services/agents/ProfileMatchingAgent.py

import asyncio
import os, sys
import time
from dotenv import load_dotenv
load_dotenv()
from typing import AsyncIterable
from typing import List, Optional, Dict
from langchain.schema import HumanMessage, AIMessage
from langchain.callbacks import AsyncIteratorCallbackHandler

# from fastapi.middleware.cors import CORSMiddleware
# from fastapi.responses import StreamingResponse
# from langchain.chat_models import ChatOpenAI
# from pydantic import BaseModel
# from langchain_google_genai import ChatGoogleGenerativeAI
from models.data_model import OutProfile, OutMatching
from services.llms.LLM import model_gemini, model_4o, model_4omini, model_5mini
from services.agents.LLMAgent import LLMAgent
from services.prompts.profile_matching import match_one_profile
from utils.utils import pdf_reader, ingest_one_profile, ingest_bulk_profile



async def helper_profile_match_prompt(profile: OutProfile) -> str:
    profile.high_edu_major_1 = profile.high_edu_major_1 if profile.high_edu_major_1 else "-"
    profile.high_edu_univ_1 = profile.high_edu_univ_1 if profile.high_edu_univ_1 else "-"
    profile.high_edu_gpa_1 = profile.high_edu_gpa_1 if profile.high_edu_gpa_1 else "-"
    profile.high_edu_major_2 = profile.high_edu_major_2 if profile.high_edu_major_2 else "-"
    profile.high_edu_univ_2 = profile.high_edu_univ_2 if profile.high_edu_univ_2 else "-"
    profile.high_edu_gpa_2 = profile.high_edu_gpa_2 if profile.high_edu_gpa_2 else "-"
    profile.high_edu_major_3 = profile.high_edu_major_3 if profile.high_edu_major_3 else "-"
    profile.high_edu_univ_3 = profile.high_edu_univ_3 if profile.high_edu_univ_3 else "-"
    profile.high_edu_gpa_3 = profile.high_edu_gpa_3 if profile.high_edu_gpa_3 else "-"
    profile.hardskills = profile.hardskills if profile.hardskills else []
    profile.softskills = profile.softskills if profile.softskills else []
    profile.certifications = profile.certifications if profile.certifications else []
    profile.business_domain_experiences = profile.business_domain_experiences if profile.business_domain_experiences else []
    
    profile_text = f"""
# Candidate Profile

**Education**
- Bachelor degree major: {profile.high_edu_major_1 or "-"}
- Bachelor university: {profile.high_edu_univ_1 or "-"}
- Bachelor GPA: {profile.high_edu_gpa_1 or "-"}

- Master degree major: {profile.high_edu_major_2 or "-"}
- Master university: {profile.high_edu_univ_2 or "-"}
- Master GPA: {profile.high_edu_gpa_2 or "-"}

- Doctor/Phd degree major: {profile.high_edu_major_3 or "-"}
- Doctor/Phd university: {profile.high_edu_univ_3 or "-"}
- Doctor/Phd GPA: {profile.high_edu_gpa_3 or "-"}

**Year of Working Experience**
{profile.yoe} year

**Hardskills**
{", ".join([p for p in profile.hardskills if p not in ['null', '']])}

**Softskills**
{", ".join([p for p in profile.softskills if p not in ['null', '']])}

**Certifications**
{", ".join([p for p in profile.certifications if p not in ['null', '']])}

**Business domain experiences**
{", ".join([p for p in profile.business_domain_experiences if p not in ['null', '']])}
""".strip()
    return profile_text
    

class ProfileMatchingAgent(LLMAgent):
    def __init__(self, model=model_5mini):
        super().__init__(model)
        self.agent_name = "ProfileMatchingAgent"
        self.prompt_template = match_one_profile

    async def _helper_matching_one(self, profile_id:int, profile_dict:Dict, profile_text:str, criteria:str) -> Dict:
        llm = self.model.with_structured_output(OutMatching)
        chain = self.prompt_template | llm
        input_chain = {
            "profile_text":profile_text,
            "criteria":criteria,
        }
        matched_profile = await chain.ainvoke(input_chain, config=None)
        print(f"_helper_matching_one/matched_profile:\n{matched_profile}")
        
        # matched_profile = matched_profile.model_dump()
        # matched_profile["profile_id"] = profile_id
        
        profile_dict["score"] = matched_profile.score
        profile_dict["reason"] = matched_profile.reason
        profile_dict["profile_id"] = profile_id

        return profile_dict
    
    async def matching_profile_bulk(self, profiles:List[OutProfile], criteria:str) -> Optional[List]:
        try:
            st = time.time()
            tasks = []
            profiles_text = []
            profiles_dict = []
            n_profiles = len(profiles)

            for profile_id, p in enumerate(profiles):
                print(f"Processing [{profile_id+1}/{n_profiles}]")
                profile_text = await helper_profile_match_prompt(p)
                profile_dict = p.model_dump()
                profiles_dict.append(profile_dict)
                profiles_text.append(profile_text)

                task = asyncio.create_task(self._helper_matching_one(profile_id=profile_id, 
                                                    profile_dict=profile_dict, 
                                                    profile_text=profile_text, 
                                                    criteria=criteria))
                tasks.append(task)
            
            profiles_with_scores = await asyncio.gather(*tasks)

            print(f"✅ Profile matching finished in {(time.time() - st)//60} min, {(time.time() - st)%60} sec")
            return profiles_with_scores
        except Exception as E:
            error_message = f"Processing profile matching error: {E}"
            print(error_message)
            exc_type, exc_obj, exc_tb = sys.exc_info()
            fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)[1]
            print(exc_type, fname, exc_tb.tb_lineno)