app.py

import gradio as gr
from huggingface_hub import InferenceClient
import os
import json
import base64
from PIL import Image
import io
import requests
from typing import Dict, List, Optional, Any, Union
import time
import logging

# Setup logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
logger = logging.getLogger(__name__)

ACCESS_TOKEN = os.getenv("HF_TOKEN")
logger.info("Access token loaded.")

# MCP Client Configuration
MCP_SERVERS = {}
try:
    mcp_config = os.getenv("MCP_CONFIG")
    if mcp_config:
        MCP_SERVERS = json.loads(mcp_config)
        logger.info(f"Loaded MCP configuration: {len(MCP_SERVERS)} servers defined")
except Exception as e:
    logger.error(f"Error loading MCP configuration: {e}")

# Function to encode image to base64
def encode_image(image_path):
    if not image_path:
        logger.warning("No image path provided")
        return None
    
    try:
        logger.info(f"Encoding image from path: {image_path}")
        
        # If it's already a PIL Image
        if isinstance(image_path, Image.Image):
            image = image_path
        else:
            # Try to open the image file
            image = Image.open(image_path)
        
        # Convert to RGB if image has an alpha channel (RGBA)
        if image.mode == 'RGBA':
            image = image.convert('RGB')
        
        # Encode to base64
        buffered = io.BytesIO()
        image.save(buffered, format="JPEG")
        img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")
        logger.info("Image encoded successfully")
        return img_str
    except Exception as e:
        logger.error(f"Error encoding image: {e}")
        return None

# MCP Client implementation
class MCPClient:
    def __init__(self, server_url: str):
        self.server_url = server_url
        self.session_id = None
        logger.info(f"Initialized MCP Client for server: {server_url}")
    
    def connect(self) -> bool:
        """Establish connection with the MCP server"""
        try:
            response = requests.post(
                f"{self.server_url}/connect",
                json={"client": "Serverless-TextGen-Hub", "version": "1.0.0"}
            )
            if response.status_code == 200:
                result = response.json()
                self.session_id = result.get("session_id")
                logger.info(f"Connected to MCP server with session ID: {self.session_id}")
                return True
            else:
                logger.error(f"Failed to connect to MCP server: {response.status_code} - {response.text}")
                return False
        except Exception as e:
            logger.error(f"Error connecting to MCP server: {e}")
            return False
    
    def list_tools(self) -> List[Dict]:
        """List available tools from the MCP server"""
        if not self.session_id:
            if not self.connect():
                return []
        
        try:
            response = requests.get(
                f"{self.server_url}/tools/list",
                headers={"X-MCP-Session": self.session_id}
            )
            if response.status_code == 200:
                result = response.json()
                tools = result.get("tools", [])
                logger.info(f"Retrieved {len(tools)} tools from MCP server")
                return tools
            else:
                logger.error(f"Failed to list tools: {response.status_code} - {response.text}")
                return []
        except Exception as e:
            logger.error(f"Error listing tools: {e}")
            return []
    
    def call_tool(self, tool_name: str, args: Dict) -> Dict:
        """Call a tool on the MCP server"""
        if not self.session_id:
            if not self.connect():
                return {"error": "Not connected to MCP server"}
        
        try:
            response = requests.post(
                f"{self.server_url}/tools/call",
                headers={"X-MCP-Session": self.session_id},
                json={"name": tool_name, "arguments": args}
            )
            if response.status_code == 200:
                result = response.json()
                logger.info(f"Successfully called tool {tool_name}")
                return result
            else:
                error_msg = f"Failed to call tool {tool_name}: {response.status_code} - {response.text}"
                logger.error(error_msg)
                return {"error": error_msg}
        except Exception as e:
            error_msg = f"Error calling tool {tool_name}: {e}"
            logger.error(error_msg)
            return {"error": error_msg}

