메뉴
从 Wikimedia Commons 抓取自由许可的真实图片素材(产品/物品/场景)。用于 U 盘素材准备、文档配图等需要真实照片但无 API key 的场景。
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Set up Alembic database migrations for FastAPI projects using async SQLAlchemy (asyncpg). Covers env.py configuration with async engine, autogenerate handling, greenfield vs brownfield migration strategies, and env var integration.
使用 CloakBrowser 绕过抖音反爬，获取真实视频 MP4 下载链接
Wan2.2 FLF2V (First/Last Frame to Video) 管线 — 生成首尾帧之间的插值视频，支持电影级镜头过渡
Use CloakBrowser (anti-detection Chromium) to access protected Chinese web platforms (Douyin/抖音, etc.) and extract structured data — intercepting API responses, extracting video/media URLs, and downloading content from sites that block standard Playwright/browser automation.
Install and diagnose Playwright-based packages (Playwright, CloakBrowser, etc.) on Linux/WSL — handles externally-managed Python, system deps, proxy-based binary downloads, and runtime troubleshooting
Exploratory QA of web apps: find bugs, evidence, reports.
| name | wikimedia-image-fetch |
| description | 从 Wikimedia Commons 抓取自由许可的真实图片素材(产品/物品/场景)。用于 U 盘素材准备、文档配图等需要真实照片但无 API key 的场景。 |
无需 API key,无需登录,所有图片 CC 自由许可可直接商用。 适用:考试素材、产品图配图、博客配图等需要"真实图片"而非占位的场景。
用:需要真实物体照片、品牌原型、场景图,且没有 API key。 不用:需要矢量图 / 矢量 logo(用 Wikimedia 后缀会搜到 SVG 但渲染要 librsvg);或需要 Unsplash 风格的高质量摄影(用 source.unsplash.com 需 key)。
import requests
PROXY = "http://127.0.0.1:7890" # WSL 下走 Clash 代理
HEADERS = {"User-Agent": "HermesBot/1.0 (your@email.com)"} # 真实 UA,默认 python-requests 会被拦
r = requests.get(
"https://commons.wikimedia.org/w/api.php",
params={
"action": "query",
"list": "search",
"srnamespace": 6, # File: 命名空间
"srsearch": "your keywords",
"srlimit": 15,
"format": "json",
},
headers=HEADERS,
proxies={"http": PROXY, "https": PROXY},
timeout=15,
)
hits = r.json()["query"]["search"] # [{"title": "File:XXX.jpg", ...}, ...]
r = requests.get(
"https://commons.wikimedia.org/w/api.php",
params={
"action": "query",
"titles": "File:XXX.jpg",
"prop": "imageinfo",
"iiprop": "url",
"iiurlwidth": 800, # 缩略图宽度,无则原图
"format": "json",
},
headers=HEADERS, proxies={"http": PROXY, "https": PROXY}, timeout=15,
)
url = r.json()["query"]["pages"][page_id]["imageinfo"][0]["thumburl"]
srsearch 跨所有 File 类型,PDF/.svg/.webm/.ogv 都会出来。必须过滤:
SKIP_EXT = (".pdf", ".svg", ".webm", ".ogv")
for h in hits:
title = h["title"]
if any(title.lower().endswith(ext) for ext in SKIP_EXT):
continue
return title # 第一张可用
否则会下到 PDF 封面而不是产品图(实测:bedside lamp 搜到 The Spirit lamp (serial) (IA spiritlampserial00doug).pdf)。
/api/rest_v1/page/summary/<title> 返回的 thumbnail.source 经常不准确(Apple Watch Ultra 2 搜到 Series 3 图)。直接用 search + imageinfo 两步法别图省事用 summary。
upload.wikimedia.org 限流。实测 sleep 60s 后恢复正常。代码里用 time.sleep(60) 重试一次即可。
见 scripts/fetch_wikimedia_images.py。
"Sony WH-1000XM4" 通常有图,但拼写要准"modern lamp" / "side-by-side refrigerator" 命中率高"product image"(太泛) / 纯品牌名(可能只搜到 logo SVG)from PIL import Image
img = Image.open(local_path)
pixels = list(img.getdata())
assert len(set(pixels[:5000])) > 10, "可能是纯色占位"
unique pixel 数 < 10 = 纯色块,> 100 = 真图。