Working local Ollama chatbot with flexible history format and conversation memory

app.py

@@ -1,4 +1,4 @@
-# app.py - Local Ollama Chatbot with DeepSeek
+# app.py - Llama 3.2 Chatbot with Ollama
 import gradio as gr
 import requests
 import logging
@@ -7,170 +7,141 @@ import traceback
 from datetime import datetime
 
 # Logging
-logging.basicConfig(level=logging.INFO, format="%(asctime)s | %(message)s", handlers=[logging.StreamHandler(sys.stdout)])
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s | %(levelname)-8s | %(message)s",
+    handlers=[logging.StreamHandler(sys.stdout)]
+)
 logger = logging.getLogger("CHATBOT")
+logging.getLogger("httpx").setLevel(logging.WARNING)
+logging.getLogger("httpcore").setLevel(logging.WARNING)
 
 logger.info("=" * 60)
-logger.info("  DEEPSEEK CHATBOT - LOCAL OLLAMA")
+logger.info("  LLAMA 3.2 CHATBOT")
 logger.info("=" * 60)
 
-# Config
 OLLAMA_URL = "http://localhost:11434"
-MODEL = "deepseek-coder:6.7b-instruct-q6_K"
+MODEL = "llama3.2:3b"
 
-# Check connection
-ollama_connected = False
+# Check Ollama
+logger.info(f"Checking Ollama...")
+ollama_ok = False
 try:
-    resp = requests.get(f"{OLLAMA_URL}/api/tags", timeout=3)
-    if resp.status_code == 200:
-        models = [m["name"] for m in resp.json().get("models", [])]
+    r = requests.get(f"{OLLAMA_URL}/api/tags", timeout=5)
+    if r.status_code == 200:
+        models = [m["name"] for m in r.json().get("models", [])]
+        logger.info(f"Models: {models}")
         if MODEL in models:
-            ollama_connected = True
-            logger.info(f"Ollama connected! Using {MODEL}")
-        else:
-            logger.warning(f"Model {MODEL} not found. Available: {models}")
-except:
-    logger.warning("Ollama not running")
-
-# Install instructions
-INSTALL_INSTRUCTIONS = """
-# 🚀 Local AI Chatbot Setup
-
-This chatbot runs on **YOUR computer** using Ollama + DeepSeek AI.
-
-## Step 1: Install Ollama
-
-**Mac:**
-```bash
-brew install ollama
-```
-
-**Linux:**
-```bash
-curl -fsSL https://ollama.com/install.sh | sh
-```
-
-**Windows:** Download from [ollama.com/download](https://ollama.com/download)
-
----
-
-## Step 2: Start Ollama & Download Model
-
-```bash
-# Start Ollama (run in terminal, keep open)
-ollama serve
-
-# Download DeepSeek model (6.7GB, one-time)
-ollama pull deepseek-coder:6.7b-instruct-q6_K
-```
-
----
-
-## Step 3: Refresh This Page
-
-Once Ollama is running with the model, **refresh this page** and the chat will work!
-
----
-
-### Alternative: Clone & Run Locally
-
-```bash
-git clone https://huggingface.co/spaces/ndwdgda/cpu
-cd cpu
-pip install gradio requests
-python app.py
-```
-
-Then open http://127.0.0.1:7860
-"""
+            ollama_ok = True
+            logger.info(f"CONNECTED - {MODEL}")
+except Exception as e:
+    logger.error(f"Connection failed: {e}")
 
-# Chat function
+# Chat function - FLEXIBLE history handling
 request_count = 0
 
-def chat_with_ai(message, history):
+def chat_fn(message, history):
     global request_count
     request_count += 1
-    rid = f"REQ-{request_count:04d}"
+    rid = f"R{request_count:04d}"
     
-    logger.info(f"[{rid}] User: {message}")
+    logger.info(f"[{rid}] User: {message[:60]}...")
+    logger.info(f"[{rid}] History: {len(history)} items")
+    logger.info(f"[{rid}] History type: {type(history)}")
     
-    if not ollama_connected:
-        return "❌ Ollama not connected! See the Setup tab for instructions."
+    if not ollama_ok:
+        return "Ollama not connected. Install Ollama and run: ollama serve"
+    
+    # Build prompt - HANDLE ANY HISTORY FORMAT
+    prompt = "You are a helpful AI assistant.\n\n"
+    
+    try:
+        for i, item in enumerate(history):
+            logger.debug(f"[{rid}] History[{i}]: type={type(item)}, value={item}")
+            
+            if isinstance(item, dict):
+                # Gradio 6.x format: {"role": "user", "content": "..."}
+                role = item.get("role", "")
+                content = str(item.get("content", ""))
+                if role == "user":
+                    prompt += f"User: {content}\n"
+                elif role == "assistant":
+                    prompt += f"Assistant: {content}\n"
+                    
+            elif isinstance(item, (list, tuple)):
+                # Old format: (user_msg, bot_msg) or [user_msg, bot_msg]
+                if len(item) == 2:
+                    user_msg, bot_msg = item[0], item[1]
+                    prompt += f"User: {user_msg}\n"
+                    if bot_msg:
+                        prompt += f"Assistant: {bot_msg}\n"
+                elif len(item) > 2:
+                    # Handle unexpected tuple/list length
+                    logger.warning(f"[{rid}] Unexpected item length: {len(item)}")
+                    for j, sub in enumerate(item):
+                        if isinstance(sub, dict):
+                            role = sub.get("role", "")
+                            content = str(sub.get("content", ""))
+                            if role == "user":
+                                prompt += f"User: {content}\n"
+                            elif role == "assistant":
+                                prompt += f"Assistant: {content}\n"
+            else:
+                logger.warning(f"[{rid}] Unknown history item type: {type(item)}")
+                
+    except Exception as e:
+        logger.error(f"[{rid}] History parsing error: {e}")
+        logger.error(traceback.format_exc())
+    
+    prompt += f"User: {message}\nAssistant:"
+    
+    logger.info(f"[{rid}] Prompt: {len(prompt)} chars")
     