# Text-to-speech client function
def text_to_speech(text: str, server_name: str = None) -> Optional[str]:
    """
    Convert text to speech using an MCP TTS server
    Returns an audio URL that can be embedded in the chat
    """
    if not server_name or server_name not in MCP_SERVERS:
        logger.warning(f"TTS server {server_name} not configured")
        return None
    
    server_url = MCP_SERVERS[server_name].get("url")
    if not server_url:
        logger.warning(f"No URL found for TTS server {server_name}")
        return None
    
    client = MCPClient(server_url)
    
    # List available tools to find the TTS tool
    tools = client.list_tools()
    tts_tool = next((t for t in tools if "text_to_audio" in t["name"] or "tts" in t["name"]), None)
    
    if not tts_tool:
        logger.warning(f"No TTS tool found on server {server_name}")
        return None
    
    # Call the TTS tool
    result = client.call_tool(tts_tool["name"], {"text": text, "speed": 1.0})
    
    if "error" in result:
        logger.error(f"TTS error: {result['error']}")
        return None
    
    # Process the result - usually a base64 encoded WAV
    audio_data = result.get("audio") or result.get("content") or result.get("result")
    
    if isinstance(audio_data, str) and audio_data.startswith("data:audio"):
        # Already a data URL
        return audio_data
    elif isinstance(audio_data, str):
        # Assume it's base64 encoded
        return f"data:audio/wav;base64,{audio_data}"
    else:
        logger.error(f"Unexpected TTS result format: {type(audio_data)}")
        return None

def respond(
    message,
    image_files,  # Changed parameter name and structure
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
    frequency_penalty,
    seed,
    provider,
    custom_api_key,
    custom_model,    
    model_search_term,
    selected_model,
    tts_enabled=False,
    tts_server=None
):
    logger.info(f"Received message: {message}")
    logger.info(f"Received {len(image_files) if image_files else 0} images")
    logger.info(f"History: {history}")
    logger.info(f"System message: {system_message}")
    logger.info(f"Max tokens: {max_tokens}, Temperature: {temperature}, Top-P: {top_p}")
    logger.info(f"Frequency Penalty: {frequency_penalty}, Seed: {seed}")
    logger.info(f"Selected provider: {provider}")         
    logger.info(f"Custom API Key provided: {bool(custom_api_key.strip())}")
    logger.info(f"Selected model (custom_model): {custom_model}")  
    logger.info(f"Model search term: {model_search_term}")
    logger.info(f"Selected model from radio: {selected_model}")
    logger.info(f"TTS enabled: {tts_enabled}, TTS server: {tts_server}")

    # Determine which token to use
    token_to_use = custom_api_key if custom_api_key.strip() != "" else ACCESS_TOKEN
    
    if custom_api_key.strip() != "":
        logger.info("USING CUSTOM API KEY: BYOK token provided by user is being used for authentication")
    else:
        logger.info("USING DEFAULT API KEY: Environment variable HF_TOKEN is being used for authentication")
    
    # Initialize the Inference Client with the provider and appropriate token
    client = InferenceClient(token=token_to_use, provider=provider)
    logger.info(f"Hugging Face Inference Client initialized with {provider} provider.")

    # Convert seed to None if -1 (meaning random)
    if seed == -1:
        seed = None

    # Create multimodal content if images are present
    if image_files and len(image_files) > 0:
        # Process the user message to include images
        user_content = []
        
        # Add text part if there is any
        if message and message.strip():
            user_content.append({
                "type": "text",
                "text": message
            })
        
        # Add image parts
        for img in image_files:
            if img is not None:
                # Get raw image data from path
                try:
                    encoded_image = encode_image(img)
                    if encoded_image:
                        user_content.append({
                            "type": "image_url",
                            "image_url": {
                                "url": f"data:image/jpeg;base64,{encoded_image}"
                            }
                        })
                except Exception as e:
                    logger.error(f"Error encoding image: {e}")
    else:
        # Text-only message
        user_content = message

    # Prepare messages in the format expected by the API
    messages = [{"role": "system", "content": system_message}]
    logger.info("Initial messages array constructed.")

