app.py

import os
import streamlit as st
from paddleocr import PaddleOCR
import cv2
from langchain.chains import LLMChain
from langchain_core.prompts import ChatPromptTemplate
from langchain_google_genai import ChatGoogleGenerativeAI
from dotenv import load_dotenv
from sqlalchemy import create_engine, Column, Integer, String, JSON
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import sessionmaker
import google.generativeai as genai

# Load environment variables
load_dotenv()

# Set up environment variables
api_key = os.getenv('API_KEY')
DATABASE_URL = "sqlite:///mydatabase.db"

# Setup database
Base = declarative_base()

class MyDataModel(Base):
    __tablename__ = 'my_data_table'
    
    id = Column(Integer, primary_key=True)
    name = Column(String)
    data = Column(JSON)

engine = create_engine(DATABASE_URL)
Session = sessionmaker(bind=engine)
session = Session()

Base.metadata.create_all(engine)

# Initialize Google Generative AI API
genai.configure(api_key=api_key)

# Define OCR function using PaddleOCR
def ocr_with_paddle(img_path):
    finaltext = ''
    ocr = PaddleOCR(lang='en', use_angle_cls=True)

    
    img = cv2.imread(img_path)

   
    result = ocr.ocr(img)

 
    for line in result[0]:  
        for word_info in line:
            
            if isinstance(word_info[1], list):
                text = word_info[1][0] 
                text=str(text) 
                finaltext += text + ' '  
            else:
               
               finaltext += str(word_info)+' '

    return finaltext.strip() 

# Define the prompt template for extracting invoice details
prompt = ChatPromptTemplate.from_messages(
    [
        ("system", "You are a helpful assistant that extracts invoice details such as invoice number, customer name, date, amount, and other relevant information from a provided invoice text."),
        ("human", "{input}"),
    ]
)

llm = ChatGoogleGenerativeAI(
    model="gemini-1.5-pro",
    temperature=0.5,
    max_tokens=None,
    timeout=None,
    max_retries=2,
    api_key=api_key
)

invoice_chain = LLMChain(prompt=prompt, llm=llm)

def extract_invoice_details(input_text):
    response = invoice_chain({"input": input_text})
    extracted_details = response["text"].strip()
    return extracted_details

# Streamlit UI
st.title("Invoice OCR and Details Extraction")

st.write(
    "Upload an image file to extract the text and invoice details such as invoice number, customer name, date, and amount."
)

# Image Upload
uploaded_image = st.file_uploader("Choose an Image", type=["jpg", "jpeg", "png"])

if uploaded_image is not None:
    # Save uploaded image to a temporary file
    img_path = "temp_image.png"
    with open(img_path, "wb") as f:
        f.write(uploaded_image.getbuffer())
    
    # Perform OCR on the uploaded image
    text = ocr_with_paddle(img_path)
    st.write("Extracted Text:")
    st.text_area("OCR Output", text, height=300)
    
    # Extract invoice details from the text
    invoice_details = extract_invoice_details(text)
    
    st.write("Extracted Invoice Details:")
    st.text_area("Invoice Details", invoice_details, height=300)
    
    # Save details to the database
    new_entry = MyDataModel(name="invoice_details", data=invoice_details)
    session.add(new_entry)
    session.commit()
    session.close()
    
    st.success("Invoice details saved to the database!")