czon 0.1.1 → 0.1.3

package/src/ai/extractMetadataFromMarkdown.ts

@@ -1,95 +0,0 @@
-import { completeMessages, OpenAIMessage } from '../services/openai';
-import { AIMetadata } from '../types';
-
-/**
- * 从 markdown 内容中提取 metadata
- * @param content Markdown 内容
- * @returns Promise<AIMetadata> 提取的元数据，失败时抛出错误
- */
-export async function extractMetadataFromMarkdown(content: string): Promise<AIMetadata> {
-  const prompt = buildMetadataPrompt(content);
-  const messages: OpenAIMessage[] = [
-    {
-      role: 'system',
-      content:
-        '你是一个专业的文档分析助手，擅长从文档中提取结构化信息。请严格按照要求的 JSON 格式返回结果。',
-    },
-    {
-      role: 'user',
-      content: prompt,
-    },
-  ];
-
-  const response = await completeMessages(messages, {
-    response_format: { type: 'json_object' },
-  });
-
-  const metadata = parseMetadataResponse(response.choices[0].message.content);
-
-  // 添加 tokens 使用情况
-  metadata.tokens_used = {
-    prompt: response.usage.prompt_tokens,
-    completion: response.usage.completion_tokens,
-    total: response.usage.total_tokens,
-  };
-
-  return metadata;
-}
-
-/**
- * 构建提取 metadata 的 prompt
- */
-function buildMetadataPrompt(content: string): string {
-  // 限制内容长度以避免 token 超限
-  const maxContentLength = Infinity; // 可根据需要调整长度限制
-  const truncatedContent =
-    content.length > maxContentLength
-      ? content.substring(0, maxContentLength) + '... [内容已截断]'
-      : content;
-
-  return `请分析以下文档内容，提取以下信息并返回 JSON 格式：
-
-文档内容：
-"""
-${truncatedContent}
-"""
-
-请提取：
-1. title: 文档的标题（简洁明了，不超过 30 个字）
-2. summary: 文档摘要（控制在 300 字以内，概括主要内容）
-3. tags: 关键词列表（3-8 个关键词，使用中文或英文）
-4. inferred_date: 文档中隐含的创建日期（如果有的话，格式：YYYY-MM-DD，没有就留空字符串）
-5. inferred_lang: 文档使用的语言代码（例如：zh-Hans 表示简体中文，en-US 表示美式英语）
-
-请严格按照以下 JSON 格式返回，不要包含任何其他文本：
-{
-  "title": "文档标题",
-  "summary": "文档摘要...",
-  "tags": ["关键词1", "关键词2", "关键词3"],
-  "inferred_date": "2023-01-01",
-  "inferred_lang": "zh-Hans"
-}`;
-}
-
-/**
- * 解析 AI 返回的 metadata
- */
-function parseMetadataResponse(responseContent: string): AIMetadata {
-  try {
-    const metadata = JSON.parse(responseContent);
-
-    // 验证和清理数据
-    return {
-      title: metadata.title?.trim() || '未命名文档',
-      summary: metadata.summary?.trim() || '',
-      tags: Array.isArray(metadata.tags)
-        ? metadata.tags.map((tag: string) => tag.trim()).filter(Boolean)
-        : [],
-      inferred_date: metadata.inferred_date?.trim() || undefined,
-      inferred_lang: metadata.inferred_lang?.trim() || 'zh-Hans',
-    };
-  } catch (error) {
-    console.error('❌ Failed to parse AI response:', error, 'Response:', responseContent);
-    throw error;
-  }
-}

