@astro-minimax/cli 0.5.0 → 0.7.1

package/template/tools/build-voice-profile.ts

@@ -1,322 +0,0 @@
-#!/usr/bin/env npx tsx
-/**
- * 构建作者表达风格画像
- *
- * 从博客标题、正文中提取作者的表达风格特征。
- * 纯本地分析，不调用 AI。
- *
- * 用法:
- *   pnpm voice:build   构建风格画像
- */
-
-import { readdir, readFile } from "node:fs/promises";
-import { join } from "node:path";
-import {
-  loadEnv,
-  writeJson,
-  DATA_DIR,
-  BLOG_DIR,
-} from "./lib/utils.js";
-import { extractFrontmatter } from "./lib/frontmatter.js";
-
-// ─── 常量 ─────────────────────────────────────────────────────
-
-const OUTPUT_FILE = join(DATA_DIR, "voice-profile.json");
-
-// ─── 文章收集 ─────────────────────────────────────────────────
-
-interface Post {
-  title: string;
-  body: string;
-  category: string;
-  lang: string;
-}
-
-async function collectPosts(): Promise<Post[]> {
-  const entries = await readdir(BLOG_DIR, { withFileTypes: true });
-  const posts: Post[] = [];
-
-  for (const entry of entries) {
-    if (!entry.isDirectory() || entry.name.startsWith("_")) continue;
-
-    const subDir = join(BLOG_DIR, entry.name);
-    const subEntries = await readdir(subDir, { withFileTypes: true });
-
-    for (const subEntry of subEntries) {
-      if (!subEntry.isFile() || !subEntry.name.endsWith(".md")) continue;
-
-      const filePath = join(subDir, subEntry.name);
-      const raw = await readFile(filePath, "utf-8");
-      const fm = extractFrontmatter(raw);
-
-      if (!fm.data.title || fm.data.draft) continue;
-
-      const relativePath = filePath.replace(BLOG_DIR + "/", "");
-      const lang = relativePath.startsWith("en/") ? "en" : "zh";
-
-      posts.push({
-        title: String(fm.data.title),
-        body: fm.body,
-        category: String(fm.data.category || ""),
-        lang,
-      });
-    }
-  }
-
-  return posts;
-}
-
-// ─── 标题风格分析 ─────────────────────────────────────────────
-
-function analyzeTitlePatterns(posts: Post[]) {
-  const patterns = {
-    usesVerticalBar: 0, // "xxx | xxx"
-    usesColon: 0, // "xxx:xxx" or "xxx：xxx"
-    usesMiddleDot: 0, // "xxx・xxx"
-    usesQuestion: 0, // 标题是问句
-    usesEmoji: 0,
-    usesEnglish: 0, // 中英混用
-    usesYear: 0, // 标题带年份
-    usesSeriesFormat: 0, // 系列文章
-  };
-
-  for (const post of posts) {
-    const title = post.title;
-
-    if (title.includes("|") || title.includes("｜")) patterns.usesVerticalBar++;
-    if (title.includes(":") || title.includes("：")) patterns.usesColon++;
-    if (title.includes("・") || title.includes("·")) patterns.usesMiddleDot++;
-    if (/[?？]/.test(title)) patterns.usesQuestion++;
-    if (/[\u{1F300}-\u{1F9FF}]/u.test(title)) patterns.usesEmoji++;
-    if (/[a-zA-Z]{3,}/.test(title)) patterns.usesEnglish++;
-    if (/20\d{2}/.test(title)) patterns.usesYear++;
-    if (
-      /[壹贰叁肆伍陆柒捌玖拾]|Day\s?\d|Part\s?\d|[（(]\d[）)]/i.test(title)
-    ) {
-      patterns.usesSeriesFormat++;
-    }
-  }
-
-  const total = posts.length || 1;
-
-  return {
-    total: posts.length,
-    patterns,
-    style: {
-      vertical_bar_rate: Math.round(
