tools/attachment_loader.py

from pathlib import Path

import requests

from config import CACHE_DIR, DEFAULT_API_URL, REQUEST_HEADERS
from tools.types import SolverResult


def local_attachment_candidates(task_id: str, file_name: str) -> list[Path]:
    return [
        Path("data") / "attachments" / task_id / file_name,
        Path("data") / "attachments" / file_name,
        Path("resource") / file_name,
        Path(file_name),
    ]


def remote_attachment_candidates(task_id: str, file_name: str) -> list[str]:
    return [
        f"{DEFAULT_API_URL}/files/{task_id}",
        f"{DEFAULT_API_URL}/files/{file_name}",
        f"{DEFAULT_API_URL}/files/{task_id}/{file_name}",
        f"https://huggingface.co/datasets/asteriadyt/2023/resolve/main/validation/{file_name}",
        f"https://huggingface.co/spaces/jitendra217/Final_Assignment/resolve/main/resource/{file_name}",
    ]


def download_task_file(task_id: str, file_name: str) -> tuple[Path | None, str]:
    """下载/定位课程附件，并缓存在 /tmp 下。"""
    if not task_id or not file_name:
        return None, "没有附件。"

    for local_path in local_attachment_candidates(task_id, file_name):
        if local_path.exists() and local_path.stat().st_size > 0:
            return local_path, f"使用本地附件：{local_path}"

    destination = CACHE_DIR / task_id / file_name
    if destination.exists() and destination.stat().st_size > 0:
        return destination, f"使用缓存附件：{destination}"

    destination.parent.mkdir(parents=True, exist_ok=True)
    errors = []
    for url in remote_attachment_candidates(task_id, file_name):
        try:
            response = requests.get(url, headers=REQUEST_HEADERS, timeout=45)
            if response.status_code == 404:
                errors.append(f"{url} -> 404")
                continue
            response.raise_for_status()
            if response.content:
                destination.write_bytes(response.content)
                return destination, f"已下载附件：{url}"
        except Exception as exc:
            errors.append(f"{url} -> {exc}")

    return None, "附件下载失败：" + " | ".join(errors[:5])