pages/Extractor.py

import streamlit as st
import fitz  # PyMuPDF
import pandas as pd
import re
import pdfplumber
from html_templates import Extracter_title, tooltip_message_extracter, logo, button_styles_fynder, upload_extracter_file, download_button_styles, extractor_placeholder



st.markdown(logo, unsafe_allow_html=True)
st.logo("alerter_4.jpeg")

def extract_text_from_pdf(uploaded_file):
    try:
        doc = fitz.open(stream=uploaded_file.read(), filetype="pdf")
        text = "\n".join([page.get_text("text") for page in doc])

        key_value_data = parse_text_as_table(text)
        table_data = extract_table_from_pdf(doc)

        return key_value_data, table_data
    except Exception as e:
        return pd.DataFrame({"Error": [f"Error extracting PDF text: {str(e)}"]}), None

def parse_text_as_table(text):
    """Extracts key-value pairs from structured text."""
    data = []
    lines = text.split("\n")
    for line in lines:
        if " - " in line:
            key, value = line.split(" - ", 1)
            data.append([key.strip(), value.strip()])
        elif ":" in line:
            key, value = line.split(":", 1)
            data.append([key.strip(), value.strip()])
    return pd.DataFrame(data, columns=["Field", "Value"])

def extract_table_from_pdf(doc):
    """Extracts tabular data using PyMuPDF."""
    table_data = []
    table_started = False

    for page in doc:
        blocks = page.get_text("blocks")  # Get structured text blocks
        blocks = sorted(blocks, key=lambda x: (x[1], x[0]))  # Sort by Y, then X

        for block in blocks:
            text = block[4].strip()
            if re.search(r"Item Name|SAC Code|Taxable Value", text, re.IGNORECASE):
                table_started = True  # Start extracting after headers
                continue  # Skip header line

            if table_started:
                columns = re.split(r"\s{2,}", text)  # Split by large spaces
                if len(columns) >= 5:  # Ensure it's a valid row
                    table_data.append(columns)

    if table_data:
        return pd.DataFrame(table_data, columns=["Item Name", "SAC Code", "Gross Amount", "Tax Type", "Tax Amount", "Total Value"])
    return None


def extract_first_table_row(uploaded_file):
    """Extracts only the first row of tabular data and reshapes it."""
    with pdfplumber.open(uploaded_file) as pdf:
        for page in pdf.pages:
            extracted_table = page.extract_table()
            if extracted_table:
                df = pd.DataFrame(extracted_table)
                df.columns = df.iloc[0]  # First row as column headers
                df = df[1:].reset_index(drop=True)  # Remove the first row (headers)

                if not df.empty:
                    first_row = df.iloc[0]  # Get only the first row
                    
                    # Reshape table into Column A (Headers) and Column B (Values)
                    reshaped_data = [[col, first_row[col]] for col in df.columns]
                    return pd.DataFrame(reshaped_data, columns=["Field", "Value"])
    
    return None

def main():
    st.markdown(Extracter_title, unsafe_allow_html = True)
    st.write("")
    st.markdown(tooltip_message_extracter, unsafe_allow_html = True)

    st.markdown(upload_extracter_file, unsafe_allow_html = True)
    uploaded_file = st.file_uploader("", type=["pdf"], accept_multiple_files=False)

    if uploaded_file:
        extracted_df, table_df = extract_text_from_pdf(uploaded_file)

        # Extract first row of table data
        first_row_df = extract_first_table_row(uploaded_file)

        # Combine key-value data and first row of table data
        if first_row_df is not None:
            combined_df = pd.concat([extracted_df, first_row_df], ignore_index=True)
        else:
            combined_df = extracted_df  # If no table row exists, use only key-value data

        # Display combined extracted data
        st.markdown(extractor_placeholder, unsafe_allow_html = True)
        st.dataframe(combined_df)

        # Allow user to download combined extracted data
        csv_combined = combined_df.to_csv(index=False).encode('utf-8')
        st.markdown(download_button_styles, unsafe_allow_html = True)
        st.download_button("Download", csv_combined, file_name="combined_extracted_data.csv", mime="text/csv")

        # Display extracted table data separately
        if table_df is not None:
            st.subheader("Extracted Table Data")
            st.dataframe(table_df)

            # Allow user to download extracted table data
            csv_table = table_df.to_csv(index=False).encode('utf-8')

            st.markdown(download_button_styles, unsafe_allow_html = True)
            st.download_button(csv_table, file_name="extracted_data.csv", mime="text/csv")

if __name__ == "__main__":
    main()