app.py

import gradio as gr
import logging
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from abc import ABC, abstractmethod
from typing import Dict, Any
from datetime import datetime
import json
import os
from huggingface_hub import login

# Configure logging
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
    handlers=[
        logging.FileHandler('wellness_assistant.log'),
        logging.StreamHandler()
    ]
)

logger = logging.getLogger("WellnessAssistant")

# Login to Hugging Face Hub
try:
    HF_TOKEN = os.getenv('HF_TOKEN')
    if HF_TOKEN:
        login(token=HF_TOKEN)
        logger.info("Successfully logged in to Hugging Face Hub")
    else:
        logger.warning("HF_TOKEN not found in environment variables")
except Exception as e:
    logger.error(f"Failed to login to Hugging Face Hub: {str(e)}")

class BaseAgent(ABC):
    def __init__(self, name: str, model_id: str):
        """Initialize base agent with common properties"""
        self.name = name
        self.model_id = model_id
        self.logger = logging.getLogger(f"Agent.{name}")
        self.logger.info(f"Initializing {name} with model {model_id}")
        
        try:
            self.model, self.tokenizer = self._load_model()
            self.logger.info(f"Successfully loaded model and tokenizer for {name}")
        except Exception as e:
            self.logger.error(f"Failed to load model for {name}: {str(e)}")
            raise
    
    def _load_model(self):
        """Load the specified model from Hugging Face"""
        self.logger.debug(f"Loading model {self.model_id}")
        try:
            tokenizer = AutoTokenizer.from_pretrained(
                self.model_id,
                token=HF_TOKEN,
                trust_remote_code=True
            )
            model = AutoModelForCausalLM.from_pretrained(
                self.model_id,
                token=HF_TOKEN,
                torch_dtype=torch.float16,
                device_map="auto",
                trust_remote_code=True
            )
            return model, tokenizer
        except Exception as e:
            self.logger.error(f"Error loading model {self.model_id}: {str(e)}")
            raise
    
    def generate_response(self, prompt: str, max_length: int = 512) -> str:
        """Generate response using the model"""
        self.logger.debug(f"Generating response for prompt: {prompt[:100]}...")
        try:
            inputs = self.tokenizer(prompt, return_tensors="pt").to(self.model.device)
            self.logger.debug("Input tokens created successfully")
            
            outputs = self.model.generate(
                **inputs,
                max_length=max_length,
                num_return_sequences=1,
                temperature=0.7,
                top_p=0.9,
                do_sample=True,
                pad_token_id=self.tokenizer.eos_token_id
            )
            self.logger.debug("Model generation completed")
            
            response = self.tokenizer.decode(outputs[0], skip_special_tokens=True)
            response = response[len(prompt):].strip()
            self.logger.debug(f"Generated response: {response[:100]}...")
            return response
            
        except Exception as e:
            self.logger.error(f"Error generating response: {str(e)}")
            return "I apologize, but I'm having trouble generating a response right now."

    @abstractmethod
    def process(self, input_data: Dict[str, Any]) -> Dict[str, Any]:
        """Process input and return response"""
        pass

class TherapeuticAgent(BaseAgent):
    def __init__(self):
        super().__init__(
            name="therapeutic_agent",
            model_id="mistralai/Mistral-7B-Instruct-v0.2"  # Using Mistral model
        )
        self.conversation_history = []
        self.logger.info("Therapeutic agent initialized")
    
    def process(self, input_data: Dict[str, Any]) -> Dict[str, Any]:
        """Process therapeutic conversations"""
        self.logger.info("Processing therapeutic input")
        self.logger.debug(f"Input data: {input_data}")
        
        prompt = self._construct_therapeutic_prompt(input_data["text"])
        response = self.generate_response(prompt)
        
        # Update conversation history
        self.conversation_history.append({
            "timestamp": datetime.now().isoformat(),
            "user": input_data["text"],
            "agent": response
        })
        
        self.logger.info("Successfully processed therapeutic input")
        self.logger.debug(f"Response: {response[:100]}...")
        
        return {
            "response": response,
            "conversation_history": self.conversation_history
        }
    
    def _construct_therapeutic_prompt(self, user_input: str) -> str:
        return f"""<s>[INST] You are a supportive and empathetic mental wellness assistant.
        Your role is to provide caring, thoughtful responses while maintaining appropriate boundaries.
        Always encourage professional help when needed.
        