package/src/ai/translateMarkdown.ts

@@ -1,60 +0,0 @@
-import { LANGUAGE_NAMES } from '../languages';
-import { completeMessages, OpenAIMessage } from '../services/openai';
-
-/**
-
- * @param content Markdown 内容
- * @param targetLang 目标语言代码（例如：zh-Hans, en-US）
- * @returns Promise<string> 翻译后的 Markdown 内容
- */
-export async function translateMarkdown(content: string, targetLang: string): Promise<string> {
-  const langName = LANGUAGE_NAMES[targetLang];
-  const lang = `${langName} (${targetLang})`;
-  const messages: OpenAIMessage[] = [
-    {
-      role: 'system',
-      content: [
-        `You are a professional document translator.`,
-        `Translate the following markdown content into ${lang}`,
-        targetLang === 'ja-JP'
-          ? [
-              //
-              '使用日语母语者自然的表达方式。',
-              '采用适合技术/专业文档的礼貌、正式语体（丁寧体/です・ます体）',
-              '确保语法和字符集完全符合日语规范。',
-              '绝对禁止使用繁体中文汉字。',
-              '所有日语汉字必须使用标准的 **日本常用汉字（Jōyō kanji）** 字形。',
-              '特别检查以下字形示例，确保使用日文标准字形：',
-              '1. “国” - 使用“国”而非“國”',
-              '2. “学” - 使用“学”而非“學”',
-              '3. “広” - 使用“広”而非“廣”',
-              '4. “円” - 使用“円”而非“圓”',
-              '5. “医” - 使用“医”而非“醫”',
-              '6. “図” - 使用“図”而非“圖”',
-              '7. “対” - 使用“対”而非“對”',
-              '8. “声” - 使用“声”而非“聲”',
-              '9. “芸” - 使用“芸”而非“藝”',
-              '10. “験” - 使用“験”而非“驗”',
-            ].join('\n')
-          : ``,
-        `Preserve the original markdown formatting, including headings, lists, code blocks, links, and images.`,
-        `Do not change any non-text elements or their formatting.`,
-        `Ensure that technical terms and code snippets remain unchanged.`,
-        `Provide a natural and fluent translation suitable for readers familiar with the subject matter.`,
-      ].join('\n'),
-    },
-    {
-      role: 'user',
-      content: content,
-    },
-  ];
-
-  const response = await completeMessages(messages);
-  const translatedContent = response.choices[0]?.message?.content?.trim() || '';
-
-  if (!translatedContent) {
-    throw new Error('Empty translation response');
-  }
-
-  return translatedContent;
-}

package/src/build/pipeline.ts

@@ -1,172 +0,0 @@
-import * as fs from 'fs/promises';
-import * as path from 'path';
-import { translateMarkdown } from '../ai/translateMarkdown';
-import { loadMetaData, MetaData, saveMetaData } from '../metadata';
-import { INPUT_DIR, ZEN_DIR, ZEN_DIST_DIR, ZEN_SRC_DIR } from '../paths';
-import { extractMetadataByAI } from '../process/extractMetadataByAI';
-import { scanSourceFiles } from '../process/scanSourceFiles';
-import { renderTemplates } from '../process/template';
-import { BuildOptions } from '../types';
-import { updateFrontmatter } from '../utils/frontmatter';
-
-/**
- * 验证构建配置
- */
-async function validateConfig(options: BuildOptions): Promise<void> {
-  const { verbose = false } = options;
-
-  if (verbose) {
-    console.log(`🚀 Starting ZEN build...`);
-    if (options.langs && options.langs.length > 0) {
-      console.log(`🌐 Target languages: ${options.langs.join(', ')}`);
-    }
-    console.log(`🔍 Verbose mode enabled`);
-  }
-
-  MetaData.options = options;
-}
-
-/**
- * 存储母语文件到 .zen/src
- */
-async function storeNativeFiles(): Promise<void> {
-  const {
-    options: { verbose },
-    files,
-  } = MetaData;
-  for (const file of MetaData.files) {
-    try {
-      if (!file.hash) throw new Error(`Missing hash`);
-      if (!file.metadata?.inferred_lang) throw new Error(`Missing inferred language`);
-      const filePath = path.join(ZEN_SRC_DIR, file.metadata.inferred_lang, file.hash + '.md');
-      const originalContent = await fs.readFile(path.join(INPUT_DIR, file.path), 'utf-8');
-
-      const enhancedContent = updateFrontmatter(originalContent, {
-        title: file.metadata.title,
-        summary: file.metadata.summary,
-        tags: file.metadata.tags,
-        inferred_date: file.metadata.inferred_date,
-        inferred_lang: file.metadata.inferred_lang,
-      });
-      await fs.mkdir(path.dirname(filePath), { recursive: true });
-
-      await fs.writeFile(filePath, enhancedContent, 'utf-8');
-    } catch (error) {
-      console.warn(`⚠️ Failed to store native file ${file.path}:`, error);
-    }
-  }
-
-  if (verbose && files.length > 0) {
-    console.log(`💾 Stored ${files.length} native language files to .zen/src`);
-  }
-}
-
-/**
- * 处理翻译
- */
-async function processTranslations(): Promise<void> {
-  const {
-    files,
-    options: { langs = [], verbose },
-  } = MetaData;
-
-  await Promise.all(
-    files.flatMap(async file => {
-      return Promise.all(
-        langs.map(async lang => {
-          if (verbose) console.info(`📄 Processing file for translation: ${file.path}`);
-          if (!file.metadata) {
-            console.warn(`⚠️ Missing metadata for file: ${file.path}, skipping translation.`);
-            return;
-          }
-          if (verbose) console.log(`🌐 Translating to ${lang}...`);
-          // 存储翻译文件到 .zen/src/{lang}
-          const sourcePath = path.join(ZEN_SRC_DIR, file.metadata.inferred_lang, file.hash + '.md'); // 使用已经加强的母语文件路径
-          const targetPath = path.join(ZEN_SRC_DIR, lang, file.hash + '.md');
-
-          try {
-            const content = await fs.readFile(sourcePath, 'utf-8');
-            if (file.metadata.inferred_lang === lang) {
-              if (verbose)
-                console.log(`ℹ️ Skipping translation for ${file.path}, already in target language`);
-              return;
-            } else {
-              // 翻译
-              // 先检查是否已经有翻译文件存在
-
-              const exists = await fs.access(targetPath).then(
-                () => true,
-                () => false
-              );
-              if (exists) {
-                if (verbose)
-                  console.log(`ℹ️ Translation already exists for ${file.path} in ${lang}`);
-                return;
-              }
-            }
-
-            const translatedContent = await translateMarkdown(content, lang);
-
-            await fs.mkdir(path.dirname(targetPath), { recursive: true });
-            await fs.writeFile(targetPath, translatedContent, 'utf-8');
-
-            if (verbose) console.log(`✅ Translated file saved: ${targetPath}`);
-          } catch (error) {
-            console.error(`❌ Failed to translate to ${lang}:`, error);
-          }
-        })
-      );
-    })
-  );
-}
-
-/**
- * 构建管道（函数组合）
- */
-async function buildPipeline(options: BuildOptions): Promise<void> {
-  // 验证配置
-  await validateConfig(options);
-
-  // 清理输出目录
-  await fs.rm(ZEN_DIST_DIR, { recursive: true, force: true });
-
-  // 确保 .zen/.gitignore 文件
-  await fs.mkdir(ZEN_DIR, { recursive: true });
-  await fs.writeFile(path.join(ZEN_DIR, '.gitignore'), 'dist\n', 'utf-8');
-
-  // 扫描源文件
-  await scanSourceFiles();
-
-  // 运行 AI 元数据提取
-  await extractMetadataByAI();
-
-  // 存储母语文件
-  await storeNativeFiles();
-
-  // 处理翻译
-  await processTranslations();
-
-  // 渲染模板
-  await renderTemplates();
-}
-
-/**
- * 主构建函数
- */
-export async function buildSite(options: BuildOptions): Promise<void> {
-  const startTime = Date.now();
-
-  try {
-    await loadMetaData();
-    await buildPipeline(options);
-
-    const endTime = Date.now();
-    const duration = ((endTime - startTime) / 1000).toFixed(2);
-    console.log(`🎉 Build completed in ${duration}s`);
-  } catch (error) {
-    console.error(`❌ Build failed:`, error);
-    throw error;
-  } finally {
-    await saveMetaData();
-  }
-}

