import streamlit as st from utils import * import uuid #Creating session variables if 'unique_id' not in st.session_state: st.session_state['unique_id'] ='' def main(): st.set_page_config(page_title="Resume Screening Assistance") st.title("HR - Resume Screening Assistance...πŸ’ ") st.subheader("I can help you in resume screening process") job_description = st.text_area("Please paste the 'JOB DESCRIPTION' here...",key="1") document_count = st.text_input("No.of 'RESUMES' to return",key="2") # Upload the Resumes (pdf files) pdf = st.file_uploader("Upload resumes here, only PDF files allowed", type=["pdf"],accept_multiple_files=True) submit=st.button("Help me with the analysis") if submit: with st.spinner('Wait for it...'): #Creating a unique ID, so that we can use to query and get only the user uploaded documents from PINECONE vector store st.session_state['unique_id']=uuid.uuid4().hex #Create a documents list out of all the user uploaded pdf files final_docs_list=create_docs(pdf,st.session_state['unique_id']) #st.write(final_docs_list) #Displaying the count of resumes that have been uploaded st.write("*Resumes uploaded* :"+str(len(final_docs_list))) #Create embeddings instance embeddings=create_embeddings_load_data() #Introducing a line separator st.write(":heavy_minus_sign:" * 30) #For each item in relavant docs - we are displaying some info of it on the UI for item in range(len(relavant_docs)): st.subheader("πŸ‘‰ "+str(item+1)) #Displaying Filepath st.write("**File** : "+relavant_docs[item][0].metadata['name']) #Introducing Expander feature with st.expander('Show me πŸ‘€'): st.info("**Match Score** : "+ str(1 - relavant_docs[item][1])) #st.write("***"+relavant_docs[item][0].page_content) #Gets the summary of the current item using 'get_summary' function that we have created which uses LLM & Langchain chain summary = get_summary(relavant_docs[item][0]) st.write("**Summary** : "+summary) st.success("Hope I was able to save your time❀️") #Invoking main function if __name__ == '__main__': main()