    # Add conversation history to the context
    for val in history:
        user_part = val[0]
        assistant_part = val[1]
        if user_part:
            # Handle both text-only and multimodal messages in history
            if isinstance(user_part, tuple) and len(user_part) == 2:
                # This is a multimodal message with text and images
                history_content = []
                if user_part[0]:  # Text
                    history_content.append({
                        "type": "text",
                        "text": user_part[0]
                    })
                
                for img in user_part[1]:  # Images
                    if img:
                        try:
                            encoded_img = encode_image(img)
                            if encoded_img:
                                history_content.append({
                                    "type": "image_url",
                                    "image_url": {
                                        "url": f"data:image/jpeg;base64,{encoded_img}"
                                    }
                                })
                        except Exception as e:
                            logger.error(f"Error encoding history image: {e}")
                
                messages.append({"role": "user", "content": history_content})
            else:
                # Regular text message
                messages.append({"role": "user", "content": user_part})
            logger.info(f"Added user message to context (type: {type(user_part)})")
        
        if assistant_part:
            messages.append({"role": "assistant", "content": assistant_part})
            logger.info(f"Added assistant message to context: {assistant_part}")

    # Append the latest user message
    messages.append({"role": "user", "content": user_content})
    logger.info(f"Latest user message appended (content type: {type(user_content)})")

    # Determine which model to use, prioritizing custom_model if provided
    model_to_use = custom_model.strip() if custom_model.strip() != "" else selected_model
    logger.info(f"Model selected for inference: {model_to_use}")

    # Start with an empty string to build the response as tokens stream in
    response = ""
    logger.info(f"Sending request to {provider} provider.")

    # Prepare parameters for the chat completion request
    parameters = {
        "max_tokens": max_tokens,
        "temperature": temperature,
        "top_p": top_p,
        "frequency_penalty": frequency_penalty,
    }
    
    if seed is not None:
        parameters["seed"] = seed

    # Use the InferenceClient for making the request
    try:
        # Create a generator for the streaming response
        stream = client.chat_completion(
            model=model_to_use,
            messages=messages,
            stream=True,
            **parameters
        )
        
        logger.info("Received tokens: ")
        
        # Process the streaming response
        for chunk in stream:
            if hasattr(chunk, 'choices') and len(chunk.choices) > 0:
                # Extract the content from the response
                if hasattr(chunk.choices[0], 'delta') and hasattr(chunk.choices[0].delta, 'content'):
                    token_text = chunk.choices[0].delta.content
                    if token_text:
                        print(token_text, end="", flush=True)
                        response += token_text
                        yield response
        
        # If TTS is enabled and we have a response, convert it to speech
        if tts_enabled and tts_server and response:
            logger.info(f"Converting response to speech using TTS server: {tts_server}")
            try:
                audio_url = text_to_speech(response, tts_server)
                if audio_url:
                    # Add audio tag to the end of the response
                    response += f"\n\n<audio src='{audio_url}' controls></audio>"
                    yield response
                else:
                    logger.warning("TTS conversion failed, continuing without audio")
            except Exception as e:
                logger.error(f"Error in TTS conversion: {e}")
                # Continue without TTS if there's an error
        
        print()
    except Exception as e:
        logger.error(f"Error during inference: {e}")
        response += f"\nError: {str(e)}"
        yield response

    logger.info("Completed response generation.")

# Function to validate provider selection based on BYOK
def validate_provider(api_key, provider):
    if not api_key.strip() and provider != "hf-inference":
        return gr.update(value="hf-inference")
    return gr.update(value=provider)

# Function to list available MCP servers
def list_mcp_servers():
    """List all configured MCP servers"""
    return list(MCP_SERVERS.keys())

# GRADIO UI
with gr.Blocks(theme="Nymbo/Nymbo_Theme") as demo:
    # Create the chatbot component
    chatbot = gr.Chatbot(
        height=600, 
        show_copy_button=True, 
        placeholder="Select a model and begin chatting. Now supports multiple inference providers, multimodal inputs, and MCP servers",
        layout="panel"
    )
    logger.info("Chatbot interface created.")
    
