agent.py

from agentpro import AgentPro
from agentpro.tools import AresInternetTool, CodeEngine
import requests
from bs4 import BeautifulSoup
import PyPDF2
import io
import docx
from typing import Dict, List, Optional

class HiringAgent:
    def __init__(self):
        self.agent = AgentPro(tools=[AresInternetTool(), CodeEngine()])
        
    def extract_text_from_pdf(self, pdf_url: str) -> str:
        """Extract text from PDF file."""
        response = requests.get(pdf_url)
        pdf_file = io.BytesIO(response.content)
        pdf_reader = PyPDF2.PdfReader(pdf_file)
        text = ""
        for page in pdf_reader.pages:
            text += page.extract_text()
        return text
    
    def extract_text_from_docx(self, docx_url: str) -> str:
        """Extract text from DOCX file."""
        response = requests.get(docx_url)
        docx_file = io.BytesIO(response.content)
        doc = docx.Document(docx_file)
        text = ""
        for paragraph in doc.paragraphs:
            text += paragraph.text + "\n"
        return text
    
    def analyze_github_profile(self, github_url: str) -> Dict:
        """Analyze GitHub profile and extract relevant information."""
        response = requests.get(github_url)
        soup = BeautifulSoup(response.text, 'html.parser')
        
        # Extract basic information
        name = soup.find('span', {'class': 'p-name'}).text.strip() if soup.find('span', {'class': 'p-name'}) else ""
        bio = soup.find('div', {'class': 'p-note'}).text.strip() if soup.find('div', {'class': 'p-note'}) else ""
        
        # Extract repositories
        repos = []
        for repo in soup.find_all('a', {'data-hovercard-type': 'repository'})[:5]:
            repos.append({
                'name': repo.text.strip(),
                'url': f"https://github.com{repo['href']}"
            })
        
        return {
            'name': name,
            'bio': bio,
            'repositories': repos
        }
    
    def analyze_candidate(self, resume_url: str, github_url: str, job_description: str, company_info: str) -> Dict:
        """Analyze candidate profile and generate assessment."""
        # Extract resume text
        if resume_url.endswith('.pdf'):
            resume_text = self.extract_text_from_pdf(resume_url)
        elif resume_url.endswith('.docx'):
            resume_text = self.extract_text_from_docx(resume_url)
        else:
            resume_text = ""
        
        # Analyze GitHub profile
        github_data = self.analyze_github_profile(github_url)
        
        # Generate assessment using AgentPro
        prompt = f"""
        Analyze this candidate profile and provide a detailed assessment:
        
        Resume Content:
        {resume_text}
        
        GitHub Profile:
        Name: {github_data['name']}
        Bio: {github_data['bio']}
        Top Repositories: {[repo['name'] for repo in github_data['repositories']]}
        
        Job Description:
        {job_description}
        
        Company Information:
        {company_info}
        
        Please provide (Generate 40 tokens max per point/topic):
        1. Skills and experience match with job requirements
        2. Technical proficiency assessment
        3. Cultural fit analysis
        4. Strengths and areas for development
        5. Overall recommendations
        """
        
        assessment = self.agent(prompt)
        
        return {
            'resume_analysis': resume_text,
            'github_analysis': github_data,
            'assessment': assessment
        }