feat: chunking retrieval updated

requirements.txt

@@ -29,7 +29,7 @@ idna==3.11
 Mako==1.3.10
 MarkupSafe==3.0.3
 mpmath==1.3.0
-numpy==2.3.5
+numpy==2.2.6
 onnxruntime==1.23.2
 packaging==25.0
 passlib==1.7.4

src/chatbot/embedding.py

@@ -9,10 +9,10 @@ MODEL_ID = "onnx-community/embeddinggemma-300m-ONNX"
 
 class EmbeddingModel:
     def __init__(self):
-        print("🔵 Loading tokenizer…")
+        print("Loading tokenizer…")
         self.tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
 
-        print("🔵 Downloading ONNX model files…")
+        print("Downloading ONNX model files…")
 
         self.model_path = hf_hub_download(
             repo_id=MODEL_ID,
@@ -25,7 +25,7 @@ class EmbeddingModel:
 
         model_dir = os.path.dirname(self.model_path)
 
-        print("🔵 Creating inference session…")
+        print("Creating inference session…")
         self.session = ort.InferenceSession(
             self.model_path,
             providers=["CPUExecutionProvider"],

src/chatbot/schemas.py

@@ -37,4 +37,4 @@ class SemanticSearchResult(BaseModel):
 
 class ManualTextRequest(BaseModel):
     kb_id: uuid.UUID
-    text: str
+    text: str

src/chatbot/service.py

@@ -68,4 +68,4 @@ async def store_manual_text(kb_id: UUID, text: str, session: AsyncSession):
         "chunk_index": next_index,
         "status": "stored",
         "text": text
-    }
+    }

src/main.py

@@ -19,10 +19,6 @@ app.include_router(auth_router)
 
 app.include_router(profile)
 
-# app.include_router(assets)
-
-# app.include_router(leave)
-
 app.include_router(chatbot)