app.py

# app.py - Robust Local Ollama Chatbot
# Features: Dynamic Connection, Model Selection, Robust History
import gradio as gr
import requests
import logging
import sys
import shutil
from datetime import datetime

# Logging
logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s | %(levelname)-8s | %(message)s",
    handlers=[logging.StreamHandler(sys.stdout)]
)
logger = logging.getLogger("CHATBOT")

OLLAMA_URL = "http://localhost:11434"

# Global state
current_model = ""
available_models = []

def get_ollama_models():
    """Fetch available models from Ollama."""
    try:
        r = requests.get(f"{OLLAMA_URL}/api/tags", timeout=2)
        if r.status_code == 200:
            return [m["name"] for m in r.json().get("models", [])]
    except:
        pass
    return []

def check_connection():
    """Check connection and update model list."""
    global current_model, available_models
    
    if shutil.which("ollama") is None:
        return "❌ **Ollama not installed**", gr.update(choices=[], value=None)
    
    models = get_ollama_models()
    available_models = models
    
    if not models:
        return "⚠️ Ollama running but **NO models found**", gr.update(choices=[], value=None)
    
    # Default logic: Prefer llama3.2, then deepseek, then first available
    if not current_model or current_model not in models:
        if "llama3.2:3b" in models:
            current_model = "llama3.2:3b"
        elif "deepseek-coder:6.7b-instruct-q6_K" in models:
            current_model = "deepseek-coder:6.7b-instruct-q6_K"
        else:
            current_model = models[0]
            
    return f"✅ Connected", gr.update(choices=models, value=current_model)

def chat_fn(message, history, model_name):
    """Chat function handling history and model selection."""
    if not model_name:
        return "Error: No model selected. Please check connection."
        
    # Build prompt
    prompt = "You are a helpful AI assistant.\n\n"
    for item in history:
        if isinstance(item, dict):
            role = item.get("role", "")
            content = item.get("content", "")
            if role == "user":
                prompt += f"User: {content}\n"
            elif role == "assistant":
                prompt += f"Assistant: {content}\n"
    prompt += f"User: {message}\nAssistant:"
    
    try:
        r = requests.post(
            f"{OLLAMA_URL}/api/generate",
            json={"model": model_name, "prompt": prompt, "stream": False},
            timeout=120
        )
        if r.status_code == 200:
            return r.json().get("response", "").strip()
        return f"Error: Ollama returned {r.status_code}"
    except Exception as e:
        return f"Error: {e}"

# UI
with gr.Blocks(title="Local Chatbot") as demo:
    gr.Markdown("# 🤖 Local AI Chatbot")
    
    with gr.Row():
        status_display = gr.Markdown("**Checking connection...**")
        refresh_btn = gr.Button("🔄 Check Connection", size="sm")
    
    with gr.Row():
        model_dropdown = gr.Dropdown(
            label="Select Model",
            choices=[],
            value=None,
            interactive=True
        )
    
    chat_interface = gr.ChatInterface(
        fn=chat_fn,
        additional_inputs=[model_dropdown],
    )
    
    with gr.Accordion("🛠️ Setup Instructions", open=False):
        gr.Markdown("""
        1. **Install Ollama:** [ollama.com](https://ollama.com)
        2. **Start Server:** `ollama serve`
        3. **Download Model:** `ollama pull llama3.2:3b`
        """)

    # Wiring
    def update_all():
        status, model_update = check_connection()
        return status, model_update

    refresh_btn.click(update_all, outputs=[status_display, model_dropdown])
    demo.load(update_all, outputs=[status_display, model_dropdown])

if __name__ == "__main__":
    demo.launch()