package/src/cli.ts

@@ -1,73 +0,0 @@
-#!/usr/bin/env node
-
-import { Cli, Command, Option } from 'clipanion';
-import { config } from 'dotenv';
-import * as fs from 'fs';
-import * as path from 'path';
-import { buildSite } from './build/pipeline';
-
-// 加载 .env 文件中的环境变量
-config();
-
-// 获取版本号 - 从 package.json 读取
-function getVersion(): string {
-  const packageJsonPath = path.join(__dirname, '..', 'package.json');
-  const packageJson = JSON.parse(fs.readFileSync(packageJsonPath, 'utf-8'));
-  return packageJson.version;
-}
-
-// Build 命令
-class BuildCommand extends Command {
-  static paths = [['build']];
-
-  template = Option.String('-t,--template');
-  verbose = Option.Boolean('-v,--verbose');
-  lang = Option.Array('--lang', {
-    description: 'Target languages for translation (e.g., en-US, ja-JP)',
-  });
-
-  static usage = Command.Usage({
-    description: 'Build documentation site from Markdown files in current directory',
-    details: `
-      This command builds a documentation site from Markdown files in the current directory.
-      The output will be placed in the .zen/dist directory.
-
-      Examples:
-        $ zengen build
-        $ zengen build --lang en-US --lang ja-JP (translate to English and Japanese)
-    `,
-  });
-
-  async execute() {
-    try {
-      await buildSite({
-        template: this.template ? path.resolve(this.template) : undefined,
-        verbose: this.verbose,
-        langs: this.lang,
-      });
-
-      return 0;
-    } catch (error) {
-      this.context.stderr.write(`❌ Build failed: ${error}\n`);
-      return 1;
-    }
-  }
-}
-
-// 创建 CLI 应用
-const cli = new Cli({
-  binaryName: 'zengen',
-  binaryLabel: 'ZEN - A minimalist Markdown documentation site builder',
-  binaryVersion: getVersion(),
-});
-
-// 注册命令
-cli.register(BuildCommand);
-
-// 运行 CLI
-cli.runExit(process.argv.slice(2), {
-  ...Cli.defaultContext,
-  stdin: process.stdin,
-  stdout: process.stdout,
-  stderr: process.stderr,
-});

