app.py

from flask import Flask, request, jsonify, render_template
import os
import random
import requests
from datetime import datetime
from haystack.document_stores import FAISSDocumentStore
from haystack.nodes import DensePassageRetriever, FARMReader
from haystack.pipelines import ExtractiveQAPipeline
from learning_data import learning_resources
from quiz_data import quiz_questions

# Initialize Flask App
app = Flask(__name__, template_folder='templates')

# API key for stock market data
API_KEY = "0L6PHZNXGSSIZAN9"
BASE_URL = "https://www.alphavantage.co/query"

# Fun financial tips
fun_facts = [
    "The 50/30/20 rule suggests allocating 50% of income to needs, 30% to wants, and 20% to savings!",
    "Start investing early to take advantage of compound interest!",
    "Avoid high-interest debt whenever possible.",
    "Automate your savings to ensure consistent contributions.",
    "Diversify your portfolio to minimize risk.",
]

# **STEP 1: Initialize FAISS Document Store**
if os.path.exists("faiss_document_store.db"):
    os.remove("faiss_document_store.db")
if os.path.exists("faiss_index"):
    os.remove("faiss_index")

document_store = FAISSDocumentStore(
    embedding_dim=768, 
    sql_url="sqlite:///faiss_document_store.db",
    faiss_index_factory_str="Flat"
)

# **STEP 2: Load Financial Knowledge into FAISS**
data = [
    {"content": "Dollar-cost averaging is an investment strategy where you invest a fixed dollar amount regularly, regardless of market conditions. It helps reduce risk by spreading purchases over time."},
    {"content": "A Roth IRA is a retirement savings account that allows tax-free withdrawals for qualified expenses. Contributions are made with after-tax dollars, which means your money grows tax-free."},
    {"content": "The 50/30/20 rule suggests allocating 50% of income to needs, 30% to wants, and 20% to savings. This budgeting method helps balance spending and saving effectively."},
    {"content": "High-yield savings accounts offer better interest rates compared to traditional savings accounts. These accounts are great for emergency funds and short-term savings."},
    {"content": "Index funds are low-cost investment funds that track a market index such as the S&P 500. They provide diversification and require minimal management, making them ideal for long-term investing."},
    {"content": "Investing in ETFs (Exchange-Traded Funds) is a great way to diversify your portfolio while maintaining flexibility. ETFs trade like stocks but offer diversification similar to mutual funds."},
    {"content": "Bonds are fixed-income investments where investors lend money to entities in exchange for periodic interest payments and the return of the principal amount at maturity."},
    {"content": "Mutual funds pool money from many investors to invest in a diversified portfolio of stocks, bonds, or other assets. They are managed by professionals."},
    {"content": "A credit score is a measure of creditworthiness, ranging from 300-850. A higher score improves your chances of getting better interest rates on loans and credit cards."},
    {"content": "Diversification reduces investment risk by spreading assets across different asset classes, such as stocks, bonds, and real estate."},
    {"content": "Compounding interest allows your savings and investments to grow exponentially over time. The earlier you start, the more significant the impact of compounding."},
    {"content": "An emergency fund should cover 3-6 months of living expenses. It should be kept in a liquid, high-yield savings account for quick access."},
    {"content": "Stock market volatility refers to the frequent rise and fall of stock prices. Long-term investors should focus on overall market trends rather than short-term fluctuations."},
    {"content": "A 401(k) is an employer-sponsored retirement savings plan that allows pre-tax contributions, reducing taxable income while saving for retirement."},
    {"content": "Annuities are financial products that provide regular payments over a specified period, often used for retirement income."},
    # New Accessibility & Literacy Content
    {"content": "Predatory lending refers to unethical practices by lenders that hinder a borrower's ability to repay a loan. Examples include payday loans with astronomical interest rates. Avoid these by seeking credit unions or community banks."},
    {"content": "Secured credit cards are a great way to build credit if you have no credit history or a low score. You provide a cash deposit that acts as your credit limit, and paying on time builds your score."},
    {"content": "Credit unions are member-owned financial cooperatives that often offer lower fees and better interest rates than traditional banks, making them a good option for accessible banking."},
    {"content": "To open a bank account without a traditional ID, look for banks that accept alternative forms of identification like consular ID cards or municipal IDs, or check 'Bank On' certified accounts."},
    {"content": "Payday loans are short-term, high-cost loans that can trap borrowers in a cycle of debt due to extremely high annual percentage rates (APRs). Alternatives include personal installment loans or borrowing from family."},
    {"content": "Building credit with low income is possible by using a secured credit card, becoming an authorized user on a trusted person's card, or using rent-reporting services to have on-time rent payments count toward your score."}
]

document_store.write_documents(data, duplicate_documents="skip")

