app.py

from pathlib import Path
import streamlit as st
import os
import requests

from form16_parser import build_parser, UnsupportedForm16Error

badges = """[![](https://img.shields.io/static/v1?label=Licence&message=MIT&color=darkgreen)](https://github.com/INF800/form16-parser)
[![stability-alpha](https://img.shields.io/badge/stability-alpha-f4d03f.svg)](https://github.com/mkenney/software-guides/blob/master/STABILITY-BADGES.md#alpha)
[![](https://img.shields.io/static/v1?label=Python&message=3.10+&color=indigo)](https://github.com/INF800/form16-parser) 
[![im](https://img.shields.io/github/stars/INF800/form16-parser.svg?style=social&)]()
"""

def main():
    st.title("Form16 Parser")

    st.markdown(f"{badges}")
    st.write("Source code available @ https://github.com/INF800/form16-parser")
    uploaded_file = st.file_uploader("Upload a PDF file", type="pdf")
    Path("temp_files").mkdir(exist_ok=True, parents=True)

    if uploaded_file is None:
        # If no file is uploaded, load the default PDF file
        default_pdf_url = "https://assets1.cleartax-cdn.com/cleartax/images/1655725194_sampleform16.pdf"
        path = Path("temp_files/sampleform16.pdf")
        # if not path.exists():
        default_pdf_response = requests.get(default_pdf_url)

        with open(str(path), "wb") as f:
            f.write(default_pdf_response.content)

        st.write(f"Parsing {default_pdf_url}.")
        st.write("Try your own form 16 by uploading above.")

        parser = build_parser()
        parsed = parser.parse(str(path), return_output=True)

        st.write("Parsed data (structured):")
        st.write(parsed)
    else:
        # Save the uploaded file to a temporary directory
        with open(os.path.join("temp_files", uploaded_file.name), "wb") as f:
            f.write(uploaded_file.getvalue())

        filepath = os.path.join("temp_files", uploaded_file.name)        
        parser = build_parser()

        try:
            parsed = parser.parse(filepath, return_output=True)
            st.write("Parsed data (structured):")
            st.write(parsed)
        except UnsupportedForm16Error as e:
            st.write(e)


        

if __name__ == "__main__":
    main()