3 commits

Author	SHA1	Message	Date
LLLin000	96fd9771ef	fix(ocr-quality): contract polish — 4 P0/P1 fixes ... 1. compute_use_cases output uses status/gates/reasons (not recommended/gate_results) 2. write_feedback validates every mark has result_hash and fulltext_hash 3. load_readiness_policy(policy=...) bypasses user override (reproducible) 4. confidence_and_fallbacks yellow for degraded=True or weak span coverage Minor: append_mark doesn't mutate caller dict; has_figure_evidence includes unmatched_legend_count and held_count.	2026-07-04 23:38:03 +08:00
LLLin000	c76f1569aa	feat(ocr-quality): add evaluate_readiness() with YAML policy evaluator ... Adds readiness policy evaluator layer: - load_readiness_policy() with deep-merge, importlib.resources, user override - _resolve_field() dotted path resolver, _apply_op() operator matcher - _check_hard_red() for hard-red rule evaluation - evaluate_readiness() with weighted scoring + status determination - compute_use_cases() for all 4 use cases (reading, qa, figure_table, chunking) - Default YAML policy with weights, hard_red rules, use_case gates - pyproject.toml: policies/*.yaml added to package-data - 7 tests (B1-B7) covering all paths 17/17 tests pass (A10 + B7).	2026-07-04 23:28:06 +08:00
LLLin000	51e16066c9	feat(ocr-quality): add build_quality_indicators() with 5 normalizers ... Creates new ocr_quality.py module as pure function layer between build_ocr_health() and evaluate_readiness(). All 5 indicators: - rendered_text_integrity - body_reference_structure - figure_table_integrity - metadata_frontmatter_quality - confidence_and_fallbacks 10 unit tests covering shape, thresholds, applicability, health_profile spelling, inventory precedence, and run_integrity.	2026-07-04 23:23:51 +08:00