Add graph expansion and cross-encoder reranking features

- Implemented GraphExpander to enhance search results with related symbols using precomputed neighbors.
- Added CrossEncoderReranker for second-stage search ranking, allowing for improved result scoring.
- Created migrations to establish necessary database tables for relationships and graph neighbors.
- Developed tests for graph expansion functionality, ensuring related results are populated correctly.
- Enhanced performance benchmarks for cross-encoder reranking latency and graph expansion overhead.
- Updated schema cleanup tests to reflect changes in versioning and deprecated fields.
- Added new test cases for Treesitter parser to validate relationship extraction with alias resolution.

codex-lens/src/codexlens/config.py

@@ -105,12 +105,22 @@ class Config:
 
     # Indexing/search optimizations
     global_symbol_index_enabled: bool = True  # Enable project-wide symbol index fast path
+    enable_merkle_detection: bool = True  # Enable content-hash based incremental indexing
+
+    # Graph expansion (search-time, uses precomputed neighbors)
+    enable_graph_expansion: bool = False
+    graph_expansion_depth: int = 2
 
     # Optional search reranking (disabled by default)
     enable_reranking: bool = False
     reranking_top_k: int = 50
     symbol_boost_factor: float = 1.5
 
+    # Optional cross-encoder reranking (second stage, requires codexlens[reranker])
+    enable_cross_encoder_rerank: bool = False
+    reranker_model: str = "cross-encoder/ms-marco-MiniLM-L-6-v2"
+    reranker_top_k: int = 50
+
     # Multi-endpoint configuration for litellm backend
     embedding_endpoints: List[Dict[str, Any]] = field(default_factory=list)
     # List of endpoint configs: [{"model": "...", "api_key": "...", "api_base": "...", "weight": 1.0}]