Files
WonderQ-Project/WonderQ-MiniAPP/scripts/download-assets.mjs
2026-08-11 19:19:11 +08:00

103 lines
3.6 KiB
JavaScript

import crypto from "node:crypto";
import fs from "node:fs/promises";
import path from "node:path";
const root = process.cwd();
const extractPath = path.join(root, "docs", "research", "source-mobile-extract.json");
const assetDir = path.join(root, "public", "assets", "source");
const manifestPath = path.join(assetDir, "manifest.json");
function extractCssUrls(value) {
if (!value) return [];
return [...String(value).matchAll(/url\((?:"|')?([^"')]+)(?:"|')?\)/g)].map((match) => match[1]);
}
function extensionFromContentType(contentType) {
if (contentType.includes("webp")) return ".webp";
if (contentType.includes("png")) return ".png";
if (contentType.includes("gif")) return ".gif";
if (contentType.includes("svg")) return ".svg";
if (contentType.includes("jpeg") || contentType.includes("jpg")) return ".jpg";
return ".bin";
}
function extensionFromUrl(url) {
try {
const parsed = new URL(url);
const ext = path.extname(parsed.pathname).toLowerCase();
if ([".jpg", ".jpeg", ".png", ".webp", ".gif", ".svg", ".avif"].includes(ext)) {
return ext === ".jpeg" ? ".jpg" : ext;
}
} catch {
return "";
}
return "";
}
const data = JSON.parse(await fs.readFile(extractPath, "utf8"));
const rawUrls = [
...data.extract.images.map((item) => item.src),
...data.extract.images.map((item) => item.src?.replace(/&/g, "&")),
...data.extract.backgroundImages.flatMap((item) => extractCssUrls(item.backgroundImage)),
...data.extract.links.filter((item) => item.as === "image").map((item) => item.href),
].filter(Boolean);
const uniqueUrls = [...new Set(rawUrls)]
.filter((url) => /^https?:\/\//.test(url))
.filter((url) => !url.includes("data:image"));
await fs.mkdir(assetDir, { recursive: true });
const manifest = [];
let cursor = 0;
async function downloadOne(originalUrl, index) {
const hash = crypto.createHash("sha1").update(originalUrl).digest("hex").slice(0, 10);
try {
const response = await fetch(originalUrl, {
headers: {
"User-Agent":
"Mozilla/5.0 (iPhone; CPU iPhone OS 17_5 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/17.5 Mobile/15E148 Safari/604.1",
Referer: "https://m.ctrip.com/",
},
});
if (!response.ok) {
manifest.push({ originalUrl, ok: false, status: response.status });
return;
}
const contentType = response.headers.get("content-type") ?? "";
const ext = extensionFromUrl(originalUrl) || extensionFromContentType(contentType);
const filename = `asset-${String(index).padStart(3, "0")}-${hash}${ext}`;
const filePath = path.join(assetDir, filename);
const arrayBuffer = await response.arrayBuffer();
await fs.writeFile(filePath, Buffer.from(arrayBuffer));
manifest.push({
originalUrl,
ok: true,
status: response.status,
contentType,
bytes: arrayBuffer.byteLength,
filename,
publicPath: `/assets/source/${filename}`,
});
} catch (error) {
manifest.push({ originalUrl, ok: false, error: error instanceof Error ? error.message : String(error) });
}
}
async function worker() {
while (cursor < uniqueUrls.length) {
const index = cursor;
cursor += 1;
await downloadOne(uniqueUrls[index], index);
}
}
await Promise.all(Array.from({ length: 5 }, () => worker()));
manifest.sort((a, b) => (a.filename ?? "").localeCompare(b.filename ?? ""));
await fs.writeFile(manifestPath, JSON.stringify({ generatedAt: new Date().toISOString(), assets: manifest }, null, 2));
const ok = manifest.filter((item) => item.ok).length;
const failed = manifest.length - ok;
console.log(`Downloaded ${ok}/${manifest.length} assets. Failed: ${failed}.`);