New recruiting agent

app.py

@@ -0,0 +1,242 @@
+import gradio as gr
+from langchain.document_loaders import PyPDFLoader, UnstructuredFileLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.embeddings import HuggingFaceEmbeddings
+from langchain.vectorstores import Chroma
+from langchain.chains import LLMChain
+from langchain_groq import ChatGroq
+from typing import List, Dict
+import os
+import tempfile
+
+# Initialize embeddings and vector store
+embeddings = HuggingFaceEmbeddings()
+vector_store = Chroma(embedding_function=embeddings, persist_directory="./chroma_db")
+
+# Initialize LLM
+llm = ChatGroq(
+    api_key=os.environ["GROQ_API_KEY"],
+    model_name="llama-3.1-8b-instant"
+)
+
+def process_candidate_submission(resume_file, job_description: str) -> str:
+    # Load and process resume
+    if resume_file.name.endswith('.pdf'):
+        loader = PyPDFLoader(resume_file.name)
+    else:
+        loader = UnstructuredFileLoader(resume_file.name)
+    
+    resume_doc = loader.load()[0]
+    
+    # Create prompt for cold email generation
+    prompt_template = """
+    Given the following resume and job description, create a professional cold email:
+    
+    Resume:
+    {resume_text}
+    
+    Job Description:
+    {job_description}
+    
+    Generate a concise, compelling cold email that highlights the candidate's relevant skills and experience.
+    """
+    
+    chain = LLMChain(
+        llm=llm,
+        prompt=prompt_template
+    )
+    
+    response = chain.run(
+        resume_text=resume_doc.page_content,
+        job_description=job_description
+    )
+    
+    return response
+
+def store_culture_docs(culture_files: List[tempfile._TemporaryFileWrapper]) -> str:
+    """Store company culture documentation in the vector store"""
+    text_splitter = RecursiveCharacterTextSplitter(
+        chunk_size=1000,
+        chunk_overlap=200
+    )
+    
+    all_docs = []
+    for file in culture_files:
+        if file.name.endswith('.pdf'):
+            loader = PyPDFLoader(file.name)
+        else:
+            loader = UnstructuredFileLoader(file.name)
+        docs = loader.load()
+        splits = text_splitter.split_documents(docs)
+        all_docs.extend(splits)
+    
+    vector_store.add_documents(all_docs, collection_name="culture_docs")
+    return f"Successfully stored {len(all_docs)} culture document chunks"
+
+def store_resumes(resume_files: List[tempfile._TemporaryFileWrapper]) -> str:
+    """Store resumes in the vector store"""
+    text_splitter = RecursiveCharacterTextSplitter(
+        chunk_size=1000,
+        chunk_overlap=200
+    )
+    
+    all_docs = []
+    for file in resume_files:
+        if file.name.endswith('.pdf'):
+            loader = PyPDFLoader(file.name)
+        else:
+            loader = UnstructuredFileLoader(file.name)
+        docs = loader.load()
+        splits = text_splitter.split_documents(docs)
+        all_docs.extend(splits)
+    
+    vector_store.add_documents(all_docs, collection_name="resumes")
+    return f"Successfully stored {len(all_docs)} resume chunks"
+
+def analyze_candidates(job_description: str) -> str:
+    """Analyze candidates against job description and culture fit"""
+    
+    # Extract skills from job description
+    skills_prompt = """
+    Extract the key technical skills and requirements from this job description:
+    
+    {job_description}
+    
+    Return the skills as a comma-separated list.
+    """
+    
+    skills_chain = LLMChain(
+        llm=llm,
+        prompt=skills_prompt
+    )
+    
+    skills = skills_chain.run(job_description=job_description)
+    
+    # Query vector store for matching resumes
+    results = vector_store.similarity_search(
+        job_description,
+        k=5,
+        collection_name="resumes"
+    )
+    
+    # Get culture documentation
+    culture_docs = vector_store.similarity_search(
+        job_description,
+        k=3,
+        collection_name="culture_docs"
+    )
+    
+    # Analysis prompt
+    analysis_prompt = """
+    Analyze these candidates for the job position and culture fit.
+    
+    Job Description:
+    {job_description}
+    
+    Required Skills:
+    {skills}
+    
+    Company Culture Context:
+    {culture_docs}
+    
+    Candidate Resumes:
+    {resumes}
+    
+    For each candidate, provide:
+    1. Skills match (percentage)
+    2. Culture fit assessment
+    3. Recommendation (move forward/reject)
+    4. Brief explanation
+    """
+    
+    analysis_chain = LLMChain(
+        llm=llm,
+        prompt=analysis_prompt
+    )
+    
+    analysis = analysis_chain.run(
+        job_description=job_description,
+        skills=skills,
+        culture_docs="\n".join([doc.page_content for doc in culture_docs]),
+        resumes="\n".join([doc.page_content for doc in results])
+    )
+    
+    return analysis
+
+def create_interface():
+    with gr.Blocks() as app:
+        gr.Markdown("# AI Recruiter Assistant")
+        
+        with gr.Tabs():
+            # Candidate View
+            with gr.Tab("Candidate View"):
+                with gr.Row():
+                    resume_upload = gr.File(label="Upload Resume")
+                    job_desc_input = gr.Textbox(
+                        label="Paste Job Description",
+                        lines=10
+                    )
+                generate_btn = gr.Button("Generate Cold Email")
+                email_output = gr.Textbox(
+                    label="Generated Cold Email",
+                    lines=10
+                )
+                
+                generate_btn.click(
+                    process_candidate_submission,
+                    inputs=[resume_upload, job_desc_input],
+                    outputs=email_output
+                )
+            
+            # Recruiter View
+            with gr.Tab("Recruiter View"):
+                with gr.Row():
+                    culture_docs_upload = gr.File(
+                        label="Upload Company Culture Documents",
+                        file_count="multiple"
+                    )
+                    store_culture_btn = gr.Button("Store Culture Docs")
+                    culture_status = gr.Textbox(label="Status")
+                
+                with gr.Row():
+                    resume_bulk_upload = gr.File(
+                        label="Upload Resumes",
+                        file_count="multiple"
+                    )
+                    store_resumes_btn = gr.Button("Store Resumes")
+                    resume_status = gr.Textbox(label="Status")
+                
+                with gr.Row():
+                    job_desc_recruiter = gr.Textbox(
+                        label="Paste Job Description",
+                        lines=10
+                    )
+                    analyze_btn = gr.Button("Analyze Candidates")
+                    analysis_output = gr.Textbox(
+                        label="Analysis Results",
+                        lines=20
+                    )
+                
+                store_culture_btn.click(
+                    store_culture_docs,
+                    inputs=culture_docs_upload,
+                    outputs=culture_status
+                )
+                
+                store_resumes_btn.click(
+                    store_resumes,
+                    inputs=resume_bulk_upload,
+                    outputs=resume_status
+                )
+                
+                analyze_btn.click(
+                    analyze_candidates,
+                    inputs=job_desc_recruiter,
+                    outputs=analysis_output
+                )
+    
+    return app
+
+if __name__ == "__main__":
+    app = create_interface()
+    app.launch()

requirements.txt

@@ -0,0 +1,18 @@
+langchain
+langchain-community
+langchain-groq
+chromadb
+sentence-transformers
+gradio
+unstructured
+pdf2image
+python-magic
+pdfminer.six
+nltk
+transformers
+torch
+numpy
+Pillow
+pypdf
+python-docx
+unstructured[pdf]