     try:
-        prompt = "You are a helpful AI coding assistant.\n\n"
-        if history:
-            for item in history:
-                if isinstance(item, dict):
-                    r = item.get("role", "")
-                    c = str(item.get("content", ""))
-                    if r == "user":
-                        prompt += f"User: {c}\n"
-                    elif r == "assistant":
-                        prompt += f"Assistant: {c}\n"
-                elif isinstance(item, (list, tuple)) and len(item) >= 2:
-                    prompt += f"User: {item[0]}\n"
-                    if item[1]:
-                        prompt += f"Assistant: {item[1]}\n"
-        
-        prompt += f"User: {message}\nAssistant:"
-        
-        logger.info(f"[{rid}] Calling DeepSeek...")
         start = datetime.now()
-        
-        resp = requests.post(
+        r = requests.post(
             f"{OLLAMA_URL}/api/generate",
             json={"model": MODEL, "prompt": prompt, "stream": False},
             timeout=120
         )
-        
         elapsed = (datetime.now() - start).total_seconds()
         
-        if resp.status_code == 200:
-            result = resp.json()
-            text = result.get("response", "")
-            logger.info(f"[{rid}] Response in {elapsed:.1f}s: {text[:100]}...")
-            return text.strip()
+        logger.info(f"[{rid}] Status: {r.status_code}, Time: {elapsed:.1f}s")
+        
+        if r.status_code == 200:
+            text = r.json().get("response", "").strip()
+            logger.info(f"[{rid}] Response: {text[:80]}...")
+            return text
         else:
-            return f"Error: {resp.status_code} - {resp.text[:200]}"
-            
+            return f"Error: Ollama returned {r.status_code}"
+    except requests.exceptions.Timeout:
+        logger.error(f"[{rid}] Timeout!")
+        return "Error: Request timed out (120s)"
     except Exception as e:
-        logger.error(f"[{rid}] Error: {e}")
+        logger.error(f"[{rid}] Error: {e}\n{traceback.format_exc()}")
         return f"Error: {e}"
 
 # UI
 logger.info("Building UI...")
 
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# 🤖 DeepSeek AI Chatbot")
+if ollama_ok:
+    description = f"**Connected** to Ollama | Model: `{MODEL}`"
+else:
+    description = """**Not connected** - Install Ollama first:
     
-    if ollama_connected:
-        gr.Markdown(f"**✅ Connected to Ollama** | Model: `{MODEL}`")
-    else:
-        gr.Markdown("**❌ Ollama not connected** - See **Setup** tab below for instructions!")
-    
-    with gr.Tabs():
-        with gr.TabItem("💬 Chat"):
-            chatbot = gr.Chatbot(height=400)
-            msg = gr.Textbox(placeholder="Ask me anything..." if ollama_connected else "Install Ollama first (see Setup tab)", label="Message")
-            with gr.Row():
-                send = gr.Button("Send", variant="primary")
-                clear = gr.ClearButton([msg, chatbot])
-            
-            def respond(message, history):
-                if not message.strip():
-                    return "", history
-                bot_response = chat_with_ai(message, history)
-                history = history + [[message, bot_response]]
-                return "", history
-            
-            msg.submit(respond, [msg, chatbot], [msg, chatbot])
-            send.click(respond, [msg, chatbot], [msg, chatbot])
-        
-        with gr.TabItem("📋 Setup"):
-            gr.Markdown(INSTALL_INSTRUCTIONS)
+**Mac:** `brew install ollama`
+**Linux:** `curl -fsSL https://ollama.com/install.sh | sh`
+**Then run:** `ollama serve` and `ollama pull llama3.2:3b`
+"""
+
+demo = gr.ChatInterface(
+    fn=chat_fn,
+    title="Llama 3.2 Chatbot",
+    description=description,
+    examples=["Hello!", "What is Python?", "Tell me a joke"],
+)
 
-logger.info("READY!")
+logger.info("READY! Open http://127.0.0.1:7860")
 
 if __name__ == "__main__":
     demo.launch()