feat: make embedding model configurable for semantic search

- Add SemanticSearchConfig to settings.py with embedding model options
- Update WorldMapper to accept embedding_model parameter
- Update main.py to pass model from config to WorldMapper
- Add documentation for available embedding models:
  - sentence-transformers/all-MiniLM-L6-v2 (default, fast)
  - sentence-transformers/all-mpnet-base-v2 (better quality)
  - google/embeddinggemma-300m (state-of-the-art)
- Update config.yaml and config.example.yaml with new options

Users can now choose the best embedding model for their needs.

docs/configuration.md

@@ -71,7 +71,12 @@ mapper:
       codec: "h264"
   
   semantic_search:
-    model: "all-MiniLM-L6-v2"    # Sentence transformer model
+    # Embedding model for semantic room search
+    # Available models:
+    #   - "sentence-transformers/all-MiniLM-L6-v2" (default, fast, 22MB)
+    #   - "sentence-transformers/all-mpnet-base-v2" (better quality, 418MB)
+    #   - "google/embeddinggemma-300m" (state-of-the-art, requires transformers>=4.40)
+    model: "sentence-transformers/all-MiniLM-L6-v2"
     top_k: 5                     # Number of search results
     similarity_threshold: 0.7    # Minimum similarity score
 ```

examples/config.example.yaml

@@ -35,10 +35,15 @@ mapper:
       fps: 30
       codec: "h264"
   
-  semantic_search:
-    model: "all-MiniLM-L6-v2"
-    top_k: 5
-    similarity_threshold: 0.7
+  semantic_search:
+    # Embedding model for semantic room search
+    # Options:
+    #   - "sentence-transformers/all-MiniLM-L6-v2" (default, fast, 22MB)
+    #   - "sentence-transformers/all-mpnet-base-v2" (better quality, 418MB)
+    #   - "google/embeddinggemma-300m" (state-of-the-art, requires transformers>=4.40)
+    model: "sentence-transformers/all-MiniLM-L6-v2"
+    top_k: 5
+    similarity_threshold: 0.7
 
 behavior:
   reaction_time_min_ms: 200

src/config/settings.py

@@ -51,24 +51,30 @@ class AIConfig(BaseModel):
     llm: LLMProviderConfig = Field(default_factory=LLMProviderConfig)
 
 
-class MapperConfig(BaseModel):
-    """Mapper configuration"""
-    class ActiveStorageConfig(BaseModel):
-        db_path: str = "data/maps/active/world.db"
-        graph_cache_size: int = 10000
-    
-    class HistoricalStorageConfig(BaseModel):
-        archive_dir: str = "data/maps/archive"
-        
-        class SnapshotTriggers(BaseModel):
-            room_count: int = 500
-            session_end: bool = True
-            time_interval_hours: int = 24
-        
-        snapshot_triggers: SnapshotTriggers = Field(default_factory=SnapshotTriggers)
-    
-    active_storage: ActiveStorageConfig = Field(default_factory=ActiveStorageConfig)
-    historical_storage: HistoricalStorageConfig = Field(default_factory=HistoricalStorageConfig)
+class MapperConfig(BaseModel):
+    """Mapper configuration"""
+    class ActiveStorageConfig(BaseModel):
+        db_path: str = "data/maps/active/world.db"
+        graph_cache_size: int = 10000
+    
+    class HistoricalStorageConfig(BaseModel):
+        archive_dir: str = "data/maps/archive"
+        
+        class SnapshotTriggers(BaseModel):
+            room_count: int = 500
+            session_end: bool = True
+            time_interval_hours: int = 24
+        
+        snapshot_triggers: SnapshotTriggers = Field(default_factory=SnapshotTriggers)
+    
+    class SemanticSearchConfig(BaseModel):
+        model: str = "sentence-transformers/all-MiniLM-L6-v2"
+        top_k: int = 5
+        similarity_threshold: float = 0.7
+    
+    active_storage: ActiveStorageConfig = Field(default_factory=ActiveStorageConfig)
+    historical_storage: HistoricalStorageConfig = Field(default_factory=HistoricalStorageConfig)
+    semantic_search: SemanticSearchConfig = Field(default_factory=SemanticSearchConfig)
 
 
 class BehaviorConfig(BaseModel):

src/main.py

@@ -60,6 +60,7 @@ async def main_loop(config):
         enable_sync=True,
         room_count_trigger=config.mapper.historical_storage.snapshot_triggers.room_count,
         time_interval_hours=config.mapper.historical_storage.snapshot_triggers.time_interval_hours,
+        embedding_model=config.mapper.semantic_search.model,
     )
 
     # Initialize Character Creator (Phase 2)

src/modules/mapper/world_mapper.py

@@ -39,6 +39,7 @@ class WorldMapper:
         enable_sync: bool = True,
         room_count_trigger: int = 500,
         time_interval_hours: float = 24.0,
+        embedding_model: str = "sentence-transformers/all-MiniLM-L6-v2",
     ) -> None:
         self.storage = ActiveStorage(db_path=db_path)
         self.pathfinder = Pathfinder(self.storage.graph)
@@ -53,7 +54,7 @@ class WorldMapper:
         self._sync_manager: Optional[SyncManager] = None
         if enable_sync:
             historical = HistoricalStorage(archive_dir=archive_dir)
-            semantic = SemanticSearch()
+            semantic = SemanticSearch(model_name=embedding_model)
             self._sync_manager = SyncManager(
                 active=self.storage,
                 historical=historical,
@@ -62,8 +63,8 @@ class WorldMapper:
                 time_interval_hours=time_interval_hours,
             )
             logger.info(
-                "SyncManager enabled: trigger=%d rooms, interval=%.1fh",
-                room_count_trigger, time_interval_hours,
+                "SyncManager enabled: trigger=%d rooms, interval=%.1fh, model=%s",
+                room_count_trigger, time_interval_hours, embedding_model,
             )
 
     # ------------------------------------------------------------------