package/src/findEntries.ts

@@ -1,37 +0,0 @@
-import { exec } from 'child_process';
-import { promisify } from 'util';
-
-const execAsync = promisify(exec);
-
-/**
- * 使用git命令查找所有Markdown文件
- * 使用git ls-files --others --cached --exclude-standard获取所有文件
- * 然后过滤掉.zen目录和只保留.md文件
- *
- * @param dirPath 要扫描的目录路径
- * @returns Promise<string[]> 返回Markdown文件的相对路径数组
- */
-export const findMarkdownEntries = async (dirPath: string): Promise<string[]> => {
-  try {
-    // 使用git命令获取所有文件（包括已跟踪和未跟踪的文件）
-    // 在指定的目录下执行git命令
-    const { stdout } = await execAsync('git ls-files --others --cached --exclude-standard', {
-      cwd: dirPath,
-    });
-
-    // 按行分割并过滤
-    const files = stdout
-      .split('\n')
-      .filter(line => line.trim() !== '') // 移除空行
-      .filter(file => !file.startsWith('.zen')) // 过滤掉.zen目录下的文件
-      .filter(file => file.endsWith('.md')); // 只保留.md文件
-
-    return files;
-  } catch (error) {
-    console.error('Error finding markdown entries:', error);
-
-    // 如果git命令失败，返回空数组
-    // 这可以处理没有git仓库或git不可用的情况
-    return [];
-  }
-};

package/src/index.ts

@@ -1 +0,0 @@
-import './cli';

package/src/languages.ts

@@ -1,37 +0,0 @@
-export const LANGUAGE_NAMES: Record<string, string> = {
-  // 现有语言
-  'zh-Hans': '简体中文',
-  'en-US': 'English',
-  'ja-JP': '日本語',
-  'ko-KR': '한국어',
-  'es-ES': 'Español',
-  'fr-FR': 'Français',
-  'de-DE': 'Deutsch',
-  'ru-RU': 'Русский',
-
-  // 补充缺失的项目支持语言
-  'pt-PT': 'Português',
-
-  // 欧洲主要语言
-  'it-IT': 'Italiano',
-  'nl-NL': 'Nederlands',
-  'pl-PL': 'Polski',
-  'sv-SE': 'Svenska',
-  'fi-FI': 'Suomi',
-  'da-DK': 'Dansk',
-  'no-NO': 'Norsk',
-
-  // 亚洲主要语言
-  'zh-Hant': '繁體中文',
-  'hi-IN': 'हिन्दी',
-  'ar-SA': 'العربية',
-  'th-TH': 'ไทย',
-  'vi-VN': 'Tiếng Việt',
-  'id-ID': 'Bahasa Indonesia',
-
-  // 其他重要语言
-  'pt-BR': 'Português (Brasil)',
-  'es-MX': 'Español (México)',
-  'tr-TR': 'Türkçe',
-  'uk-UA': 'Українська',
-};

package/src/metadata.ts

@@ -1,44 +0,0 @@
-import * as fs from 'fs/promises';
-import * as path from 'path';
-import { ZEN_META_PATH } from './paths';
-import { MetaDataStore } from './types';
-
-/**
- * 全局 MetaDataStore 单例
- */
-export const MetaData: MetaDataStore = {
-  // 稍后覆盖
-  version: '1.0.0',
-  options: {},
-  files: [],
-};
-
-/**
- * 从文件中读取数据，覆盖 store，但是要保持它仍然是同一个对象
- */
-export async function loadMetaData(): Promise<void> {
-  try {
-    await fs.access(ZEN_META_PATH);
-    const content = await fs.readFile(ZEN_META_PATH, 'utf-8');
-    const newData = JSON.parse(content);
-
-    // 使用 Object.assign 保持同一个对象引用
-    Object.assign(MetaData, newData);
-  } catch (error) {
-    // 如果文件不存在，初始化默认值
-    MetaData.version = '1.0.0';
-    MetaData.files = [];
-  }
-}
-
-/**
- * 将 MetaData 写入 store
- */
-export async function saveMetaData(): Promise<void> {
-  // 确保 .zen 目录存在
-  const zenDir = path.dirname(ZEN_META_PATH);
-  await fs.mkdir(zenDir, { recursive: true });
-
-  // 保存文件
-  await fs.writeFile(ZEN_META_PATH, JSON.stringify(MetaData, null, 2), 'utf-8');
-}

