Update app.py

app.py

@@ -56,9 +56,13 @@ from sklearn.metrics.pairwise import cosine_similarity
 from concurrent.futures import ThreadPoolExecutor
 
 # --- Setup Logging ---
-logging.basicConfig(level=logging.INFO,
-                    format='[%(asctime)s] [%(levelname)s] [%(threadName)s] %(message)s',
-                    datefmt='%Y-%m-%d %H:%M:%S')
+logging.basicConfig(
+    level=logging.INFO,
+    format='[%(asctime)s] [%(levelname)s] [%(threadName)s] %(message)s',
+    datefmt='%Y-%m-%d %H:%M:%S',
+    stream=sys.stdout,
+    force=True
+)
 
 try:
     import pvporcupine
@@ -1298,7 +1302,7 @@ class Hive:
         # This can be manually overridden by setting HIVE_USE_HF_INFERENCE.
         is_hf_space = "SPACE_ID" in os.environ
         use_remote_default = is_hf_space
-
+        print(f"[Hive] Detected Hugging Face Space: {is_hf_space}. Defaulting to remote inference: {use_remote_default}.")
         # Check for manual override from environment variable
         if "HIVE_USE_HF_INFERENCE" in os.environ:
             use_remote = CFG["HIVE_USE_HF_INFERENCE"]
@@ -1306,7 +1310,7 @@ class Hive:
             use_remote = use_remote_default
 
         if use_remote:
-            print("[Hive] Using remote Hugging Face Inference endpoint.")
+            print("[Hive] Using remote Hugging Face Inference endpoint.", flush=True)
             from huggingface_hub import InferenceClient; endpoint = CFG["HIVE_HF_ENDPOINT"] or None; token = CFG["HF_READ_TOKEN"] or os.getenv("HF_TOKEN") or os.getenv("HUGGING_FACE_HUB_TOKEN") or None
             self.client = InferenceClient(model=self.model_id if endpoint is None else None, token=token, timeout=60, base_url=endpoint) # type: ignore
             def _remote_pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, **kw):
@@ -1323,7 +1327,7 @@ class Hive:
             self.model = None # No local model needed
             self.stopping_criteria = None # Not used with InferenceClient
         else:
-            print("[Hive] Using local LLM for inference.")
+            print("[Hive] Using local LLM for inference.", flush=True)
             self.tok = AutoTokenizer.from_pretrained(self.model_id, trust_remote_code=trust, chat_template=None)
             if self.tok.pad_token is None:
                 self.tok.pad_token = self.tok.eos_token
@@ -1454,7 +1458,7 @@ class Hive:
         if hasattr(self, 'client') and self.client: # Remote Inference
             stop_sequences = ["</s>", "Assistant:"] + [self.tok.decode(st) for st in self.stop_tokens]
             try:
-                for token in self.client.text_generation(prompt, max_new_tokens=int(max_new_tokens), temperature=float(temperature), do_sample=True, stop_sequences=stop_sequences, stream=True):
+                for token in self.client.text_generation(prompt, max_new_tokens=int(max_new_tokens), temperature=float(temperature), do_sample=True, stop_sequences=stop_sequences, stream=True): # type: ignore
                     yield token
             except Exception as e:
                 print(f"[ModelBridge] Remote inference stream failed: {e}")
@@ -1671,7 +1675,7 @@ def launch_ui(bootstrap_instance: "Bootstrap"):
                 
                 current_history.append({"role": "assistant", "content": ""})
                 try:
-                    # The dialogue manager needs the full history to maintain context.
+                    # The dialogue manager needs the full history to maintain context. # type: ignore
                     for chunk in hive_instance.dialogue_manager.process_turn(current_history, current_user_id, effective_role, session_id):
                         if chunk["type"] == "token":
                             current_history[-1]["content"] += chunk["content"]
@@ -1679,7 +1683,7 @@ def launch_ui(bootstrap_instance: "Bootstrap"):
                     # After the stream is complete, re-enable the textbox.
                     yield current_history, gr.Textbox(placeholder=f"Talk to {CFG['AGENT_NAME']}", interactive=True)
                 except Exception as e:
-                    error_msg = f"Error in DialogueManager: {e}"
+                    error_msg = f"Error in DialogueManager: {e}" # type: ignore
                     print(f"[ERROR] {error_msg}")
                     current_history[-1]["content"] = f"An error occurred: {error_msg}"
                     yield current_history, gr.Textbox(value="", interactive=True)
@@ -1704,7 +1708,7 @@ def launch_ui(bootstrap_instance: "Bootstrap"):
 