-        (patterns.usesVerticalBar / total) * 100
-      ),
-      colon_rate: Math.round((patterns.usesColon / total) * 100),
-      question_rate: Math.round((patterns.usesQuestion / total) * 100),
-      english_mix_rate: Math.round((patterns.usesEnglish / total) * 100),
-      year_in_title_rate: Math.round((patterns.usesYear / total) * 100),
-    },
-  };
-}
-
-// ─── 正文风格分析 ─────────────────────────────────────────────
-
-function analyzeContentPatterns(posts: Post[]) {
-  let totalLength = 0;
-  let postsWithCode = 0;
-  let postsWithLinks = 0;
-  let postsWithImages = 0;
-  let postsWithHeadings = 0;
-  let postsWithLists = 0;
-  let postsWithQuotes = 0;
-
-  for (const post of posts) {
-    const body = post.body;
-    totalLength += body.length;
-
-    if (/```[\s\S]*?```/.test(body)) postsWithCode++;
-    if (/\[[^\]]+\]\([^)]+\)/.test(body)) postsWithLinks++;
-    if (/!\[[^\]]*\]\([^)]+\)/.test(body)) postsWithImages++;
-    if (/^#{1,6}\s/m.test(body)) postsWithHeadings++;
-    if (/^[\-\*]\s/m.test(body)) postsWithLists++;
-    if (/^>\s/m.test(body)) postsWithQuotes++;
-  }
-
-  const total = posts.length || 1;
-
-  return {
-    avgLength: Math.round(totalLength / total),
-    codeBlockRate: Math.round((postsWithCode / total) * 100),
-    linkRate: Math.round((postsWithLinks / total) * 100),
-    imageRate: Math.round((postsWithImages / total) * 100),
-    headingRate: Math.round((postsWithHeadings / total) * 100),
-    listRate: Math.round((postsWithLists / total) * 100),
-    quoteRate: Math.round((postsWithQuotes / total) * 100),
-  };
-}
-
-// ─── 表达习惯提取 ─────────────────────────────────────────────
-
-function extractExpressionHabits(posts: Post[]) {
-  const connectors: Record<string, number> = {};
-
-  const connectorPatterns = [
-    "其实",
-    "说实话",
-    "不过",
-    "但是",
-    "所以",
-    "然后",
-    "感觉",
-    "突然",
-    "顺手",
-    "顺便",
-    "随手",
-    "折腾",
-    "踩坑",
-    "跳坑",
-    "体验",
-    "分享",
-    "推荐",
-    "安利",
-    "种草",
-    "注意",
-    "重要",
-    "关键",
-    "总结",
-  ];
-
-  for (const post of posts) {
-    const text = `${post.title} ${post.body}`;
-
-    for (const pattern of connectorPatterns) {
-      const count = (text.match(new RegExp(pattern, "g")) || []).length;
-      if (count > 0) {
-        connectors[pattern] = (connectors[pattern] || 0) + count;
-      }
-    }
-  }
-
-  const sortedConnectors = Object.entries(connectors)
-    .sort(([, a], [, b]) => b - a)
-    .slice(0, 15)
-    .map(([word, count]) => ({ word, count }));
-
-  return {
-    frequent_expressions: sortedConnectors,
-    style_notes: [
-      "标题常用竖线分隔格式：「主题 | 标题」",
-      "技术类标题直接用英文术语，不强行翻译",
-      "生活类标题偏感性，常用问句或感叹",
-      "喜欢用「折腾」「踩坑」等口语化表达描述技术探索",
-      "分享经验时常用「攻略」「指南」「记」等实用性标题词",
-    ],
-  };
-}
-
-// ─── 分类风格分析 ─────────────────────────────────────────────
-
-function analyzeCategoryStyles(posts: Post[]) {
-  const categoryMap = new Map<string, Post[]>();
-
-  for (const post of posts) {
-    const cat = post.category || "其他";
-    if (!categoryMap.has(cat)) categoryMap.set(cat, []);
