perf: optimize cache freshness, indexing, and search hot paths

[DevvGwardo]

Summary

This PR improves runtime performance and cache correctness across core indexing/search paths.

✅ Changes made

• peppy/cache.py

  • Replaced md5 cache keying with sha256.
  • Added optional ttl_seconds support to expire stale cache entries.
  • Added cache freshness signature validation based on indexed file stats (count, max mtime, total size).
  • Added safer UTF-8 reads/writes for cache files.

• peppy/indexer.py

  • Optimized ignore-path checks with precomputed exact-vs-glob rule buckets.
  • Switched gitignore matching to stable relative POSIX paths.
  • Added adaptive default worker sizing when max_workers is not provided.
  • Precomputes and stores symbol_types and file_extensions in index metadata.

• peppy/searcher.py

  • Added robust file-pattern matching against relative path, basename, and full path.
  • Reduced grep overhead with tighter matching flow and lower per-line allocations.
  • Builds context payload only when context_lines > 0.
  • Reuses precomputed index statistics when present.

Why

• Reduce repeated compute in indexing/search hot paths.
• Improve cache correctness so changed codebases don’t silently reuse stale index data.
• Improve scaling behavior for larger repositories.

Notes

• These are non-breaking internal performance/correctness improvements.
• Existing API surface remains unchanged.