract_scores
test_integration.py	+49/-1	Regression test for context validation; Path import moved under TYPE_CHECKING

Review

Correctness: The logic is sound. _extract_validated_contexts filters to isinstance(context, str) and _extract_scores now does the same for counting/length metrics. Metadata is validated to dict or None before passing downstream. The regression test directly verifies non-string entries are stripped.

Minor issues:

1. Duplicated filtering logic — _extract_scores (lines 198-207) and _extract_validated_contexts (lines 220-235) perform identical context extraction + string filtering. Consider having _extract_scores delegate to _extract_validated_contexts to DRY it up.

2. Indentation inconsistency — analyzer.py:130-134:

                chunking_issues = self._chunking_analyzer.analyze(
                        question,        # 8-space indent (was 4)
                        contexts,
                        metadata,
                )

This uses an 8-space hanging indent while the rest of the file uses 4-space (e.g., the DiagnosisReport call at line 150). Should be consistent with project style.

Tests: The regression test test_chunking_receives_validated_contexts is well-written — uses a capture proxy instead of mocking the full analyzer, and covers string/dict/int/None cases.

Verdict

Approved with minor nits. Fix the indentation and optionally DRY up the duplicated context filtering.

opencode session  |  github run

[github-actions]

🎉 Congratulations @Nithin00614!

Your pull request has been successfully merged into main. 🚀

Thank you for contributing to OpenAgentHQ and helping improve the project.

We truly appreciate your contribution and hope to see you back with more amazing PRs!

Happy Open Sourcing! ❤️

🌟 This is your first merged contribution to this repository.
Welcome to the OpenAgentHQ contributors family!