@xingwangzhe/stalux 1.25.5 → 1.25.7

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@xingwangzhe/stalux",
-    "version": "1.25.5",
+    "version": "1.25.7",
     "description": "A powerful, modern Astro blog theme — use as template or install as plugin",
     "keywords": [
         "astro",
@@ -75,8 +75,11 @@
         "prepare": "git config core.hooksPath scripts/git-hooks"
     },
     "dependencies": {
+        "@astrojs/markdown-satteri": "0.3.7",
         "@astrojs/rss": "^4.0.19",
         "@astrojs/sitemap": "^3.7.3",
+        "@echogarden/icu-segmentation-wasm": "^0.4.0",
+        "@echogarden/text-segmentation": "^0.7.0",
         "@expressive-code/plugin-line-numbers": "^0.44.1",
         "@lucide/astro": "^1.33.0",
         "@mcp-b/webmcp-polyfill": "^4.0.0",
@@ -90,6 +93,7 @@
         "astro-seo": "^1.1.0",
         "astro-typewriter": "^1.2.0",
         "badge-maker": "^6.0.0",
+        "js-tokens": "^10.0.0",
         "pagefind": "^1.5.2",
         "photoswipe": "^5.4.4",
         "satteri": "^0.10.5",

package/src/components/stalux/posts/postContent.astro

@@ -12,7 +12,6 @@ interface Props {
     transitionName?: string;
     date?: string;
     updated?: string;
-    readingTime?: string;
     readingMinutes?: number;
     wordCount?: number;
     hasImage?: boolean;
@@ -62,13 +61,13 @@ const mdSourceUrl = exportMd
         </div>
             {
                 <div class={style.pageStats}>
-                    {(Boolean(props.readingTime) || (props.readingMinutes ?? 0) > 0) && (
+                    {(props.readingMinutes ?? 0) > 0 && (
                         <span class={style.readingTime}>
                             {t("post.readingTime")}{props.readingMinutes && props.readingMinutes > 0
                                 ? (props.readingMinutes < 1
                                     ? t("common.lessThanOneMinute")
                                     : t("common.minute", { n: props.readingMinutes }))
-                                : props.readingTime}
+                                : ""}
                         </span>
                     )}
                     {(props.wordCount ?? 0) > 0 && (

package/src/layouts/PostLayout.astro

@@ -14,7 +14,6 @@ interface Props {
     title: string;
     date?: string;
     updated?: string;
-    readingTime?: string;
     readingMinutes?: number;
     wordCount?: number;
     tags?: string[];
@@ -57,7 +56,6 @@ const walineConfig = walineSection && walineSection.serverURL ? {
             title={props.title}
             date={props.date}
             updated={props.updated}
-            readingTime={props.readingTime}
             readingMinutes={props.readingMinutes}
             wordCount={props.wordCount}
             hasImage={props.hasImage}

package/src/pages/about.astro

@@ -37,7 +37,6 @@ const author = authorConfig.name;
             transitionName="about-content"
             date={undefined}
             updated={undefined}
-            readingTime={undefined}
         >
             <Content />
             <PostCC

package/src/pages/api/posts.json.ts

@@ -1,3 +1,4 @@
+import { analyzeFeatureFlags } from "@plugins/feature-flags";
 import type { APIRoute } from "astro";
 import { getCollection } from "astro:content";
 
@@ -12,17 +13,19 @@ export const prerender = true;
 export const GET: APIRoute = async () => {
     const posts = await getCollection("posts", ({ data }) => !data.draft);
 
-    const payload = posts.map((post) => ({
-        title: post.data.title,
-        abbrlink: String(post.data.abbrlink),
-        date: post.data.date ?? undefined,
-        updated: post.data.updated ?? undefined,
-        tags: post.data.tags ?? [],
-        categories: post.data.categories ?? [],
-        desc: post.data.desc ?? "",
-        wordCount: post.data.wordCount ?? undefined,
-        url: `/posts/${post.data.abbrlink}/`,
-    }));
+    const payload = await Promise.all(
+        posts.map(async (post) => ({
+            title: post.data.title,
+            abbrlink: String(post.data.abbrlink),
+            date: post.data.date ?? undefined,
+            updated: post.data.updated ?? undefined,
+            tags: post.data.tags ?? [],
+            categories: post.data.categories ?? [],
+            desc: post.data.desc ?? "",
+            wordCount: (await analyzeFeatureFlags(post.body)).wordCount,
+            url: `/posts/${post.data.abbrlink}/`,
+        })),
+    );
 
