#!/usr/bin/env python3
"""按 GitHub API 的文件清单，用 raw 直取仓库文件（免 clone，应对慢镜像）。"""
import json, time, urllib.request
from pathlib import Path
ROOT = Path(__file__).resolve().parent
REPOS = ['ziniao-open/skills', 'ziniao-open/skill-hub', 'hubstudio-Max/hubstudio-skill']
UA = {'User-Agent': 'Mozilla/5.0'}
ok = fail = 0
for r in REPOS:
    name = r.split('/')[-1]
    tree = json.load(open(ROOT / f'tree-{name}.json'))
    for t in tree.get('tree', []):
        if t['type'] != 'blob':
            continue
        dest = ROOT / 'repos' / name / t['path']
        if dest.exists() and dest.stat().st_size > 0:
            ok += 1
            continue
        dest.parent.mkdir(parents=True, exist_ok=True)
        for br in ('main', 'master'):
            url = f"https://raw.githubusercontent.com/{r}/{br}/{t['path']}"
            try:
                data = urllib.request.urlopen(urllib.request.Request(url, headers=UA), timeout=30).read()
                dest.write_bytes(data); ok += 1
                break
            except Exception:  # noqa: BLE001
                time.sleep(0.5)
        else:
            fail += 1
            print('  ❌', r, t['path'], flush=True)
        time.sleep(0.1)
    print(f'  ✅ {r} 完成', flush=True)
print(f'仓库文件: ok={ok} fail={fail}', flush=True)