Ollama-only chatbot - working locally with llama3.2:3b

app.py

@@ -1,135 +1,110 @@
-# app.py - Main Gradio Application with Ollama Backend
+# app.py - Ollama-Only Chatbot
 import gradio as gr
-import os
-from logging_config import setup_logging, log_banner, log_section, log_startup_info
-from ollama_client import OllamaClient
+import requests
+import logging
+import sys
 
 # ============================================================================
-#                    INITIALIZATION
+#                    LOGGING SETUP
 # ============================================================================
-logger = setup_logging()
-log_startup_info(logger)
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s | %(levelname)-8s | %(message)s",
+    handlers=[logging.StreamHandler(sys.stdout)]
+)
+logger = logging.getLogger("ChatbotBrain")
+
+logger.info("=" * 60)
+logger.info("  OLLAMA CHATBOT STARTING")
+logger.info("=" * 60)
 
 # ============================================================================
 #                    OLLAMA CONFIGURATION
 # ============================================================================
-# Models to try in order of preference (smaller = faster, more reliable)
-MODELS = [
-    "llama3.2:3b",      # Fast, small
-    "gemma3:1b",        # Very fast, tiny
-    "phi3:mini",        # Good quality, medium
-    "deepseek-coder:6.7b-instruct-q6_K",  # Good for code
-]
-
-log_section(logger, "OLLAMA CLIENT SETUP")
-ollama = OllamaClient(logger, model=MODELS[0])
+OLLAMA_URL = "http://localhost:11434"
+MODEL = "llama3.2:3b"
 
-# Check connection and find working model
-logger.info("Checking Ollama connection...")
-if ollama.check_connection():
-    logger.info("Ollama is running!")
-    available = ollama.list_models()
-    
-    # Find first available preferred model
-    for model in MODELS:
-        if model in available:
-            ollama.model = model
-            logger.info(f"Selected model: {model}")
-            break
-else:
-    logger.warning("Ollama not available - running in limited mode")
+# Check Ollama connection
+logger.info(f"Checking Ollama at {OLLAMA_URL}...")
+try:
+    response = requests.get(f"{OLLAMA_URL}/api/tags", timeout=5)
+    if response.status_code == 200:
+        models = [m["name"] for m in response.json().get("models", [])]
+        logger.info(f"Ollama connected! Found {len(models)} models")
+        if MODEL in models:
+            logger.info(f"Using model: {MODEL}")
+        else:
+            logger.warning(f"Model {MODEL} not found, available: {models[:5]}")
+    else:
+        logger.error(f"Ollama returned {response.status_code}")
+except Exception as e:
+    logger.error(f"Cannot connect to Ollama: {e}")
+    logger.error("Run: ollama serve")
 
 # ============================================================================
-#                    CHAT RESPONSE FUNCTION
+#                    CHAT FUNCTION
 # ============================================================================
-def respond(message, history):
-    """Generate AI response using Ollama."""
-    log_section(logger, "NEW USER MESSAGE")
-    logger.info(f"User: {message}")
-    logger.info(f"History: {len(history)} previous messages")
+def chat_with_ollama(message, history):
+    """Send message to Ollama and get response."""
+    logger.info("-" * 40)
+    logger.info(f"USER: {message}")
+    logger.info(f"History: {len(history)} messages")
     
-    # Build messages array
-    messages = [{"role": "system", "content": "You are a helpful AI assistant. Be concise and helpful."}]
-    for user_msg, bot_msg in history[-5:]:  # Last 5 exchanges for context
-        messages.append({"role": "user", "content": user_msg})
+    # Build prompt from history (tuple format: [(user, bot), ...])
+    prompt = "You are a helpful AI assistant.\n\n"
+    for user_msg, bot_msg in history:
+        prompt += f"User: {user_msg}\n"
         if bot_msg:
-            messages.append({"role": "assistant", "content": bot_msg})
-    messages.append({"role": "user", "content": message})
+            prompt += f"Assistant: {bot_msg}\n"
+    prompt += f"User: {message}\nAssistant:"
     
-    # Call Ollama
-    result = ollama.chat(messages)
+    logger.info(f"Sending to Ollama ({MODEL})...")
     
