app.py

import os
import gradio as gr
import openai
from langdetect import detect

# Set up OpenAI API with your custom endpoint
openai.api_key = os.getenv("API_KEY")
openai.api_base = "https://api.groq.com/openai/v1"

# Import datasets from the Python files in your project
from datasets.company_profile import company_profile
from datasets.workforce import workforce
from datasets.financials import financials
from datasets.investors import investors
from datasets.products_services import products_services
from datasets.market_trends import market_trends
from datasets.partnerships_collaborations import partnerships_collaborations
from datasets.legal_compliance import legal_compliance
from datasets.customer_insights import customer_insights
from datasets.news_updates import news_updates
from datasets.social_media import social_media
from datasets.tech_stack import tech_stack

# Command handler for specific queries
def command_handler(user_input):
    if user_input.lower().startswith("define "):
        term = user_input[7:].strip()
        definitions = {
            "market analysis": (
                "Market analysis is the process of evaluating a business's position in the market by assessing "
                "competitors, trends, and customer behavior to make informed decisions for growth and investment. "
                "Think of it as gathering the insights needed to craft a winning strategy. 📊"
            ),
            "financials": (
                "Financial analysis provides a snapshot of a company's fiscal health. It involves reviewing metrics like "
                "profit margins, revenues, and expenditures to assess if the business is sustainable and scalable. 💵"
            ),
            "investors": (
                "Investors are the capital partners that fuel innovation and expansion. They provide financial resources "
                "in exchange for equity or debt, aiming to generate a return on their investment. It's all about leveraging "
                "their capital to scale. 🏦"
            )
        }
        return definitions.get(term.lower(), "I don't have a definition for that term yet. Let’s dive into your query directly!")
    return None

# Function to get the response from OpenAI with professionalism and energy
def get_groq_response(message, user_language):
    try:
        response = openai.ChatCompletion.create(
            model="llama-3.1-70b-versatile",
            messages=[
                {
                    "role": "system",
                    "content": (
                        f"You are a professional, energetic Private Market Analyst AI. Your task is to explain market trends, "
                        f"company insights, and investment strategies in a clear, engaging, and authoritative manner. "
                        f"While providing top-tier market analysis, keep the tone polished, insightful, and relevant to business leaders. "
                        f"Always maintain professionalism while making the content informative and engaging."
                    )
                },
                {"role": "user", "content": message}
            ]
        )
        return response.choices[0].message["content"]
    except Exception as e:
        return f"Oops, looks like something went wrong! Error: {str(e)}"

# Function to handle the interaction and queries
def market_analysis_agent(user_input, history=[]):
    try:
        # Detect the language of the user's input
        detected_language = detect(user_input)
        user_language = "Hindi" if detected_language == "hi" else "English"
        
        # Handle special commands like "Define [term]"
        command_response = command_handler(user_input)
        if command_response:
            history.append((user_input, command_response))
            return history, history
        
        # Handle private market queries with datasets
        if "company" in user_input.lower():
            response = company_profile
        elif "financials" in user_input.lower():
            response = financials
        elif "investors" in user_input.lower():
            response = investors
        elif "products" in user_input.lower():
            response = products_services
        elif "workforce" in user_input.lower():
            response = workforce
        else:
            # Get dynamic AI response if query doesn't match predefined terms
            response = get_groq_response(user_input, user_language)

        # Add some professional and engaging replies for the user
        cool_replies = [
            "That’s a great insight! Keep the questions coming. 🔍",
            "Excellent direction! Let’s explore that further. 📈",
            "Insightful! Let’s dive into the numbers. 📊",
            "You've got the right focus. Let's sharpen those strategies. 🧠",
            "You're on the right track. Let’s optimize that idea! 🔧"
        ]
        response = f"{response} {cool_replies[hash(user_input) % len(cool_replies)]}"
        
        # Add to chat history
        history.append((user_input, response))
        return history, history

    except Exception as e:
        return [(user_input, f"Oops, something went wrong: {str(e)}")], history

# Gradio Interface setup
chat_interface = gr.Interface(
    fn=market_analysis_agent,  # Function for handling user interaction
    inputs=["text", "state"],  # Inputs: user message and chat history
    outputs=["chatbot", "state"],  # Outputs: chatbot messages and updated history
    live=False,  # Disable live responses; show after submit
    title="Private Market AI Agent",  # Title of the app
    description=(
        "Welcome to your professional Private Market Analyst AI! 📊\n\n"
        "Ask me anything about market trends, company profiles, financial analysis, investors, and more! "
        "I’ll provide you with actionable insights, backed by data and clear explanations, to help you make informed decisions. "
        "Let’s break down the complexities of the market with precision and insight. 🔍"
    )
)

# Launch the Gradio interface
if __name__ == "__main__":
    chat_interface.launch()