     return new Response(JSON.stringify(payload), {
         status: 200,

package/src/pages/posts/[post].astro

@@ -2,6 +2,7 @@
 import Stalux from "@layouts/Stalux.astro";
 import PostLayout from "@layouts/PostLayout.astro";
 import { getCollection, type CollectionEntry, render } from "astro:content";
+import { analyzeFeatureFlags } from "@plugins/feature-flags";
 import { toTimestamp } from "@utils/dayjs";
 import Navs from "@components/stalux/layout/navs.astro";
 import { getSiteData, getAuthorData, getHeadData, getNavsData, getLinksData, getCommentData, getFooterData, getPromoteData, getTypetextsData, getMediaLinksData } from "@utils/config-utils";
@@ -14,19 +15,11 @@ const configEntries = await getCollection("config");
 const siteConfig = getSiteData(configEntries);
 const authorConfig = getAuthorData(configEntries);
 const props = Astro.props as Props;
-const { Content, headings, remarkPluginFrontmatter } = await render(props.post);
+const { Content, headings } = await render(props.post);
 
-// 从 remarkPluginFrontmatter 读取 feature flags 和字数统计（由 satteri 插件在构建时写入）
-const flags = {
-    hasImage: remarkPluginFrontmatter?.hasImage ?? false,
-};
-
-// 字数统计直接从 frontmatter 读取
-const minutesRead = (remarkPluginFrontmatter?.minutesRead as string) ?? "";
-const readingMinutes = (remarkPluginFrontmatter?.readingMinutes as number) ?? 0;
-const wordCount = (remarkPluginFrontmatter?.wordCount as number) ?? 0;
-// desc 可能为空字符串，需用 || 做“空串回退”，不能用 ??（?? 会把 "" 当作有效值）
-const autoDesc = props.post.data.desc || (remarkPluginFrontmatter?.desc as string) || "";
+// 文章统计完全由 Sätteri AST 分析，不读取 render metadata。
+const flags = await analyzeFeatureFlags(props.post.body);
+const autoDesc = props.post.data.desc || "";
 
 export async function getStaticPaths() {
     const posts = await getCollection("posts", ({ data }) => !data.draft);
@@ -71,7 +64,7 @@ export async function getStaticPaths() {
 >
     <div style="height:3em"></div>
     <Navs /><div style="height:3em"></div>
-        <PostLayout title={props.post.data.title} date={props.post.data.date} updated={props.post.data.updated} readingTime={minutesRead} readingMinutes={readingMinutes} tags={props.post.data.tags||[]} categories={props.post.data.categories||[]} headings={headings} author={authorConfig.name} authorAvatar={authorConfig.avatar} prev={props.prev} next={props.next} cc={props.post.data.cc} wordCount={wordCount} hasImage={flags.hasImage}>
+        <PostLayout title={props.post.data.title} date={props.post.data.date} updated={props.post.data.updated} readingMinutes={flags.readingMinutes} tags={props.post.data.tags||[]} categories={props.post.data.categories||[]} headings={headings} author={authorConfig.name} authorAvatar={authorConfig.avatar} prev={props.prev} next={props.next} cc={props.post.data.cc} wordCount={flags.wordCount} hasImage={flags.hasImage}>
         <Content/>
     </PostLayout>
 </Stalux>

package/src/plugins/feature-flags.ts

@@ -1,9 +1,12 @@
+import { createSatteriMarkdownProcessor } from "@astrojs/markdown-satteri";
+import { splitToWords } from "@echogarden/text-segmentation";
+import jsTokens from "js-tokens";
 /**
  * Sätteri 插件：在构建时完成字数统计和特性标记，
- * 通过 ctx.data.astro.frontmatter 把最终结果传递到 frontmatter，运行时无需再算。
+ * 只从 Sätteri 的 MDAST/HAST 节点读取信息，并通过 ctx 注入最终结果。
  *
  * 字数统计策略：
- * - 只数 heading 和 paragraph 的内容（用 ctx.textContent 获取纯文本）。
+ * - 从 root 递归收集正文文本，覆盖列表、引用、表格等嵌套块。
  * - 独立的 code 块、math/displayMath 公式不参与计数
  *   （代码是"看"不是"读"，符号表达式不按词计）。
  * - 行内 code 和行内 math 会被 textContent(paragraph) 自然包含。
@@ -15,34 +18,84 @@
  */
 import { defineMdastPlugin, defineHastPlugin } from "satteri";
 