-    categoryMap.get(cat)!.push(post);
-  }
-
-  const categoryStyles: Record<string, { count: number; avgLength: number }> =
-    {};
-
-  for (const [cat, catPosts] of categoryMap.entries()) {
-    const avgLength =
-      catPosts.reduce((sum, p) => sum + p.body.length, 0) / catPosts.length;
-    categoryStyles[cat] = {
-      count: catPosts.length,
-      avgLength: Math.round(avgLength),
-    };
-  }
-
-  return categoryStyles;
-}
-
-// ─── 主流程 ───────────────────────────────────────────────────
-
-async function main() {
-  await loadEnv();
-
-  console.log("🎤 构建作者表达风格画像");
-  console.log("━".repeat(50));
-  console.log("");
-
-  console.log("📂 扫描博客文章...");
-  const posts = await collectPosts();
-  console.log(`   找到 ${posts.length} 篇文章`);
-
-  const zhPosts = posts.filter((p) => p.lang === "zh");
-  const enPosts = posts.filter((p) => p.lang === "en");
-  console.log(`   - 中文: ${zhPosts.length} 篇`);
-  console.log(`   - 英文: ${enPosts.length} 篇`);
-
-  // 分析各维度
-  console.log("\n🔍 分析风格特征...");
-  const titleAnalysis = analyzeTitlePatterns(posts);
-  const contentAnalysis = analyzeContentPatterns(posts);
-  const expressionHabits = extractExpressionHabits(posts);
-  const categoryStyles = analyzeCategoryStyles(posts);
-
-  // 构建画像
-  const profile = {
-    $schema: "voice-profile-v1",
-    generatedAt: new Date().toISOString(),
-    overall_tone: {
-      description:
-        "技术博客风格，注重实践与可操作性，语言简洁明了",
-      primary_persona: "技术博主",
-      communication_style: "先给结论，再补细节",
-    },
-    blog_title_style: titleAnalysis,
-    content_style: contentAnalysis,
-    expression_habits: expressionHabits,
-    category_styles: categoryStyles,
-    style_modes: {
-      technical: {
-        description: "技术类文章",
-        traits: [
-          "直接给方案",
-          "会提到具体工具和版本",
-          "不回避踩过的坑",
-          "代码片段简洁",
-        ],
-      },
-      tutorial: {
-        description: "教程类文章",
-        traits: [
-          "步骤清晰",
-          "配合截图或代码",
-          "注意事项突出",
-          "有总结部分",
-        ],
-      },
-      lifestyle: {
-        description: "生活类文章",
-        traits: ["语气更随意", "偶尔自嘲", "不说教", "分享真实体验"],
-      },
-    },
-  };
-
-  await writeJson(OUTPUT_FILE, profile);
-
-  console.log("\n✅ 风格画像构建完成");
-  console.log(`📄 输出文件: ${OUTPUT_FILE}`);
-  console.log("\n📊 风格概览:");
-  console.log(`   标题竖线分隔率: ${titleAnalysis.style.vertical_bar_rate}%`);
-  console.log(`   标题中英混用率: ${titleAnalysis.style.english_mix_rate}%`);
-  console.log(`   平均文章长度: ${contentAnalysis.avgLength} 字符`);
-  console.log(
-    `   高频表达: ${expressionHabits.frequent_expressions
-      .slice(0, 5)
-      .map((e) => e.word)
-      .join("、")}`
-  );
-}
-
-main().catch((error) => {
-  console.error("❌ 构建失败:", error.message);
-  process.exit(1);
-});

package/template/tools/generate-author-profile.ts

@@ -1,369 +0,0 @@
-#!/usr/bin/env npx tsx
-/**
- * 生成作者画像报告
- *
- * 基于作者上下文数据生成用于 About 页面的结构化简介。
- * 支持 AI 生成和规则模板两种模式。
- *
- * 用法:
- *   pnpm profile:generate              AI 生成画像报告
