Update app.py

app.py

@@ -11,13 +11,17 @@ MODEL_ID = "drwlf/PsychoQwen14b"
 HF_TOKEN = os.getenv("HF_TOKEN")
 
 # Initialize client, handle potential missing token
-try:
-    client = InferenceClient(model=MODEL_ID, token=HF_TOKEN)
-    print("InferenceClient initialized successfully.")
-except Exception as e:
-    print(f"Error initializing InferenceClient: {e}")
-    print("Please ensure HF_TOKEN is set in your environment/secrets.")
-    client = None # Set client to None if initialization fails
+client = None # Initialize client to None
+if not HF_TOKEN:
+    print("Warning: HF_TOKEN secret not found. Cannot initialize InferenceClient.")
+    # Optionally raise an error or handle this case in the respond function
+else:
+    try:
+        client = InferenceClient(model=MODEL_ID, token=HF_TOKEN)
+        print("InferenceClient initialized successfully.")
+    except Exception as e:
+        print(f"Error initializing InferenceClient: {e}")
+        # Client remains None
 
 
 def respond(
@@ -33,7 +37,10 @@ def respond(
     Generator function to stream responses from the HF Inference API.
     """
     if not client:
-         yield "Error: Inference Client not initialized. Check HF_TOKEN."
+         yield "Error: Inference Client not initialized. Check HF_TOKEN secret."
+         return
+    if not message or not message.strip():
+         yield "Please enter a message."
          return
 
     messages = [{"role": "system", "content": system_message}]
@@ -52,6 +59,14 @@ def respond(
     # Handle Top-K value (API often expects None to disable, not 0)
     top_k_val = top_k if top_k > 0 else None
 
+    # Debugging: Print parameters being sent
+    print(f"--- Sending Request ---")
+    print(f"Model: {MODEL_ID}")
+    print(f"Messages: {messages}")
+    print(f"Max Tokens: {max_tokens}, Temp: {temperature}, Top-P: {top_p}, Top-K: {top_k_val}")
+    print(f"-----------------------")
+
+
     try:
         stream = client.chat_completion(
             messages,
@@ -73,16 +88,50 @@ def respond(
                 token = message_chunk.choices[0].delta.content
                 if token: # Ensure token is not None or empty
                     response += token
+                    # print(token, end="") # Debugging stream locally
                     yield response
             # Optional: Add error checking within the loop if needed
 
     except Exception as e:
          print(f"Error during chat completion: {e}")
          yield f"Sorry, an error occurred: {str(e)}"
-    finally:
-        # Ensure the stream object is properly handled if it exists
-        # (Though InferenceClient might handle cleanup internally)
-        if stream is not None:
-            # Potential cleanup if required by the library, often not needed explicitly
-            pass
+    # No finally block needed unless specific cleanup is required
+
+
+"""
+For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
+"""
+demo = gr.ChatInterface(
+    respond,
+    chatbot=gr.Chatbot(height=500), # Set chatbot height
+    additional_inputs=[
+        gr.Textbox(value="You are a friendly psychotherapy AI capable of thinking.", label="System message"),
+        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=0.1, maximum=2.0, value=0.7, step=0.1, label="Temperature"), # Adjusted max temp based on common usage
+        gr.Slider(
+            minimum=0.05, # Min Top-P often > 0
+            maximum=1.0,
+            value=0.95,
+            step=0.05,
+            label="Top-P (nucleus sampling)",
+        ),
+        # Added Top-K slider
+        gr.Slider(
+            minimum=0, # 0 disables Top-K
+            maximum=100, # Common range, adjust if needed
+            value=0, # Default to disabled
+            step=1,
+            label="Top-K (0 = disabled)",
+        ),
+    ],
+     title="PsychoQwen Chat",
+     description=f"Chat with {MODEL_ID}. Adjust generation parameters below.",
+     retry_btn="Retry",
+     undo_btn="Undo",
+     clear_btn="Clear Chat",
+)
+
+# --- Launch the app directly ---
+# The if __name__ == "__main__": block is removed or commented out
+demo.queue().launch(debug=True) # debug=True is useful for seeing logs in the Space