utils.py

from typing import Dict, List, Tuple
import re
import tempfile
from pathlib import Path
import pandas as pd
import requests
from pandas import DataFrame

# --- Constants ---
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
QUESTIONS_URL = f"{DEFAULT_API_URL}/questions"
SUBMIT_URL = f"{DEFAULT_API_URL}/submit"
FILE_PATH = f"{DEFAULT_API_URL}/files/"


def process_file(task_id: str, question_text: str) -> str:
    """
    Attempt to download a file associated with a task from the API.
    - If the file exists (HTTP 200), it is saved to a temp directory and the local file path is returned.
    - If no file is found (HTTP 404), returns None.
    - For all other HTTP errors, the exception is propagated to the caller.
    """
    file_url = f"{FILE_PATH}{task_id}"

    try:
        response = requests.get(file_url, timeout=30)
        response.raise_for_status()
    except requests.exceptions.RequestException as exc:
        print(f"Exception in download_file>> {str(exc)}")
        return question_text # Unable to get the file

    # Determine filename from 'Content-Disposition' header, fallback to task_id
    content_disposition = response.headers.get("content-disposition", "")
    filename = task_id
    match = re.search(r'filename="([^"]+)"', content_disposition)
    if match:
        filename = match.group(1)

    # Save file in a temp directory
    temp_storage_dir = Path(tempfile.gettempdir()) / "gaia_cached_files"
    temp_storage_dir.mkdir(parents=True, exist_ok=True)

    file_path = temp_storage_dir / filename
    file_path.write_bytes(response.content)
    return (
                f"{question_text}\n\n"
                f"---\n"
                f"A file was downloaded for this task and saved locally at:\n"
                f"{str(file_path)}\n"
                f"---\n\n"
            )