         def do_get_vocab_word(uid, request: gr.Request):
             hive_instance = get_hive_instance() # type: ignore
-            if hive_instance.lite_mode: return "Vocabulary features are disabled in Lite Mode."
+            if hive_instance.lite_mode: return "Vocabulary features are disabled in Lite Mode." # type: ignore
             current_user_id = uid or request.session_hash
             log_path = os.path.join(CFG["HIVE_HOME"], "users", "conversations", f"{current_user_id}.jsonl")
             if not os.path.exists(log_path): return "No conversation history to find words from."
@@ -1720,7 +1724,7 @@ def launch_ui(bootstrap_instance: "Bootstrap"):
 
         def wait_for_memory_features():
             """Waits for the full Hive core and enables memory-related UI features."""
-            bootstrap_instance.hive_ready.wait()
+            bootstrap_instance.hive_ready.wait() # type: ignore
             hive_instance = get_hive_instance() # Ensure the UI's HIVE_INSTANCE is updated to full
             return (
                 "✅ **Full Hive Core is Ready.** Advanced features are now online.",
@@ -1742,7 +1746,7 @@ def launch_ui(bootstrap_instance: "Bootstrap"):
         demo.load(wait_for_memory_features, None, [core_status, summary_output, msg, summary_btn, vocab_output, vocab_btn, progress_output, online_now, ingest_now_btn, mem_compress_btn, hotpatch_apply, propose_btn, test_btn, apply_btn, network_status_md])
         def wait_for_lite_core():
             """Waits for the lite Hive core and enables basic chat."""
-            bootstrap_instance.lite_core_ready.wait()
+            bootstrap_instance.lite_core_ready.wait() # type: ignore
             return gr.Textbox(placeholder=f"Talk to {CFG['AGENT_NAME']} (Lite Mode)", interactive=True)
 
         demo.load(wait_for_lite_core, None, [msg])
@@ -1752,12 +1756,12 @@ def launch_ui(bootstrap_instance: "Bootstrap"):
             global HIVE_INSTANCE
 
             # If the full hive is ready, ensure we are using it, and it's a valid instance.
-            if bootstrap_instance.hive_ready.is_set():
+            if bootstrap_instance.hive_ready.is_set(): # type: ignore
                 if bootstrap_instance.hive_instance is not None and (HIVE_INSTANCE is None or HIVE_INSTANCE.lite_mode):
                     HIVE_INSTANCE = bootstrap_instance.hive_instance
                     print("[UI] Full Hive instance attached.")
                 return HIVE_INSTANCE
-            
+            # type: ignore
             # Otherwise, use the lite instance.
             if HIVE_INSTANCE is None:
                 if bootstrap_instance.lite_core_ready.is_set() and bootstrap_instance.hive_lite_instance is not None:
@@ -1772,8 +1776,8 @@ def launch_ui(bootstrap_instance: "Bootstrap"):
 
         def wait_for_voice_features(request: gr.Request):
             """Waits for ASR/TTS models and enables voice-related UI elements."""
-            bootstrap_instance.voice_ready.wait()
-            bootstrap_instance.hive_ready.wait() # Also wait for full core for voice features
+            bootstrap_instance.voice_ready.wait() # type: ignore
+            bootstrap_instance.hive_ready.wait() # Also wait for full core for voice features # type: ignore
             hive_instance = get_hive_instance(bootstrap_instance)
 
             voice_ready = not hive_instance.lite_mode and hasattr(hive_instance, 'asr_service') and hasattr(hive_instance, 'tts_service')