        User message: {user_input}
        
        Provide a helpful and empathetic response: [/INST]"""

class MindfulnessAgent(BaseAgent):
    def __init__(self):
        super().__init__(
            name="mindfulness_agent",
            model_id="mistralai/Mistral-7B-Instruct-v0.2"  # Using Mistral model
        )
        self.session_history = []
        self.logger.info("Mindfulness agent initialized")
    
    def process(self, input_data: Dict[str, Any]) -> Dict[str, Any]:
        """Process mindfulness-related requests"""
        self.logger.info("Processing mindfulness input")
        self.logger.debug(f"Input data: {input_data}")
        
        prompt = self._construct_mindfulness_prompt(input_data["text"])
        response = self.generate_response(prompt)
        
        # Update session history
        self.session_history.append({
            "timestamp": datetime.now().isoformat(),
            "user": input_data["text"],
            "agent": response
        })
        
        self.logger.info("Successfully processed mindfulness input")
        self.logger.debug(f"Response: {response[:100]}...")
        
        return {
            "response": response,
            "session_history": self.session_history
        }
    
    def _construct_mindfulness_prompt(self, user_input: str) -> str:
        return f"""<s>[INST] You are a mindfulness and meditation guide.
        Your role is to provide calming guidance, meditation instructions, and mindfulness exercises.
        Focus on present-moment awareness and gentle guidance.
        
        User request: {user_input}
        
        Provide mindfulness guidance: [/INST]"""

class WellnessApp:
    def __init__(self):
        self.logger = logging.getLogger("WellnessApp")
        self.logger.info("Initializing Wellness App")
        
        try:
            self.therapeutic_agent = TherapeuticAgent()
            self.mindfulness_agent = MindfulnessAgent()
            self.logger.info("Successfully initialized all agents")
        except Exception as e:
            self.logger.error(f"Failed to initialize agents: {str(e)}")
            raise
        
        self.current_agent = "therapeutic"  # Default agent
    
    def switch_agent(self, agent_type: str) -> str:
        """Switch between therapeutic and mindfulness agents"""
        self.logger.info(f"Switching to {agent_type} agent")
        self.current_agent = agent_type
        return f"Switched to {agent_type} mode"
    
    def respond(self, message: str, history: list) -> str:
        """Process user message and return agent response"""
        self.logger.info(f"Processing message with {self.current_agent} agent")
        self.logger.debug(f"Message: {message}")
        
        try:
            if self.current_agent == "therapeutic":
                response = self.therapeutic_agent.process({"text": message})
            else:
                response = self.mindfulness_agent.process({"text": message})
            
            self.logger.info("Successfully generated response")
            return response["response"]
            
        except Exception as e:
            self.logger.error(f"Error processing message: {str(e)}")
            return "I apologize, but I'm having trouble processing your message right now."
    
    def create_interface(self):
        """Create Gradio interface"""
        self.logger.info("Creating Gradio interface")
        
        with gr.Blocks(theme=gr.themes.Soft()) as demo:
            gr.Markdown("# Mental Wellness Assistant")
            
            with gr.Row():
                therapeutic_btn = gr.Button("Therapeutic Mode")
                mindfulness_btn = gr.Button("Mindfulness Mode")
            
            chatbot = gr.ChatInterface(
                fn=self.respond,
                examples=[
                    "I've been feeling anxious lately",
                    "Guide me through a breathing exercise",
                    "I need help managing stress",
                    "Can you teach me meditation?"
                ],
                title="",
            )
            
            therapeutic_btn.click(
                fn=lambda: self.switch_agent("therapeutic"),
                outputs=gr.Textbox(label="Status")
            )
            mindfulness_btn.click(
                fn=lambda: self.switch_agent("mindfulness"),
                outputs=gr.Textbox(label="Status")
            )
            
            gr.Markdown("""
            ### Important Notice
            This is a demo AI assistant and not a substitute for professional mental health care.
            If you're experiencing a mental health crisis, please contact emergency services or a mental health professional.
            """)
        
        self.logger.info("Gradio interface created successfully")
        return demo

# Create and launch the app
def main():
    logger.info("Starting Wellness Assistant application")
    
    try:
        app = WellnessApp()
        demo = app.create_interface()
        logger.info("Application initialized successfully")
        
        if __name__ == "__main__":
            logger.info("Launching Gradio interface")
            demo.launch()
            
    except Exception as e:
        logger.error(f"Failed to start application: {str(e)}")
        raise

main()