incremental-indexing

name	incremental-indexing
description	Implement hash-based incremental indexing for codebase intelligence MCP to avoid full re-index and keep graph data fresh efficiently.

Incremental Indexing

When to Use

• Re-index repositories frequently with minimal cost.

Procedure

1. Detect changed files by commit SHA and content hash.
2. Recompute entities/edges only for impacted files.
3. Re-link affected transitive edges (imports/calls) safely.
4. Mark stale nodes/edges and clean up in controlled phase.
5. Store run stats and diff summary.

Correctness Checks

• No orphan edges after incremental run.
• Deterministic results between full vs incremental for same revision.
• Safe rollback strategy on partial failure.