-import { countWords } from "../utils/count-words.ts";
-
 const WORDS_PER_MINUTE = 400;
+const CODE_SECONDS_PER_NON_EMPTY_LINE = 2;
+
+type FeatureFlagsState = {
+    proseText: string;
+    codeTokens: number;
+    codeCharacters: number;
+    codeNonEmptyLines: number;
+    codeBlocks: number;
+    hasImage: boolean;
+};
+
+function getState(ctx: any): FeatureFlagsState {
+    const data = ctx.data as Record<string, unknown>;
+    return (data.staluxFeatureFlags ??= {
+        proseText: "",
+        codeTokens: 0,
+        codeCharacters: 0,
+        codeNonEmptyLines: 0,
+        codeBlocks: 0,
+        hasImage: false,
+    }) as FeatureFlagsState;
+}
+
+function injectState(ctx: any) {
+    const state = getState(ctx);
+    const injected = ctx.data.astro.frontmatter as Record<string, unknown>;
+    injected.proseText = state.proseText;
+    injected.codeTokens = state.codeTokens;
+    injected.codeCharacters = state.codeCharacters;
+    injected.codeNonEmptyLines = state.codeNonEmptyLines;
+    injected.codeBlocks = state.codeBlocks;
+    injected.hasImage = state.hasImage;
+}
 
-/** 从纯文本统计字数并累加到 frontmatter。 */
-function flushWordCount(ctx: any, text: string) {
-    const prev = (ctx.data.astro.frontmatter.wordCount as number) ?? 0;
-    const wc = prev + countWords(text);
-    const fm = ctx.data.astro.frontmatter;
-    fm.wordCount = wc;
-    fm.readingMinutes = Math.ceil(wc / WORDS_PER_MINUTE);
-    // 搜索全文：拼接所有 heading + paragraph 的纯文本
-    fm.searchText = ((fm.searchText as string) || "") + " " + text;
+function countCodeTokens(value: string, lang?: string): number {
+    if (/^(?:js|jsx|javascript|ts|tsx|typescript)$/.test(lang ?? "")) {
+        return Array.from(jsTokens(value, { jsx: /jsx|tsx/.test(lang ?? "") })).filter(
+            (token) =>
+                ![
+                    "WhiteSpace",
+                    "LineTerminatorSequence",
+                    "MultiLineComment",
+                    "SingleLineComment",
+                    "HashbangComment",
+                ].includes(token.type),
+        ).length;
+    }
+    return value.match(/\p{L}[\p{L}\p{N}_]*|\p{N}+(?:\.\p{N}+)?|[^\s]/gu)?.length ?? 0;
 }
 
 export const featureFlagsMdast = defineMdastPlugin({
     name: "feature-flags-mdast",
 
-    code(_node, _ctx) {
-        // 独立的 code 块不参与字数统计
+    code(node, ctx) {
+        const state = getState(ctx);
+        state.codeBlocks++;
+        state.codeCharacters += [...node.value].length;
+        state.codeNonEmptyLines += node.value
+            .split(/\r?\n/u)
+            .filter((line) => /\S/u.test(line)).length;
+        state.codeTokens += countCodeTokens(node.value, node.lang);
+        injectState(ctx);
     },
 
     heading(node, ctx) {
-        flushWordCount(ctx, ctx.textContent(node));
+        getState(ctx).proseText += ` ${ctx.textContent(node)}`;
+        injectState(ctx);
     },
 
     paragraph(node, ctx) {
-        flushWordCount(ctx, ctx.textContent(node));
+        getState(ctx).proseText += ` ${ctx.textContent(node)}`;
+        injectState(ctx);
+    },
+
+    tableCell(node, ctx) {
+        getState(ctx).proseText += ` ${ctx.textContent(node)}`;
+        injectState(ctx);
     },
 });
 
@@ -55,7 +108,9 @@ export const featureFlagsHast = defineHastPlugin({
     element: {
         filter: ["img"],
         visit(node, ctx) {
-            ctx.data.astro.frontmatter.hasImage = true;
+            const state = getState(ctx);
+            state.hasImage = true;
+            (ctx.data.astro.frontmatter as Record<string, unknown>).hasImage = true;
             // 第一张正文图片可能成为 LCP，优先加载；其余图片延迟加载。
             const data = ctx.data as Record<string, unknown>;
             const firstImage = !data.staluxFirstImageSeen;
@@ -72,3 +127,32 @@ export const featureFlagsHast = defineHastPlugin({
         },
     },
 });
+
+/**
+ * 直接用 Sätteri AST 分析一篇文章，供文章页、全局统计和 API 复用。
+ * 这里不复用内容集合的 data，也不读取 Astro 的渲染 metadata。
+ */
+export async function analyzeFeatureFlags(body: string | undefined): Promise<{
+    hasImage: boolean;
+    readingMinutes: number;
+    wordCount: number;
+}> {
+    const processor = await createSatteriMarkdownProcessor({
+        mdastPlugins: [featureFlagsMdast],
+        hastPlugins: [featureFlagsHast],
+    });
+    const result = await processor.render(body ?? "");
+    const metadata = result.metadata.frontmatter as Record<string, unknown>;
+    const proseWords = (
+        await splitToWords(String(metadata.proseText ?? ""))
+    ).nonPunctuationEntries.filter((entry) => /\S/u.test(entry.text)).length;
+    const codeNonEmptyLines = Number(metadata.codeNonEmptyLines ?? 0);
+    return {
+        hasImage: metadata.hasImage === true,
+        readingMinutes: Math.ceil(
+            proseWords / WORDS_PER_MINUTE +
+                (codeNonEmptyLines * CODE_SECONDS_PER_NON_EMPTY_LINE) / 60,
+        ),
+        wordCount: proseWords,
+    };
+}