    # Multimodal textbox for messages (combines text and file uploads)
    msg = gr.MultimodalTextbox(
        placeholder="Type a message or upload images...",
        show_label=False,
        container=False,
        scale=12,
        file_types=["image"],
        file_count="multiple",
        sources=["upload"]
    )
    
    # Create accordion for settings
    with gr.Accordion("Settings", open=False):
        # System message
        system_message_box = gr.Textbox(
            value="You are a helpful AI assistant that can understand images and text.", 
            placeholder="You are a helpful assistant.",
            label="System Prompt"
        )
        
        # Generation parameters
        with gr.Row():
            with gr.Column():
                max_tokens_slider = gr.Slider(
                    minimum=1,
                    maximum=4096,
                    value=512,
                    step=1,
                    label="Max tokens"
                )
                
                temperature_slider = gr.Slider(
                    minimum=0.1,
                    maximum=4.0,
                    value=0.7,
                    step=0.1,
                    label="Temperature"
                )
                
                top_p_slider = gr.Slider(
                    minimum=0.1,
                    maximum=1.0,
                    value=0.95,
                    step=0.05,
                    label="Top-P"
                )
                
            with gr.Column():
                frequency_penalty_slider = gr.Slider(
                    minimum=-2.0,
                    maximum=2.0,
                    value=0.0,
                    step=0.1,
                    label="Frequency Penalty"
                )
                
                seed_slider = gr.Slider(
                    minimum=-1,
                    maximum=65535,
                    value=-1,
                    step=1,
                    label="Seed (-1 for random)"
                )
        
        # Provider selection
        providers_list = [
            "hf-inference",  # Default Hugging Face Inference
            "cerebras",      # Cerebras provider
            "together",      # Together AI
            "sambanova",     # SambaNova
            "novita",        # Novita AI
            "cohere",        # Cohere
            "fireworks-ai",  # Fireworks AI
            "hyperbolic",    # Hyperbolic
            "nebius",        # Nebius
        ]
        
        provider_radio = gr.Radio(
            choices=providers_list,
            value="hf-inference",
            label="Inference Provider",
        )
        
        # New BYOK textbox
        byok_textbox = gr.Textbox(
            value="",
            label="BYOK (Bring Your Own Key)",
            info="Enter a custom Hugging Face API key here. When empty, only 'hf-inference' provider can be used.",
            placeholder="Enter your Hugging Face API token",
            type="password"  # Hide the API key for security
        )
        
        # Custom model box
        custom_model_box = gr.Textbox(
            value="",
            label="Custom Model",
            info="(Optional) Provide a custom Hugging Face model path. Overrides any selected featured model.",
            placeholder="meta-llama/Llama-3.3-70B-Instruct"
        )
        
        # Model search
        model_search_box = gr.Textbox(
            label="Filter Models",
            placeholder="Search for a featured model...",
            lines=1
        )
        
        # Featured models list
        # Updated to include multimodal models
        models_list = [
            "meta-llama/Llama-3.2-11B-Vision-Instruct",
            "meta-llama/Llama-3.3-70B-Instruct",
            "meta-llama/Llama-3.1-70B-Instruct",
            "meta-llama/Llama-3.0-70B-Instruct",
            "meta-llama/Llama-3.2-3B-Instruct",
            "meta-llama/Llama-3.2-1B-Instruct",
            "meta-llama/Llama-3.1-8B-Instruct",
            "NousResearch/Hermes-3-Llama-3.1-8B",
            "NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO",
            "mistralai/Mistral-Nemo-Instruct-2407",
            "mistralai/Mixtral-8x7B-Instruct-v0.1",
            "mistralai/Mistral-7B-Instruct-v0.3",
            "mistralai/Mistral-7B-Instruct-v0.2",
            "Qwen/Qwen3-235B-A22B",
            "Qwen/Qwen3-32B",
            "Qwen/Qwen2.5-72B-Instruct",
            "Qwen/Qwen2.5-3B-Instruct",
            "Qwen/Qwen2.5-0.5B-Instruct",
            "Qwen/QwQ-32B",
            "Qwen/Qwen2.5-Coder-32B-Instruct",
            "microsoft/Phi-3.5-mini-instruct",
            "microsoft/Phi-3-mini-128k-instruct",
            "microsoft/Phi-3-mini-4k-instruct",
        ]