- *   pnpm profile:generate --no-ai      使用规则模板
- *   pnpm profile:generate --force      强制重新生成（不回退）
- */
-
-import { readdir, readFile } from "node:fs/promises";
-import { join } from "node:path";
-import {
-  loadEnv,
-  readJson,
-  writeJson,
-  truncate,
-  parseCliArgs,
-  DATA_DIR,
-  BLOG_DIR,
-  } from "./lib/utils.js";
-import { stripMarkdown } from "./lib/markdown.js";
-import { extractFrontmatter } from "./lib/frontmatter.js";
-import { chatCompletion, hasAPIKey, getConfig } from "./lib/ai-provider.js";
-
-// ─── 常量 ─────────────────────────────────────────────────────
-
-const OUTPUT_REPORT = join(DATA_DIR, "author-profile-report.json");
-const OUTPUT_CONTEXT = join(DATA_DIR, "author-profile-context.json");
-const DEFAULT_SITE_URL = "https://example.com";
-
-// ─── CLI 参数 ─────────────────────────────────────────────────
-
-interface CliFlags {
-  force: boolean;
-  noAI: boolean;
-}
-
-function parseArgs(): CliFlags {
-  return parseCliArgs({ force: false, noAI: false });
-}
-
-// ─── 文章收集 ─────────────────────────────────────────────────
-
-interface Post {
-  title: string;
-  date: string;
-  lang: string;
-  category: string;
-  tags: string[];
-  description: string;
-  summary?: string;
-  keyPoints?: string[];
-  url: string;
-}
-
-async function collectPosts(_siteUrl: string): Promise<Post[]> {
-  const entries = await readdir(BLOG_DIR, { withFileTypes: true });
-  const aiSummaries = await readJson<{
-    articles?: Record<string, { data?: { summary?: string; keyPoints?: string[] } }>
-  }>(join(DATA_DIR, "ai-summaries.json"), {
-    articles: {},
-  });
-  const posts: Post[] = [];
-
-  for (const entry of entries) {
-    if (!entry.isDirectory() || entry.name.startsWith("_")) continue;
-
-    const subDir = join(BLOG_DIR, entry.name);
-    const subEntries = await readdir(subDir, { withFileTypes: true });
-
-    for (const subEntry of subEntries) {
-      if (!subEntry.isFile() || !subEntry.name.endsWith(".md")) continue;
-
-      const filePath = join(subDir, subEntry.name);
-      const raw = await readFile(filePath, "utf-8");
-      const fm = extractFrontmatter(raw);
-
-      if (!fm.data.title || fm.data.draft) continue;
-
-      const relativePath = filePath.replace(BLOG_DIR + "/", "");
-      const id = relativePath.replace(/\.md$/, "");
-      const lang = relativePath.startsWith("en/") ? "en" : "zh";
-      const slug = id.split("/").slice(1).join("/");
-      const summaryEntry = aiSummaries.articles?.[id]?.data;
-
-      posts.push({
-        title: String(fm.data.title),
-        date: String(fm.data.pubDatetime),
-        lang,
-        category: String(fm.data.category || ""),
-        tags: Array.isArray(fm.data.tags) ? (fm.data.tags as string[]) : [],
-        description: String(fm.data.description || ""),
-        summary: summaryEntry?.summary || truncate(stripMarkdown(fm.body), 100),
-        keyPoints: summaryEntry?.keyPoints || [],
-        url: `/${lang}/posts/${slug}/`,
-      });
-    }
-  }
-
-  posts.sort((a, b) => new Date(b.date).getTime() - new Date(a.date).getTime());
