Future Enhancements¶

Back to main README

Coverage path lookup indexing¶

Problem¶

Single-file tools (summary, raw, detailed, uncovered) call CoverageLineResolver#resolve_key, which normalizes every key in the coverage map on each lookup. That makes per-file queries O(n) in the number of files and can become O(n^2) across repeated MCP/CLI requests.

Proposed approach¶

Precompute a normalized-key index once per resolver (or per cached model data). Use that index for O(1) lookups while preserving ambiguity detection when multiple original keys normalize to the same value.

Why this matters¶

Large resultsets or frequent interactive queries can feel sluggish due to repeated normalization and full-map scans. Indexing would improve responsiveness without changing output semantics.