app.py

import streamlit as st

from pdf_to_image import pdf_to_image
from image_to_text import image_to_text
from mirascope_extractor import extractor

import google.generativeai as genai
import pandas as pd

import glob
import os
# from dotenv import load_dotenv
import streamlit as st

from resume_recommender import recommend_resume



# load_dotenv()

global empty_df
openai_api_key = os.environ['OPENAI_API_KEY']

print('This is API key: ', openai_api_key)
genai.configure(api_key=openai_api_key)


st.set_page_config(page_title="CV Recommendor Gen AI")
st.title("Gen AI CV Recommender")
st.write("""
**Find the Best Candidates for Your Job!**

🔍 **What It Does:**
- Analyzes multiple resumes in PDF format.
- Matches them with your job description.

📋 **How It Works:**
1. Upload resumes 📄
2. Upload the job description 📝
3. Get a table of top recommended candidates 🏆

Ready to find the perfect fit? Let's get started!
""")

uploaded_files = st.sidebar.file_uploader("Choose PDF files", accept_multiple_files=True, type="pdf")
job_description = st.sidebar.text_input('Enter Job Description')

if uploaded_files and job_description:
    if st.sidebar.button('AI Recommendation'):
        image_bytes = pdf_to_image(uploaded_files)
        
        all_texts = []
        for image_byte in image_bytes:
            print('This is image_byte: ', image_byte)
            
            combine_text = ''
            for image in image_byte:
                text = image_to_text(image)
                combine_text += text
            print('This is the text from single PDF: ', combine_text)
            all_texts.append(combine_text)    
            
        empty_df = pd.DataFrame()
        
        for text in all_texts:
            extracted_text = extractor(text)
            task_details_dict = extracted_text.dict()
            df = pd.DataFrame([task_details_dict])
            empty_df = pd.concat([empty_df, df])
            

        recommend_df = recommend_resume(empty_df, job_description)
        if 'Unnamed: 0' in recommend_df.columns:
            recommend_df = recommend_df.drop('Unnamed: 0', axis=1)

        print('Recommendation process done successfully')
            
        st.write(recommend_df)
        # csv = empty_df.to_csv(index=False)
        # st.download_button(
        #     label = 'Click to Download CSV',
        #     data = csv,
        #     file_name = 'Extracted_data.csv',
        #     mime='text/csv',
        # )