-  return posts;
-}
-
-// ─── 上下文构建 ───────────────────────────────────────────────
-
-function buildContext(posts: Post[], siteUrl: string) {
-  const selectedPosts = posts.slice(0, 12);
-  const zhPosts = posts.filter((p) => p.lang === "zh");
-  const enPosts = posts.filter((p) => p.lang === "en");
-
-  // 标签聚合
-  const tagCounts = new Map<string, number>();
-  for (const post of posts) {
-    for (const tag of post.tags) {
-      tagCounts.set(tag, (tagCounts.get(tag) || 0) + 1);
-    }
-  }
-  const topTags = [...tagCounts.entries()]
-    .sort((a, b) => b[1] - a[1])
-    .slice(0, 15)
-    .map(([tag]) => tag);
-
-  // 分类聚合
-  const categoryCounts = new Map<string, number>();
-  for (const post of posts) {
-    if (post.category) {
-      categoryCounts.set(
-        post.category,
-        (categoryCounts.get(post.category) || 0) + 1
-      );
-    }
-  }
-  const topCategories = [...categoryCounts.entries()]
-    .sort((a, b) => b[1] - a[1])
-    .slice(0, 5)
-    .map(([cat]) => cat);
-
-  return {
-    generatedAt: new Date().toISOString(),
-    siteUrl,
-    sourceInfo: {
-      totalPosts: posts.length,
-      zhPosts: zhPosts.length,
-      enPosts: enPosts.length,
-      selectedPosts: selectedPosts.length,
-    },
-    profile: {
-      name: process.env.SITE_AUTHOR || "博主",
-      siteUrl,
-    },
-    posts: selectedPosts.map((post) => ({
-      title: post.title,
-      date: post.date,
-      categories: [post.category].filter(Boolean),
-      tags: post.tags.slice(0, 5),
-      summary: post.summary,
-      keyPoints: post.keyPoints,
-      url: post.url,
-    })),
-    topTags,
-    topCategories,
-    contentStats: {
-      totalPosts: posts.length,
-      avgPostPerMonth: Math.round(
-        posts.length / Math.max(1, calculateMonthsSpan(posts))
-      ),
-    },
-  };
-}
-
-function calculateMonthsSpan(posts: Post[]): number {
-  if (posts.length < 2) return 1;
-  const latest = new Date(posts[0].date);
-  const earliest = new Date(posts[posts.length - 1].date);
-  return (
-    (latest.getFullYear() - earliest.getFullYear()) * 12 +
-    (latest.getMonth() - earliest.getMonth()) +
-    1
-  );
-}
-
-// ─── AI 生成 ───────────────────────────────────────────────────
-
-async function generateReportWithAI(
-  context: ReturnType<typeof buildContext>
-): Promise<unknown> {
-  if (!hasAPIKey()) {
-    throw new Error("未配置 AI API Key");
-  }
-
-  const config = getConfig();
-
-  const systemPrompt = `你是一位中文科技写作编辑。请基于给定上下文，以第三方视角生成作者画像 JSON。
-要求：
-1. 严格输出 JSON，不要输出 Markdown 或多余文本。
-2. 语气客观、克制、具体，不要夸张和空泛。
-3. 结论必须可由上下文支撑，避免编造。
-4. 文案使用中文，第三人称，不使用"我"。
-
-输出 schema:
-{
-  "report": {
-    "hero": {"title":"AI 视角下的作者","summary":"...","intro":"..."},
-    "identities":[{"name":"...","description":"...","evidence":"..."}],
-    "strengths":[{"title":"...","points":["..."]}],
-    "styles":[{"trait":"...","description":"..."}],
-    "proofs":{
-      "posts":[{"title":"...","url":"...","reason":"...","date":"YYYY-MM-DD"}]
-    },
-    "disclaimer":"..."
