core/parser/task.py

from core.models.parser import ApplicantResume
from core.parser.file_parser import BaseParser
from core.parser.prompt import generate_context_prompt, generate_identity_prompt
from core.resources.azure_openai import generate_response
import json

def extractor_task(file_name: str):
    # Parse all data in it's raw format.
    parser = BaseParser(file_name)
    raw_txt = parser.parse_pdf()
    messages = generate_messages(raw_txt)

    # generate response and put it in a pydantic model that validates what gpt has given us
    # it will also serve as our baseline on which data we can access from the resume
    # if it has enough fields for us to do an context similarity on, we can generate a proper report.
    # The model will also serve in our reporting as we can raise this as an issue if they do not have
    # enough context for us to work on. that means it is either it is not a complete resume.
    response = generate_response(messages)
    json_response = json.loads(response)
    model_response = ApplicantResume(**json_response)

    return json_response


def generate_messages(raw_txt: str):
    print("raw_txt in generate_messages", raw_txt)
    messages = []

    identity_prompt = generate_identity_prompt()
    context_prompt = generate_context_prompt(raw_txt)
    
    messages.append({"role": "system", "content": identity_prompt})
    messages.append({"role": "system", "content": context_prompt})

    return messages

def list_dict_to_str_parser(items):
    result = ""
    for item in items:
        for val in item.dict().values():
            result += f"{val}\n"
    
    return result