升级

scripts/export_article_covers.py

@@ -0,0 +1,64 @@
+import hashlib
+import json
+from pathlib import Path
+
+from PIL import Image, ImageOps
+
+
+ROOT = Path(__file__).resolve().parents[1]
+CACHE = ROOT / ".tmp" / "gzhh-images"
+RAW = CACHE / "raw"
+MANIFEST = CACHE / "manifest.json"
+ARTICLES = ROOT / "src" / "data" / "articles.generated.json"
+OUTPUT = ROOT / "public" / "assets" / "article-covers"
+
+manifest = json.loads(MANIFEST.read_text(encoding="utf-8"))
+by_url = {
+    item["url"]: item
+    for item in manifest
+    if item.get("status") == "ok" and item.get("file")
+}
+articles = json.loads(ARTICLES.read_text(encoding="utf-8"))
+OUTPUT.mkdir(parents=True, exist_ok=True)
+
+exported = 0
+missing = []
+used_files = set()
+
+for article in articles:
+    source_url = article.get("coverSourceUrl") or article.get("cover", "")
+    if not source_url.startswith("http"):
+        continue
+    cache_entry = by_url.get(source_url)
+    if not cache_entry:
+        missing.append({"title": article.get("title", ""), "url": source_url})
+        continue
+
+    key = hashlib.sha256(source_url.encode("utf-8")).hexdigest()[:12]
+    output_name = f"cover-{key}.webp"
+    output_path = OUTPUT / output_name
+    with Image.open(RAW / cache_entry["file"]) as image:
+        frame = ImageOps.exif_transpose(image).convert("RGB")
+        frame.thumbnail((960, 960), Image.Resampling.LANCZOS)
+        frame.save(output_path, "WEBP", quality=82, method=6)
+
+    article["coverSourceUrl"] = source_url
+    article["cover"] = f"/assets/article-covers/{output_name}"
+    used_files.add(output_name)
+    exported += 1
+
+ARTICLES.write_text(
+    json.dumps(articles, ensure_ascii=False, indent=2) + "\n",
+    encoding="utf-8",
+)
+
+summary = {
+    "articles": len(articles),
+    "exported": exported,
+    "uniqueFiles": len(used_files),
+    "missing": len(missing),
+    "bytes": sum((OUTPUT / name).stat().st_size for name in used_files),
+}
+print(json.dumps(summary, ensure_ascii=False))
+if missing:
+    raise RuntimeError(f"Missing {len(missing)} cached cover images")

scripts/import-gzhh-articles.mjs

@@ -1,3 +1,4 @@
+import { createHash } from "node:crypto";
 import { readdir, readFile, writeFile } from "node:fs/promises";
 import { fileURLToPath } from "node:url";
 import path from "node:path";
@@ -14,6 +15,12 @@ const cleanMarkdown = (value) =>
     .replace(/\s+/g, " ")
     .trim();
 
+const getLocalCoverPath = (url) => {
+  if (!url) return "";
+  const key = createHash("sha256").update(url).digest("hex").slice(0, 12);
+  return `/assets/article-covers/cover-${key}.webp`;
+};
+
 const getExcerpt = (body, title) => {
   const ignored = /(?:data:image|%3Csvg|transform=|阅读|点赞|分享|推荐|留言|2024届612|font-family|__bottom-bar__|sns_opr_btn|picture_content|page_content)/i;
   const lines = body
@@ -60,16 +67,18 @@ const parseArticle = (fileName, raw) => {
   const contentStart = raw.search(/>\s*原文地址:/);
   const body = contentStart >= 0 ? raw.slice(contentStart).split(/\r?\n/).slice(2).join("\n") : raw;
   const images = [...body.matchAll(/!\[[^\]]*\]\((https?:\/\/[^\s\)]+)\)/g)]
-    .map((match) => match[1])
+    .map((match) => match[1].replaceAll("&", "&"))
     .filter((url) => /mmbiz\.qpic\.cn/.test(url));
   const excerptText = getExcerpt(body, title);
+  const coverSourceUrl = images[0] ?? "";
 
   return {
     title,
     date,
     displayDate: date ? date.replaceAll("-", ".") : "日期待补",
     sourceUrl,
-    cover: images[0] ?? "",
+    cover: getLocalCoverPath(coverSourceUrl),
+    coverSourceUrl,
     excerpt: excerptText.length > 88 ? `${excerptText.slice(0, 88)}…` : excerptText || "一页从公众号找回的班级记录。",
     category: getCategory(title),
     fileName