-  }
-}`;
-
-  const contextText = JSON.stringify(context, null, 2).slice(0, 25000);
-  const userPrompt = `上下文数据如下，请根据这些信息生成报告：\n${contextText}`;
-
-  const content = await chatCompletion(
-    [
-      { role: "system", content: systemPrompt },
-      { role: "user", content: userPrompt },
-    ],
-    { maxTokens: 3000, responseFormat: "json" }
-  );
-
-  const parsed = JSON.parse(content);
-  return {
-    meta: {
-      lastUpdated: new Date().toISOString(),
-      model: config.model,
-      generatedBy: "ai",
-    },
-    report: parsed.report || parsed,
-  };
-}
-
-// ─── 规则模板生成 ─────────────────────────────────────────────
-
-function buildRuleBasedReport(
-  context: ReturnType<typeof buildContext>
-): unknown {
-  const posts = context.posts.slice(0, 5).map((post) => ({
-    title: post.title,
-    url: post.url,
-    reason:
-      post.summary?.slice(0, 60) || "该文章体现了作者的写作风格。",
-    date: post.date,
-  }));
-
-  const topCategories = context.topCategories.slice(0, 3).join("、");
-  const authorName = context.profile.name;
-
-  return {
-    meta: {
-      lastUpdated: new Date().toISOString(),
-      model: "rule-based-template",
-      generatedBy: "rule-based",
-    },
-    report: {
-      hero: {
-        title: `AI 视角下的 ${authorName}`,
-        summary: `一位专注于${topCategories}领域的博主，持续输出高质量内容。`,
-        intro: `博客已发布 ${context.sourceInfo.totalPosts} 篇文章，涵盖 ${context.topTags.length} 个主题标签。`,
-      },
-      identities: [
-        {
-          name: "技术博主",
-          description: "持续分享技术经验与实践心得。",
-          evidence: `已发布 ${context.sourceInfo.totalPosts} 篇文章。`,
-        },
-        {
-          name: "内容创作者",
-          description: "注重内容质量与读者体验。",
-          evidence: posts[0]
-            ? `近期文章《${posts[0].title}》体现了专业的写作风格。`
-            : "",
-        },
-      ],
-      strengths: [
-        {
-          title: "内容深度",
-          points: [
-            "文章结构清晰，逻辑性强",
-            "注重实践，配合代码示例",
-            "持续更新，覆盖多个技术领域",
-          ],
-        },
-        {
-          title: "表达风格",
-          points: ["语言简洁明了", "注重可读性", "善于总结提炼"],
-        },
-      ],
-      styles: [
-        {
-          trait: "技术导向",
-          description: "专注于技术内容的深度讲解与实践分享。",
-        },
-        {
-          trait: "结构化表达",
-          description: "文章结构清晰，便于读者理解和学习。",
-        },
-      ],
-      proofs: { posts },
-      disclaimer:
-        "该页面由 AI 归纳与规则模板联合生成，旨在帮助访客快速建立认知，可能存在概括偏差，请以原始文章信息为准。",
-    },
-  };
-}
-
-// ─── 主流程 ───────────────────────────────────────────────────
-
-async function main() {
-  const args = parseArgs();
-  await loadEnv();
-
-  console.log("👤 生成作者画像报告");
-  console.log("━".repeat(50));
-
-  const siteUrl = process.env.SITE_URL || DEFAULT_SITE_URL;
-  console.log(`   站点 URL: ${siteUrl}`);
-  console.log(`   输出目录: ${DATA_DIR}`);
-  console.log("");
-
-  // 收集数据
-  console.log("📂 收集文章数据...");
-  const posts = await collectPosts(siteUrl);
-  console.log(`   找到 ${posts.length} 篇文章`);
-
-  const context = buildContext(posts, siteUrl);
-  await writeJson(OUTPUT_CONTEXT, context);
-  console.log(`   上下文已保存: ${OUTPUT_CONTEXT}`);
-
-  // 生成报告
-  let report: unknown;
-
-  if (args.noAI) {
-    console.log("\n📝 使用规则模板生成...");
-    report = buildRuleBasedReport(context);
-  } else {
-    console.log("\n🤖 使用 AI 生成...");
-    try {
-      report = await generateReportWithAI(context);
-      console.log("   ✅ AI 生成成功");
-    } catch (error) {
-      const err = error as Error;
-      if (!args.force) {
-        console.warn(`   ⚠️  AI 生成失败: ${err.message}`);
-        console.log("   📝 回退使用规则模板...");
-        report = buildRuleBasedReport(context);
-      } else {
-        throw error;
-      }
-    }
-  }
-
-  await writeJson(OUTPUT_REPORT, report);
-
-  console.log("\n✅ 画像报告生成完成");
-  console.log(`📄 报告文件: ${OUTPUT_REPORT}`);
-  console.log(`🧩 上下文文件: ${OUTPUT_CONTEXT}`);
-}
-
-main().catch((error) => {
-  console.error("❌ 生成失败:", error.message);
-  process.exit(1);
-});