data_processor.py

import os
from langchain_community.document_loaders import PyPDFLoader

async def load_pages(file_dir):
    pages = []
    for file_path in os.listdir(file_dir):
        file_path = os.path.join(file_dir, file_path)
        if not file_path.endswith(".pdf"):
            continue

        loader = PyPDFLoader(file_path)
        async for page in loader.alazy_load():
            pages.append(page)
    return pages



async def process_claim(agent, claim):
    response = await agent.ainvoke({"claim": claim, "mistake_message": ""})
    return {
        "claim": claim,
        "match_source": [
            {
                "document_name": detail.document_name,
                "matching_text": detail.matching_text
            }
            for detail in response["answer"].details
        ]
    }