# **STEP 3: Initialize Retriever & Update FAISS Embeddings**
retriever = DensePassageRetriever(
    document_store=document_store,
    query_embedding_model="facebook/dpr-question_encoder-single-nq-base",
    passage_embedding_model="facebook/dpr-ctx_encoder-single-nq-base",
    use_gpu=False
)

print("Updating FAISS document embeddings...")
document_store.update_embeddings(retriever)
print("FAISS embeddings updated successfully!")

# **STEP 4: Initialize Reader & QA Pipeline**
reader = FARMReader(model_name_or_path="deepset/roberta-large-squad2", use_gpu=False, max_seq_len=512, doc_stride=128)
qa_pipeline_rag = ExtractiveQAPipeline(reader, retriever)

# **STEP 5: Fetch Market Data Function**
def fetch_market_data():
    indices = {"S&P 500": "SPY", "Dow Jones": "DIA", "NASDAQ": "QQQ"}
    market_data = {}
    try:
        for name, symbol in indices.items():
            response = requests.get(BASE_URL, params={
                "function": "TIME_SERIES_INTRADAY",
                "symbol": symbol,
                "interval": "5min",
                "apikey": API_KEY
            })
            data = response.json()
            if "Time Series (5min)" in data:
                latest_time = sorted(data["Time Series (5min)"].keys())[-1]
                latest_close = data["Time Series (5min)"][latest_time]["4. close"]
                market_data[name] = latest_close
            else:
                market_data[name] = "N/A"
    except Exception as e:
        print(f"Error fetching market data: {e}")
    return market_data

def fetch_top_stocks():
    try:
        response = requests.get(BASE_URL, params={
            "function": "TOP_GAINERS_LOSERS",
            "apikey": API_KEY
        })
        data = response.json()
        if "top_gainers" in data:
            return data["top_gainers"][:5]
        return []
    except Exception as e:
        print(f"Error fetching top stocks: {e}")
        return []

# **STEP 6: Serve Frontend UI**
@app.route("/")
def home():
    return render_template("index.html")

# **STEP 7: API Endpoints**
@app.route("/api/stocks/top5", methods=["GET"])
def get_top_stocks():
    return jsonify(fetch_top_stocks())

@app.route("/api/learning", methods=["GET"])
def get_learning_resources():
    return jsonify(learning_resources)

@app.route("/api/quiz", methods=["GET"])
def get_quiz():
    # Return a random selection of 3 questions or all of them
    return jsonify(random.sample(quiz_questions, min(len(quiz_questions), 5)))

@app.route("/chat", methods=["POST"])
def chat():
    user_input = request.json.get("message", "").lower()

    # Handle chatbot commands
    if user_input == "help":
        return jsonify({"response": """
        🤖 **Chatbot Commands:**
        1️⃣ **Ask finance questions** (e.g., "How do I save more money?")
        2️⃣ **Get live stock market updates** (Type 'market')
        3️⃣ **Receive a financial tip** (Type 'tip')
        """})
    elif user_input == "tip":
        return jsonify({"response": random.choice(fun_facts)})
    elif "market" in user_input:
        return jsonify({"response": fetch_market_data()})

    # Run FAISS query
    try:
        result = qa_pipeline_rag.run(query=user_input, params={"Retriever": {"top_k": 20}, "Reader": {"top_k": 5}})
        
        retrieved_answers = [ans.answer for ans in result.get("answers", [])]
        
        if retrieved_answers:
            detailed_answers = list(set([ans.strip() for ans in retrieved_answers if ans and len(ans) > 20]))

            if len(detailed_answers) > 1:
                full_response = "\n\n".join(detailed_answers[:3])
            elif len(detailed_answers) == 1:
                full_response = detailed_answers[0]
            else:
                full_response = "I couldn't find a clear answer. Try rewording your question."

            return jsonify({"response": full_response})
    except Exception as e:
        print(f"Error in RAG pipeline: {e}")

    return jsonify({"response": "I'm not sure, but you can try rewording your question."})

# **Run Flask App**
if __name__ == "__main__":
    from waitress import serve
    # Hugging Face Spaces defaults to port 7860
    serve(app, host="0.0.0.0", port=7860)