package/src/schemas/collections.ts

@@ -61,8 +61,6 @@ const postSchema = z.object({
             invalid_type_error: "desc must be a string",
         })
         .min(1, "desc is required"),
-    minutesRead: z.string().optional(),
-    wordCount: z.number().optional(),
     cc: z.string().optional().default("CC-BY-NC-SA-4.0"),
     cover: z.string().optional(),
 });

package/src/utils/word-count-utils.ts

@@ -1,34 +1,12 @@
+import { analyzeFeatureFlags } from "@plugins/feature-flags";
 /**
  * 站点字数统计工具
  *
  * Astro 7.1 启用 deferRender 后，post.rendered 不再在 content sync 阶段预计算。
- * 因此改为直接从 post.body（原始 markdown）统计字数，不再依赖 Sätteri 的渲染结果。
+ * 因此统一通过 Sätteri AST 分析 post.body，确保各出口使用同一套结果。
  */
 import { getCollection } from "astro:content";
 
-// ---------------------------------------------------------------------------
-// 字数统计算法（基于 W3C Intl.Segmenter，body-based）
-// ---------------------------------------------------------------------------
-
-const bodySegmenter = new Intl.Segmenter("zh", { granularity: "word" });
-
-/**
- * 从原始 markdown 正文估算字数。
- * - 中文：`Intl.Segmenter` 逐字分割，每个汉字计 1
- * - 英文：按词计 1
- * - 纯数字序列：不计入（与旧正则行为一致）
- * - 标点/空白不计
- */
-function countWordsFromBody(body: string | undefined | null): number {
-    if (!body) return 0;
-    let count = 0;
-    for (const { segment, isWordLike } of bodySegmenter.segment(body)) {
-        // 排除纯数字序列（如年份、代码行号等），与旧正则行为保持一致
-        if (isWordLike && !/^\d+$/.test(segment)) count++;
-    }
-    return count;
-}
-
 // ---------------------------------------------------------------------------
 // 模块级缓存（整个构建周期只计算一次）
 // ---------------------------------------------------------------------------
@@ -43,7 +21,8 @@ export async function getTotalWordCount(): Promise<number> {
     if (_cachedTotal !== null) return _cachedTotal;
     try {
         const posts = await getCollection("posts", ({ data }) => !data.draft);
-        _cachedTotal = posts.reduce((sum, post) => sum + countWordsFromBody(post.body), 0);
+        const results = await Promise.all(posts.map((post) => analyzeFeatureFlags(post.body)));
+        _cachedTotal = results.reduce((sum, result) => sum + result.wordCount, 0);
         return _cachedTotal;
     } catch (error) {
         console.error("计算文章总字数时出错:", error);
@@ -56,18 +35,20 @@ let _cachedDescriptions: Map<string, { desc: string; wordCount: number }> | null
 /**
  * 获取所有已发布文章的描述和字数缓存，供 RSS/Atom/LLMs feed 共享。
  * - desc   ：直接从 frontmatter 读取（post.data.desc）
- * - wordCount：从 post.body 估算
+ * - wordCount：从 Sätteri AST 分析 post.body
  */
 export async function getPostDescriptions(): Promise<
     Map<string, { desc: string; wordCount: number }>
 > {
     if (_cachedDescriptions !== null) return _cachedDescriptions;
     const posts = await getCollection("posts", ({ data }) => !data.draft);
-    const results = posts.map((post) => ({
-        id: String(post.data.abbrlink ?? post.id),
-        desc: post.data.desc || "",
-        wordCount: countWordsFromBody(post.body),
-    }));
+    const results = await Promise.all(
+        posts.map(async (post) => ({
+            id: String(post.data.abbrlink ?? post.id),
+            desc: post.data.desc || "",
+            wordCount: (await analyzeFeatureFlags(post.body)).wordCount,
+        })),
+    );
     _cachedDescriptions = new Map(
         results.map((r) => [r.id, { desc: r.desc, wordCount: r.wordCount }]),
     );