change

app.py

@@ -2,9 +2,12 @@ import gradio as gr
 import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM
 from fastapi import FastAPI, Request
+from fastapi.middleware.cors import CORSMiddleware
+from gradio.routes import mount_gradio_app
+import uvicorn
 
 # -------------------------------------------------
-# 1. Load model (same as your old code)
+# 1. Load model
 # -------------------------------------------------
 print("Initializing DialoGPT-medium model...")
 model_name = "microsoft/DialoGPT-medium"
@@ -14,25 +17,23 @@ model = AutoModelForCausalLM.from_pretrained(model_name)
 if tokenizer.pad_token is None:
     tokenizer.pad_token = tokenizer.eos_token
 
-print("DialoGPT-medium loaded!")
+print("DialoGPT-medium loaded successfully!")
 
 # -------------------------------------------------
-# 2. Generation helper (your old logic, cleaned up)
+# 2. Helper: Generate a response
 # -------------------------------------------------
 def generate_response(message: str, chat_history: list):
     if not message.strip():
         return "Please enter a message."
 
-    # Build conversation string
-    conv = ""
+    # Build the conversation context
+    conversation = ""
     for user, bot in chat_history:
-        conv += f"User: {user}\nBot: {bot}\n"
-    conv += f"User: {message}\nBot:"
+        conversation += f"User: {user}\nBot: {bot}\n"
+    conversation += f"User: {message}\nBot:"
 
-    # Encode
-    inputs = tokenizer.encode(conv, return_tensors="pt", max_length=1024, truncation=True)
+    inputs = tokenizer.encode(conversation, return_tensors="pt", max_length=1024, truncation=True)
 
-    # Generate
     with torch.no_grad():
         outputs = model.generate(
             inputs,
@@ -53,20 +54,20 @@ def generate_response(message: str, chat_history: list):
     return response
 
 # -------------------------------------------------
-# 3. Gradio chat function (used by /run/predict)
+# 3. Gradio chat handler
 # -------------------------------------------------
 def chat_fn(message: str, history: list):
     response = generate_response(message, history or [])
     history.append((message, response))
-    return "", history   # clear textbox, update chat
+    return "", history  # clear textbox, update chat
 
 # -------------------------------------------------
-# 4. Build the UI (your Blocks layout)
+# 4. Build the Gradio UI
 # -------------------------------------------------
 example_questions = [
     "Hello! How are you today?",
     "What can you help me with?",
-    "Tell me about artificial intelligence",
+    "Tell me about artificial intelligence.",
     "What's your favorite programming language?",
     "Can you explain machine learning?",
     "How does a neural network work?"
@@ -77,8 +78,8 @@ with gr.Blocks(
     title="GihonTech - AI Conversation Assistant"
 ) as demo:
 
-    gr.Markdown("# GihonTech AI Conversation Assistant")
-    gr.Markdown("Chat with an AI powered by **DialoGPT-medium**")
+    gr.Markdown("# 🤖 GihonTech AI Conversation Assistant")
+    gr.Markdown("Chat with an AI powered by **DialoGPT-medium**.")
 
     with gr.Row():
         with gr.Column(scale=3):
@@ -104,7 +105,7 @@ with gr.Blocks(
             gr.Markdown("---")
             gr.Markdown("### Model Info")
             gr.Textbox(
-                value="DialoGPT-medium: Loaded",
+                value="DialoGPT-medium: Loaded ✅",
                 label="Model Status",
                 interactive=False,
             )
@@ -115,39 +116,41 @@ with gr.Blocks(
                 - Conversation memory  
 
                 **Tips**  
-                - Ask clear questions  
-                - Use *Clear Chat* to start over
+                - Ask clear, simple questions  
+                - Use *Clear Chat* to start over  
                 """
             )
 
-    # Event wiring
+    # Wire up events
     send.click(chat_fn, inputs=[msg, chatbot], outputs=[msg, chatbot])
     msg.submit(chat_fn, inputs=[msg, chatbot], outputs=[msg, chatbot])
     clear.click(lambda: ([], ""), outputs=[chatbot, msg])
 
 # -------------------------------------------------
-# 5. OPTIONAL: expose /lambda (same JSON format)
+# 5. FastAPI app + Lambda route
 # -------------------------------------------------
 fastapi_app = FastAPI()
 
+# Allow AnythingLLM / frontend CORS access
+fastapi_app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
 @fastapi_app.post("/lambda")
 async def lambda_endpoint(req: Request):
     payload = await req.json()
-    # Gradio sends {"data": [...]} ; we accept anything
     user_msg = payload.get("data", [""])[0]
-    # Use the same generation logic (no history for this endpoint)
-    resp = generate_response(user_msg, [])
-    return {"data": [resp]}
+    response = generate_response(user_msg, [])
+    return {"data": [response]}
 
-demo.mount_app(fastapi_app)   # makes /lambda reachable
+# Mount Gradio app inside FastAPI
+mount_gradio_app(fastapi_app, demo, path="/")
 
 # -------------------------------------------------
-# 6. Launch with queue (critical for API!)
+# 6. Run the combined FastAPI + Gradio app
 # -------------------------------------------------
 if __name__ == "__main__":
-    demo.queue().launch(
-        server_name="0.0.0.0",
-        server_port=7860,
-        share=False,
-        show_error=True,
-    )
+    uvicorn.run(fastapi_app, host="0.0.0.0", port=7860)