-    if result["success"]:
-        return result["response"].strip()
-    else:
-        error_msg = result.get("error", "Unknown error")
-        return f"Error: {error_msg}\n\nMake sure Ollama is running: `ollama serve`"
+    try:
+        response = requests.post(
+            f"{OLLAMA_URL}/api/generate",
+            json={
+                "model": MODEL,
+                "prompt": prompt,
+                "stream": False
+            },
+            timeout=120
+        )
+        
+        if response.status_code == 200:
+            result = response.json()
+            text = result.get("response", "")
+            duration = result.get("total_duration", 0) / 1_000_000_000
+            logger.info(f"SUCCESS: Got response in {duration:.1f}s")
+            logger.info(f"AI: {text[:100]}...")
+            return text.strip()
+        else:
+            error = f"Ollama returned {response.status_code}: {response.text}"
+            logger.error(error)
+            return f"Error: {error}"
+            
+    except requests.exceptions.ConnectionError:
+        logger.error("Cannot connect to Ollama. Is it running?")
+        return "Error: Cannot connect to Ollama. Run: ollama serve"
+    except Exception as e:
+        logger.error(f"Error: {e}")
+        return f"Error: {e}"
 
 # ============================================================================
 #                    GRADIO UI
 # ============================================================================
-log_section(logger, "BUILDING GRADIO UI")
+logger.info("Building Gradio UI...")
 
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# CPU Chatbot")
-    gr.Markdown(f"### Powered by Ollama ({ollama.model})")
-    gr.Markdown("*Using local AI - no cloud required!*")
-    
-    with gr.Row():
-        with gr.Column(scale=4):
-            chatbot = gr.Chatbot(height=500, label="Chat")
-            msg = gr.Textbox(placeholder="Ask me anything...", label="Your message")
-            with gr.Row():
-                submit_btn = gr.Button("Send", variant="primary")
-                clear = gr.ClearButton([msg, chatbot], value="Clear")
-        
-        with gr.Column(scale=1):
-            gr.Markdown("### Status")
-            status_box = gr.JSON(
-                label="Ollama Stats",
-                value=ollama.get_stats()
-            )
-            model_dropdown = gr.Dropdown(
-                choices=ollama.available_models or MODELS,
-                value=ollama.model,
-                label="Model"
-            )
-            refresh_btn = gr.Button("Refresh")
-    
-    def user_submit(message, history):
-        if not message.strip():
-            return "", history, ollama.get_stats()
-        return "", history + [[message, None]], ollama.get_stats()
-    
-    def bot_respond(history):
-        if not history:
-            return history, ollama.get_stats()
-        user_message = history[-1][0]
-        bot_response = respond(user_message, history[:-1])
-        history[-1][1] = bot_response
-        return history, ollama.get_stats()
-    
-    def change_model(model):
-        ollama.model = model
-        logger.info(f"Switched to model: {model}")
-        return ollama.get_stats()
-    
-    def refresh_stats():
-        ollama.check_connection()
-        return ollama.get_stats()
-    
-    msg.submit(user_submit, [msg, chatbot], [msg, chatbot, status_box], queue=False).then(
-        bot_respond, chatbot, [chatbot, status_box]
-    )
-    submit_btn.click(user_submit, [msg, chatbot], [msg, chatbot, status_box], queue=False).then(
-        bot_respond, chatbot, [chatbot, status_box]
-    )
-    model_dropdown.change(change_model, model_dropdown, status_box)
-    refresh_btn.click(refresh_stats, outputs=status_box)
+demo = gr.ChatInterface(
+    fn=chat_with_ollama,
+    title="CPU Chatbot",
+    description=f"**Powered by Ollama** ({MODEL})\n\nUsing local AI - no cloud required!",
+    examples=["Hello!", "What is Python?", "Tell me a joke"],
+)
 
-log_banner(logger, "SYSTEM READY - USING OLLAMA")
-logger.info(f"Model: {ollama.model}")
-logger.info("Run 'ollama serve' if not already running")
+logger.info("=" * 60)
+logger.info("  READY - Open http://127.0.0.1:7860")
+logger.info("=" * 60)
 
 if __name__ == "__main__":
     demo.launch()