finance_agent.py

import os
from groq import Groq
import yfinance as yf
import datetime
import pandas as pd
import json
from dotenv import load_dotenv
load_dotenv()
from scrapeNews import extract_news
from tickerMap import company_ticker_map
from functions import analyze_financials,analyze_sector,get_Full_ticker,get_stock_price



client = Groq(api_key=os.environ.get("GROQ_API_KEY"))


tools=[
        {
            "type":"function",
            "function":{
                "name":"get_stock_price",
                "description":"Get the stock price of the given stock",
                "parameters":{
                    "type":"object",
                    "properties":{
                        "symbol":{
                            "type":"string",
                            "description":"Fetch real-time stock price by using the ticker symbol. Always use this when the user asks about a stock or share price."
                        }
                    },
                    "required":["symbol"]
                }
            }
        },

        {
            "type":"function",
            "function":{
                "name":"analyze_financials",
                "description":"Summarize the key financial metrics for a given stock including FCF, Net Income, P/E Ratio, etc.",
                "parameters":{
                    "type":"object",
                    "properties":{
                        "symbol":{
                            "type":"string",
                            "description":"Analyze key financial metrics of a company using its stock ticker. Use this to provide investment insights based on revenue, profit, growth trends and different ratios"
                        }
                    },
                    "required":["symbol"]
                }
            }
        },
        {
                "type": "function",
                "function": {
                    "name": "analyze_sector",
                    "description": "Analyze the sector by inferring it from a stock ticker (e.g., 'TCS sector analysis','Tell me how the sector related to CIPLA is doing ').",

                    "parameters": {
                    "type": "object",
                    "properties": {
                        "symbol": {
                        "type": "string",
                        "description": "A stock ticker (e.g., TCS.NS or AAPL). If provided, the sector will be inferred from this ticker."
                        },
                        "sector": {
                        "type": "string",
                        "description": "The name of the sector (e.g., 'it', 'pharma', 'banking'). If no ticker is provided, this will be used directly."
                        }
                    },
                    "required": []
                }
            }
        }

    ]


def sentiment_analysis(symbol):
    news = extract_news(symbol)

    chat_completion = client.chat.completions.create(
        messages=[
            {
                "role":"system",
                "content":"You are a financial news sentiment analyzer and give brief explantion as to how you classified them"
            },
            {
                "role":"user",
                "content":f"""Classify the market sentiment as Neutral,Positive or Negative according to the 

                //{news}//
                Also, briefly explain why.

                When giving the explantions be crisp and to point
                
                Summarize the most influential news item

                Use emojis like ✅ 📈 💰 🚀 🟢 for positive, ❌ ⚠️ 📉 🔻 🔴 for negative and📎 🟡 🤝 for neutral and give good spacing
                
                """
            }
        ],
        
        model="llama3-70b-8192",
        max_tokens=300
    )

    return chat_completion.choices[0].message.content





def classify_levl(user_prompt):
    level_identifier_prompt= f""""
             Classify the user's knowledge as:

            - **Beginner**: Asks basic questions like "Whats the price of the stock",,asks to explain basic finance terms like P/E ratio,free cash flow,Income statement and similar terms and says "what to buy",phrases like "I am new to this","explain in simple terms" and so on.
            
            -**Amateur**: Uses terms like "risk vs return","undervalued","overvalued","Company financials and stability" etc.

            --**Seasoned**: Asks detailed questions based on phrases such as "capital allocation","long term stability of stock","FCF","beta" and similar advanced financial terms and short forms.

            Analyze // {user_prompt} //
            
             Which category does it most closely reflect? Respond only with one of: **beginner**, **amateur**, or **seasoned**.

             If unclear dont make any assumption and carry on

                    """
    
    chat_completion = client.chat.completions.create(
        messages=[
            {
                "role":"system",
                "content":"You classify the level of the investor based on user queries"
            },
            {
                "role":"user",
                "content":level_identifier_prompt
            }
        ],
        
        model="llama3-70b-8192",
        max_tokens=5
    )
    return chat_completion.choices[0].message.content.strip().lower()


def run_finance_agent(user_prompt,conversation_hist):
    level = classify_levl(user_prompt)
    print(len(conversation_hist))
    sector_analysis_added = False

    print(user_prompt)
    
    
    conversation_hist=conversation_hist[-3:]
    chat_completion = client.chat.completions.create(
        messages=[
        {    "role":"system",
                "content":f"""
        You are a smart AI finance chatbot.

        The user is a **{level}** investor:
        
        -  Beginner: explain financial terms briefly and clearly.
        -  Amateur: use relevant financial vocabulary, but keep things clear.
        - Seasoned: go deep into financial metrics, use ratios, and be concise.

