app.py

import os
import streamlit as st
import fitz
import sqlite3
import pdfplumber
from langchain_community.vectorstores import FAISS
from langchain_openai import OpenAIEmbeddings
from langchain.text_splitter import RecursiveCharacterTextSplitter  # Now works

@st.cache_resource
def init_system():
    process_pdf("Q1FY24.pdf")
    embeddings = OpenAIEmbeddings(openai_api_key=os.getenv("OPENAI_API_KEY"))
    vector_store = FAISS.load_local("faiss_index", embeddings)
    conn = sqlite3.connect('metric_table.db')
    return vector_store, conn

    
def process_pdf(pdf_path):
    # Structured Data Extraction
    conn = sqlite3.connect('metric_table.db')
    cursor = conn.cursor()
    cursor.execute('''CREATE TABLE IF NOT EXISTS metric_table 
                    (metric TEXT, quarter TEXT, value REAL)''')

    # Example metric insertion (add full extraction logic)
    cursor.execute("INSERT INTO metric_table VALUES ('Revenue', 'Q1 FY24', 19.8)")
    conn.commit()

    # Unstructured Data Processing
    full_text = ""
    with fitz.open(pdf_path) as doc:
        for page in doc:
            full_text += page.get_text()

    # Text Chunking & Embedding
    splitter = RecursiveCharacterTextSplitter(chunk_size=1000)
    chunks = splitter.split_text(full_text)
    embeddings = OpenAIEmbeddings(
        openai_api_key=os.getenv("OPENAI_API_KEY"),
        model="text-embedding-ada-002"
    )
    FAISS.from_texts(chunks, embeddings).save_local("faiss_index")

# Streamlit UI
def main():
    st.title("Fundrev Financial Analyzer")
    
    # Initialize system
    vector_store, conn = init_system()
    
    query = st.text_input("Ask financial question:")
    
    if query:
        # Structured data queries
        if any(kw in query.lower() for kw in ["trend", "margin", "revenue"]):
            cursor = conn.cursor()
            cursor.execute(f"SELECT * FROM metric_table WHERE metric LIKE '%{query}%'")
            results = cursor.fetchall()
            st.table(results if results else "No matching metrics found")
        # Unstructured data queries
        else:
            docs = vector_store.similarity_search(query, k=1)
            st.write(docs[0].page_content if docs else "No relevant information found")

if __name__ == "__main__":
    main()