        featured_model_radio = gr.Radio(
            label="Select a model below",
            choices=models_list,
            value="meta-llama/Llama-3.2-11B-Vision-Instruct",  # Default to a multimodal model
            interactive=True
        )
        
        gr.Markdown("[View all Text-to-Text models](https://huggingface.co/models?inference_provider=all&pipeline_tag=text-generation&sort=trending) | [View all multimodal models](https://huggingface.co/models?inference_provider=all&pipeline_tag=image-text-to-text&sort=trending)")
        
        # MCP TTS integration
        with gr.Accordion("MCP Integration", open=False):
            gr.Markdown("## Model Context Protocol (MCP) Integration")
            gr.Markdown("Connect to MCP servers to extend functionality.")
            
            tts_enabled = gr.Checkbox(
                label="Enable Text-to-Speech",
                value=False,
                info="When enabled, responses will be converted to speech using the selected MCP TTS server"
            )
            
            # Create dropdown for available MCP servers
            available_servers = list_mcp_servers()
            tts_server = gr.Dropdown(
                label="TTS Server",
                choices=available_servers,
                value=available_servers[0] if available_servers else None,
                interactive=True,
                visible=len(available_servers) > 0
            )
            
            # If no servers configured, show a message
            if not available_servers:
                gr.Markdown("""
                No MCP servers configured. Add them using the MCP_CONFIG environment variable:
                ```json
                {
                  "kokoroTTS": {
                    "url": "https://your-kokoro-tts-server/gradio_api/mcp/sse"
                  }
                }
                ```
                """)

    # Chat history state
    chat_history = gr.State([])
    
    # Function to filter models
    def filter_models(search_term):
        logger.info(f"Filtering models with search term: {search_term}")
        filtered = [m for m in models_list if search_term.lower() in m.lower()]
        logger.info(f"Filtered models: {filtered}")
        return gr.update(choices=filtered)

    # Function to set custom model from radio
    def set_custom_model_from_radio(selected):
        logger.info(f"Featured model selected: {selected}")
        return selected

    # Function for the chat interface
    def user(user_message, history):
        # Debug logging for troubleshooting
        logger.info(f"User message received: {user_message}")
        
        # Skip if message is empty (no text and no files)
        if not user_message or (not user_message.get("text") and not user_message.get("files")):
            logger.info("Empty message, skipping")
            return history
        
        # Prepare multimodal message format
        text_content = user_message.get("text", "").strip()
        files = user_message.get("files", [])
        
        logger.info(f"Text content: {text_content}")
        logger.info(f"Files: {files}")
        
        # If both text and files are empty, skip
        if not text_content and not files:
            logger.info("No content to display")
            return history
        
        # Add message with images to history
        if files and len(files) > 0:
            # Add text message first if it exists
            if text_content:
                # Add a separate text message
                logger.info(f"Adding text message: {text_content}")
                history.append([text_content, None])
            
            # Then add each image file separately
            for file_path in files:
                if file_path and isinstance(file_path, str):
                    logger.info(f"Adding image: {file_path}")
                    # Add image as a separate message with no text
                    history.append([f"![Image]({file_path})", None])
            
            return history
        else:
            # For text-only messages
            logger.info(f"Adding text-only message: {text_content}")
            history.append([text_content, None])
            return history
    