        Your goal is to analyze the stock data, adapt your tone accordingly, and provide a summary + investment verdict.
        Focus on Explainibility rather than only data
        - Use good spacing and give section wise content 
        -Please include relevant emojis to make the response more engaging.
            Use emojis like 📈 for growth, 💰 for profits, ⚠️ for risk, ✅ for strong fundamentals, 📉 for decline, and ⭐️ for final verdicts.
            """
            
        },
        *conversation_hist,
        {
                "role":"user",
            "content":user_prompt
    ,
        }
        ],
        model="llama3-70b-8192"
    ,
        tools=tools,
        tool_choice="auto",
        max_completion_tokens=500
    )
    #print(chat_completion.choices[0].message)
    message = chat_completion.choices[0].message
    if message.tool_calls is None:
        return message.content  # This is a true fallback — no tool was ever planned


    tool_outputs = []
    sentiment_added = False
    stock_price=False
    analyze_financials_added=False

    

    for tool_call in message.tool_calls:
        function_name = tool_call.function.name
        argument = json.loads(tool_call.function.arguments)
        print(function_name)
        symbol = argument.get('symbol')
        symbol=get_Full_ticker(symbol)
        print(symbol)
        if argument.get('symbol') and not sentiment_added:
            sentiment = sentiment_analysis(argument['symbol'])
            #print(sentiment)
            tool_outputs.append(f"📰 **Market Sentiment Analysis for {argument['symbol']}**:\n{sentiment}")
            sentiment_added=True

        
        if function_name=='get_stock_price' and not stock_price:
            result = get_stock_price(symbol)
            tool_outputs.append(f"📈 Stock Price Info:\n{result}")
            stock_price=True
            print(result)
            
        
            
        elif function_name=='analyze_sector':
            if argument.get('symbol'):
                sector_analysis=analyze_sector(argument['symbol'],"")
            
                prompt = f"""
                        📊 **Sector Financial Analysis**

                        Here is a summary of the financial performance of the top companies in this sector:

                        {sector_analysis}

                        ✅ Use this information to evaluate the overall health of the sector based on profitability, valuation, and growth metrics.

                        📌 If a stock ticker was provided, compare its performance with these top companies across key metrics like Free Cash Flow, P/E Ratio, and Revenue Growth like:
                                        
                            📌 Mention 1–2 quick insights at the end such as:
                            - Which company leads in revenue growth?
                            - Which has the strongest cash flow or best valuation?

                        """
                tool_outputs.append(prompt)
                sector_analysis_added = True

    
    if symbol!=None and not stock_price:
            result = get_stock_price(symbol)
            tool_outputs.append(f"📈 Stock Price Info:\n{result}")
            stock_price=True
            print(result)
            
    if symbol!=None and not analyze_financials_added:
        summary = analyze_financials(symbol)
        prompt = f"""
                Here is the financial summary of {argument['symbol']}:

                {summary}

                Give an investment attractiveness score out of 100 and explain briefly.
                - Use **bold headings** for sections like Stock Price,Sentiment Analysis,Financial Insights, Verdict, etc.
                "Use emojis like ✅ 📈 💰 🚀 🟢 for positive, ❌ ⚠️ 📉 🔻 🔴 for negative, and 📎 🟡 🤝 for neutral. Add them at the start of bullet points or section headers."
                - Use bullet points * or - for each fact or insight.
                - Add line breaks between paragraphs for better readability.
                -Please include relevant emojis to make the response more engaging.
                Use emojis like 📈 for growth, 💰 for profits, ⚠️ for risk, ✅ for strong fundamentals, 📉 for decline, and ⭐️ for final verdicts.            """
        tool_outputs.append(prompt)
    if message.tool_calls is None and "sector" in user_prompt.lower() and not sector_analysis_added:
        # Try to extract ticker from prompt
        from difflib import get_close_matches
    
        def extract_ticker_from_prompt(prompt):
            prompt = prompt.lower()
            company_names = list(company_ticker_map.keys())
            matches = get_close_matches(prompt, company_names, n=1, cutoff=0.6)
            if matches:
                return company_ticker_map[matches[0]]
            return None

        fallback_ticker = extract_ticker_from_prompt(user_prompt)
        if fallback_ticker:
            sector_analysis = analyze_sector(fallback_ticker, "")
            prompt = f"""
    📊 **Sector Financial Analysis (via {fallback_ticker})**

    Here is a summary of the financial performance of the top companies in this sector:

    {sector_analysis}

    ✅ Use this information to evaluate the overall health of the sector based on profitability, valuation, and growth metrics.

    📌 Compare this company with sector peers on:
    - Free Cash Flow
    - P/E Ratio
    - Revenue Growth
    - Net Income
    - Debt-to-Equity
    - Market Cap

    🎯 Format insights cleanly with a markdown table if possible.
    """
            tool_outputs.append(prompt)


    final_prompt = "\n\n".join(tool_outputs)

    followup_messages = [
        {
            "role":"system",
            "content": """
            You are a helpful financial assistant. Use clear, engaging, and structured formatting with relevant emojis.

            - Use emojis like ✅ 📈 💰 🚀 🟢 for positive, ❌ ⚠️ 📉 🔻 🔴 for negative, and 📎 🟡 🤝 for neutral.
            - Start section headings and bullet points with appropriate emojis.
           
            -Use a strict rule based scoring out of 100 using financial data:
                Reduce scoring as much as possible and be realistic
                25 for Profitability - Above 20 only if Net Profit is positive
                20 for Cash Flow
                15 for Valuation
                15 for Growth
                20 for Shareholder Returns
            - Add section headers like **Stock Price**, **Market Sentiment**, **Financials**, and **Final Verdict**.
            - Make your response not less than 300 words.
            - When doing sentiment analysis try to explain them in sentences rather than giving short verdicts
            -If relevant news not found skip it.
            """
        },
       
        {
            "role":"user",
            "content": final_prompt,
        }
    ]


    summary_response = client.chat.completions.create(
            model="llama3-70b-8192",
            messages=followup_messages,
            max_tokens=800
        )
    
    assistant_reply= summary_response.choices[0].message.content

    return assistant_reply

#print(run_finance_agent("Can you analyze TCS stock price",[]))
#print(sentiment_analysis("TCS"))