Force CPU for embeddings and add ingest logging

apps/web/app.py

@@ -119,6 +119,9 @@ def admin_ingest(files):
         init_services()
         status += "Services reloaded. Ready to chat."
     except Exception as e:
+        print(f"Ingestion Failed: {e}") # Print to server logs
+        import traceback
+        traceback.print_exc()
         status += f"Error: {e}"
         
     return status

services/rag/embed.py

@@ -16,7 +16,8 @@ class Embedder:
             self.client = OpenAI(api_key=api_key)
         else:
             print(f"Loading local embedding model: {model_name}")
-            self.model = SentenceTransformer(model_name)
+            # Force CPU to avoid ZeroGPU conflicts during ingestion
+            self.model = SentenceTransformer(model_name, device="cpu")
 
     def embed(self, texts: List[str]) -> np.ndarray:
         if not texts: