feat(codexlens): Improve search defaults and add explicit SPLADE mode

Config changes:
- Disable SPLADE by default (slow ~360ms), use FTS instead
- Enable use_fts_fallback by default for faster sparse search

CLI improvements:
- Fix duplicate index_app typer definition
- Add cascade_search dispatch for cascade method
- Rename 'mode' to 'method' in search output
- Mark embeddings-status, splade-status as deprecated
- Add enable_splade and enable_cascade to search options

Hybrid search:
- Add enable_splade parameter for explicit SPLADE mode
- Add fallback handling when SPLADE is requested but unavailable

codex-lens/src/codexlens/config.py

@@ -118,13 +118,13 @@ class Config:
     embedding_use_gpu: bool = True  # For fastembed: whether to use GPU acceleration
 
     # SPLADE sparse retrieval configuration
-    enable_splade: bool = True  # Enable SPLADE as default sparse backend
+    enable_splade: bool = False  # Disable SPLADE by default (slow ~360ms, use FTS instead)
     splade_model: str = "naver/splade-cocondenser-ensembledistil"
     splade_threshold: float = 0.01  # Min weight to store in index
     splade_onnx_path: Optional[str] = None  # Custom ONNX model path
 
     # FTS fallback (disabled by default, available via --use-fts)
-    use_fts_fallback: bool = False  # Use FTS instead of SPLADE
+    use_fts_fallback: bool = True  # Use FTS for sparse search (fast, SPLADE disabled)
 
     # Indexing/search optimizations
     global_symbol_index_enabled: bool = True  # Enable project-wide symbol index fast path