app.py

import gradio as gr
import requests
import os


# For Hugging Face Spaces deployment:
# Set these environment variables in your Space settings:
# OLLAMA_HOST: Your Ollama server URL
# OLLAMA_MODEL: Model name (e.g., "llama2")

# Get environment variables for deployment
OLLAMA_HOST = os.getenv("OLLAMA_HOST", "http://localhost:11434")
OLLAMA_MODEL = os.getenv("OLLAMA_MODEL", "llama2")
API_URL = f"{OLLAMA_HOST}/api/generate"

def generate_response(prompt):
    try:
        # Make request to local Ollama server
        response = requests.post(
            API_URL, 
            json={
                "model": OLLAMA_MODEL,
                "prompt": prompt,
                "stream": False
            }
        )
        
        # Check for specific error cases
        if response.status_code == 404:
            return "Error: Ollama server not found. Make sure Ollama is running and OLLAMA_HOST is set correctly."
        elif response.status_code == 500:
            return "Error: Server error. Check if the model is loaded in Ollama and OLLAMA_MODEL is set correctly."
        
        response.raise_for_status()
        return response.json()['response']
    except requests.exceptions.RequestException as e:
        return f"Error: {str(e)}"
    except Exception as e:
        return f"Unexpected error: {str(e)}"

# Create Gradio interface
demo = gr.Interface(
    fn=generate_response,
    inputs=gr.Textbox(lines=2, placeholder="Enter your prompt here..."),
    outputs=gr.Textbox(lines=10),
    title="Llama 2 Chat",
    description="Chat with Llama 2 model using local Ollama server. Enter your prompt below and get a response.",
    examples=[
        ["What is artificial intelligence?"],
        ["Tell me a short story about a robot."],
        ["Explain quantum computing in simple terms."]
    ]
)

# Launch the app
if __name__ == "__main__":
    # For local development
    #demo.launch()
    # For Hugging Face Spaces:
    demo.launch(server_name="0.0.0.0", server_port=7860)