Added cover letter generator module

app.py

@@ -2,11 +2,13 @@
     This is the main file of the app. This file contains the code for the streamlit app.
 '''
 
+import time
 import streamlit as st
 from streamlit_chat import message
 
 from job_description_generator import predict_job_description, get_job_description_conversation
 from job_description_fixer import fix_job_description, get_job_description_fixer_conversation
+from cover_letter_generator import get_cover_letter
 
 conversation = get_job_description_conversation()
 if 'generator_conversation' not in st.session_state:
@@ -67,13 +69,14 @@ def main():
     st.sidebar.markdown("---")
     #selector
     page = st.sidebar.selectbox(
-        "Select a page", ["Home", "Job Description Generator", "Job Description Fixer"])
+        "Select a page", ["Home", "Job Description Generator", "Job Description Fixer", "Cover Letter Generator"])
     if page == "Home":
         st.title("JobGPT")
         st.write("Select a page in the sidebar to get started.")
         st.write("### Available options:")
         st.write("1. Job Description Generator")
         st.write("2. Job Description Fixer")
+        st.write("3. Cover Letter Generator")
         st.markdown("---")
 
     elif page == "Job Description Generator":
@@ -107,7 +110,7 @@ def main():
             message_writer(input_text, response)
     elif page == "Job Description Fixer":
         container_two = st.container()
-        container_two.title("JobGPT: A Job Description Fixing Chatbot")
+        container_two.title("A Job Description Fixing Chatbot")
         container_two.markdown(
             "JobGPT is a chatbot that fixes job descriptions. This is built just for demo purpose."
         )
@@ -134,4 +137,41 @@ def main():
                 input_text, st.session_state['fixer_conversation'])
             message_writer(input_text, response)
 
+    elif page == "Cover Letter Generator":
+        container_three = st.container()
+        container_three.title("A Cover Letter Generating Chatbot")
+        container_three.markdown( "JobGPT is a chatbot that generates cover letters. \
+                                 This is built just for demo purpose.")
+        container_three.markdown("---")
+        #upload file
+        uploaded_file = container_three.file_uploader("Upload your resume", type=["pdf"])
+        if uploaded_file is not None:
+            #store it as a pdf
+            with open("resume.pdf", "wb") as file_io:
+                file_io.write(uploaded_file.getbuffer())
+           
+            #show a loading bar
+            with st.spinner('Uploading...'):
+                time.sleep(1)
+            container_three.success('Uploaded!')
+            container_three.markdown("---")
+            #add a form to get title, company and more info from the user
+            form = container_three.form(key='my_form')
+            title = form.text_input("Job Title (required)", placeholder="VP of Engineering")
+            company = form.text_input("Company Name (required)", placeholder="Google")
+            more_info = form.text_area("More Info",
+                                       help="Add more info about you or the job in natural language",
+                                       placeholder="I am a software engineer with 5 years of experience. The job focuses on building a new product in healthcare sector.")
+            submit_button = form.form_submit_button(label='Submit')
+            if submit_button:
+                if title == "":
+                    st.error("Please enter a job title")
+                elif company == "":
+                    st.error("Please enter a company name")
+                else:
+                    cover_letter = get_cover_letter(title, company, more_info, "resume.pdf")
+                    container_three.markdown("---")
+                    container_three.markdown("### Cover Letter:")
+                    container_three.write(cover_letter)
+
 main()

cover_letter_generator.py

@@ -0,0 +1,23 @@
+'''
+    This module generates cover letter from resume
+'''
+
+from loaders import get_chain_for_pdf
+
+def generate_query(title:str, company:str, more_info = None):
+    '''
+    Generate a query from a title and company
+    '''
+    query = f"Write a professional and well-formatted cover letter based on the given resume for the job of {title} at {company}.\n"
+    if more_info:
+        query += f"More info: {more_info}\n"
+    return query
+
+def get_cover_letter(title:str, company:str, more_info = None, resume_path = "resume.pdf"):
+    '''
+    Generate a cover letter from a title and company
+    '''
+    query = generate_query(title, company, more_info)
+    chain = get_chain_for_pdf(resume_path)
+    response = chain({"query": query})
+    return response['result']

loaders.py

@@ -0,0 +1,59 @@
+'''
+    This module contains all the loaders
+'''
+
+import os
+from langchain.document_loaders import PyPDFLoader
+from langchain.text_splitter import CharacterTextSplitter
+from langchain.embeddings import OpenAIEmbeddings
+from langchain.vectorstores import Chroma
+from langchain.chains import RetrievalQA
+from langchain.chat_models import ChatOpenAI
+
+openai_api_key=os.environ['OPENAI_API_KEY']
+
+def load_pdf(path: str = "resume.pdf"):
+    '''
+    Load a pdf file from a stringio object
+    '''
+    pdf_loader = PyPDFLoader(path)
+    documents = pdf_loader.load()
+    return documents
+    
+def get_embeddings(documents):
+    '''
+    Get embeddings from a list of documents
+    '''
+    splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
+    texts = splitter.split_documents(documents)
+    embeddings = OpenAIEmbeddings(openai_api_key=openai_api_key)
+    return texts, embeddings
+
+def get_db(texts, embeddings):
+    '''
+    Get a vectorstore from a list of texts and embeddings
+    '''
+    db = Chroma.from_documents(texts, embeddings)
+    return db
+
+def get_retriever(db):
+    '''
+    Get a retriever from a vectorstore
+    '''
+    retriever = db.as_retriever(search_type="similarity", search_kwargs={"k":1})
+    return retriever
+
+def get_chain_for_pdf(path):
+    '''
+    Get a conversation chain from a path
+    '''
+    documents = load_pdf(path)
+    texts, embeddings = get_embeddings(documents)
+    db = get_db(texts, embeddings)
+    retriever = get_retriever(db)
+    chain = RetrievalQA.from_chain_type(
+    llm=ChatOpenAI(temperature=0, openai_api_key=openai_api_key),
+        chain_type="stuff",
+        retriever=retriever,
+        return_source_documents=True)
+    return chain

requirements.txt

@@ -1,4 +1,7 @@
-langchain==0.0.194
+langchain==0.0.198
 openai==0.27.8
 streamlit==1.23.1
 streamlit_chat==0.0.2.2
+pypdf==3.9.1
+chromadb==0.3.26
+tiktoken==0.4.0