package/src/paths.ts

@@ -1,7 +0,0 @@
-import { join } from 'path';
-
-export const INPUT_DIR = process.cwd();
-export const ZEN_DIR = join(process.cwd(), '.zen');
-export const ZEN_DIST_DIR = join(ZEN_DIR, 'dist');
-export const ZEN_SRC_DIR = join(ZEN_DIR, 'src');
-export const ZEN_META_PATH = join(ZEN_DIR, 'meta.json');

package/src/process/extractMetadataByAI.ts

@@ -1,31 +0,0 @@
-import { readFile } from 'fs/promises';
-import { extractMetadataFromMarkdown } from '../ai/extractMetadataFromMarkdown';
-import { MetaData } from '../metadata';
-
-/**
- * 运行 AI 元数据提取
- */
-export async function extractMetadataByAI(): Promise<void> {
-  const { files } = MetaData;
-
-  if (MetaData.options.verbose) console.log(`🤖 Running AI metadata extraction...`);
-  console.log(`🤖 Processing ${files.length} files with AI...`);
-
-  await Promise.all(
-    files.map(async file => {
-      try {
-        if (file.metadata) {
-          console.info(`ℹ️ Skipping ${file.path}, already has metadata`);
-          return;
-        }
-        const content = await readFile(file.path, 'utf-8');
-        file.metadata = await extractMetadataFromMarkdown(content);
-        console.log(`✅ Extracted AI metadata for ${file.path}`, file.metadata.tokens_used);
-      } catch (error) {
-        console.error(`⚠️ Failed to process file ${file.path}:`, error);
-      }
-    })
-  );
-
-  console.log(`✅ AI processing completed for ${files.length} files`);
-}

package/src/process/scanSourceFiles.ts

@@ -1,73 +0,0 @@
-import { readFile } from 'fs/promises';
-import path from 'path';
-import { findMarkdownEntries } from '../findEntries';
-import { MetaData } from '../metadata';
-import { INPUT_DIR } from '../paths';
-import { sha256 } from '../utils/sha256';
-
-const extractLinksFromMarkdown = (content: string): string[] => {
-  const linkRegex = /\[.*?\]\((.*?)\)/g;
-  const links: string[] = [];
-  let match;
-  while ((match = linkRegex.exec(content)) !== null) {
-    links.push(match[1]);
-  }
-  return links;
-};
-
-/**
- * 扫描源文件
- */
-export async function scanSourceFiles(): Promise<void> {
-  console.log(`🔍 Scanning source directory...`);
-  const markdownFiles = await findMarkdownEntries(INPUT_DIR);
-  const hashes = new Set<string>();
-
-  for (const relativePath of markdownFiles) {
-    const fullPath = path.join(INPUT_DIR, relativePath);
-
-    try {
-      // 检查文件是否存在
-
-      const content = await readFile(fullPath, 'utf-8'); // 确保文件可读
-
-      const hash = sha256(content);
-      const links = extractLinksFromMarkdown(content);
-      console.info(`  - Found file: ${relativePath} (hash: ${hash})`);
-      console.info(`    Links: ${links.join(', ') || 'None'}`);
-
-      hashes.add(hash);
-
-      const metaWithSameHash = MetaData.files.find(f => f.hash === hash);
-      if (metaWithSameHash) {
-        metaWithSameHash.path = relativePath;
-        metaWithSameHash.links = links;
-      } else {
-        // 如果没有相同哈希的元数据，则添加一个新的占位符
-        MetaData.files.push({
-          hash,
-          path: relativePath,
-          links,
-        });
-      }
-    } catch (error) {
-      console.warn(`⚠️ File not found or inaccessible: ${fullPath}`, error);
-    }
-  }
-  // 移除不再存在的文件元数据
-  MetaData.files = MetaData.files.filter(f => hashes.has(f.hash));
-  // 按路径降序排序 (通常外层目录优先)
-  MetaData.files.sort(
-    (a, b) =>
-      // 第一级按目录排序
-      path.dirname(a.path).localeCompare(path.dirname(b.path)) ||
-      // 第二级按文件名排序
-      a.path.localeCompare(b.path)
-  );
-
-  console.log(`✅ Found ${MetaData.files.length} Markdown files`);
-
-  if (MetaData.files.length === 0) {
-    console.warn(`⚠️ No Markdown files found in ${INPUT_DIR}`);
-  }
-}