    # Define bot response function
    def bot(history, system_msg, max_tokens, temperature, top_p, freq_penalty, seed, provider, api_key, custom_model, search_term, selected_model, tts_enabled, tts_server):
        # Check if history is valid
        if not history or len(history) == 0:
            logger.info("No history to process")
            return history
        
        # Get the most recent message and detect if it's an image
        user_message = history[-1][0]
        logger.info(f"Processing user message: {user_message}")
        
        is_image = False
        image_path = None
        text_content = user_message
        
        # Check if this is an image message (marked with ![Image])
        if isinstance(user_message, str) and user_message.startswith("![Image]("):
            is_image = True
            # Extract image path from markdown format ![Image](path)
            image_path = user_message.replace("![Image](", "").replace(")", "")
            logger.info(f"Image detected: {image_path}")
            text_content = ""  # No text for image-only messages
        
        # Look back for text context if this is an image
        text_context = ""
        if is_image and len(history) > 1:
            # Use the previous message as context if it's text
            prev_message = history[-2][0]
            if isinstance(prev_message, str) and not prev_message.startswith("![Image]("):
                text_context = prev_message
                logger.info(f"Using text context from previous message: {text_context}")
        
        # Process message through respond function
        history[-1][1] = ""
        
        # Use either the image or text for the API
        if is_image:
            # For image messages
            for response in respond(
                text_context,           # Text context from previous message if any
                [image_path],           # Current image
                history[:-1],           # Previous history
                system_msg,
                max_tokens,
                temperature,
                top_p,
                freq_penalty,
                seed,
                provider,
                api_key,
                custom_model,
                search_term,
                selected_model,
                tts_enabled,
                tts_server
            ):
                history[-1][1] = response
                yield history
        else:
            # For text-only messages
            for response in respond(
                text_content,           # Text message
                None,                   # No image
                history[:-1],           # Previous history
                system_msg,
                max_tokens,
                temperature,
                top_p,
                freq_penalty,
                seed,
                provider,
                api_key,
                custom_model,
                search_term,
                selected_model,
                tts_enabled,
                tts_server
            ):
                history[-1][1] = response
                yield history

    # Event handlers - only using the MultimodalTextbox's built-in submit functionality
    msg.submit(
        user,
        [msg, chatbot],
        [chatbot],
        queue=False
    ).then(
        bot,
        [chatbot, system_message_box, max_tokens_slider, temperature_slider, top_p_slider, 
         frequency_penalty_slider, seed_slider, provider_radio, byok_textbox, custom_model_box, 
         model_search_box, featured_model_radio, tts_enabled, tts_server],
        [chatbot]
    ).then(
        lambda: {"text": "", "files": []},  # Clear inputs after submission
        None,
        [msg]
    )
    
    # Connect the model filter to update the radio choices
    model_search_box.change(
        fn=filter_models,
        inputs=model_search_box,
        outputs=featured_model_radio
    )
    logger.info("Model search box change event linked.")

    # Connect the featured model radio to update the custom model box
    featured_model_radio.change(
        fn=set_custom_model_from_radio,
        inputs=featured_model_radio,
        outputs=custom_model_box
    )
    logger.info("Featured model radio button change event linked.")
    
    # Connect the BYOK textbox to validate provider selection
    byok_textbox.change(
        fn=validate_provider,
        inputs=[byok_textbox, provider_radio],
        outputs=provider_radio
    )
    logger.info("BYOK textbox change event linked.")

    # Also validate provider when the radio changes to ensure consistency
    provider_radio.change(
        fn=validate_provider,
        inputs=[byok_textbox, provider_radio],
        outputs=provider_radio
    )
    logger.info("Provider radio button change event linked.")
    
    # Update TTS server dropdown visibility based on the TTS toggle
    tts_enabled.change(
        lambda enabled: gr.update(visible=enabled and len(list_mcp_servers()) > 0),
        inputs=tts_enabled,
        outputs=tts_server
    )

logger.info("Gradio interface initialized.")

if __name__ == "__main__":
    logger.info("Launching the demo application.")
    demo.launch(show_api=True)