hololake-system-architecture/product-source/hololake-desktop/electron/folder-import.ts

import fs from 'fs';
import path from 'path';
import simpleGit from 'simple-git';

const DOCUMENT_EXTENSIONS = new Set(['.md', '.markdown', '.txt', '.csv', '.json', '.yaml', '.yml']);
const ASSET_EXTENSIONS = new Set(['.png', '.jpg', '.jpeg', '.gif', '.webp', '.svg']);
const IGNORED_DIRECTORIES = new Set(['.git', 'node_modules', '.idea', '.vscode', '__macosx']);
const MAX_FILES = 1000;
const MAX_FILE_BYTES = 10 * 1024 * 1024;

export interface FolderImportResult {
  cancelled: boolean;
  sourceName?: string;
  destination?: string;
  imported: number;
  assets: number;
  skipped: number;
  failed: string[];
  firstDocument?: string;
}

function safeSegment(value: string): string {
  const cleaned = value.normalize('NFC').replace(/[\\/:*?"<>|\u0000-\u001f]/g, '-').trim();
  return cleaned || '未命名文件夹';
}

function markdownFrontmatter(title: string, source: string): string {
  const now = new Date().toISOString();
  return `---\ntitle: ${JSON.stringify(title)}\nauthor: HoloLake Importer\ncreatedAt: ${now}\nupdatedAt: ${now}\nsource: ${JSON.stringify(source)}\ntags:\n  - imported\n---\n\n`;
}

function csvToMarkdown(input: string): string {
  const rows = input
    .split(/\r?\n/)
    .filter(Boolean)
    .map(line => line.split(',').map(cell => cell.trim().replace(/\|/g, '\\|')));
  if (!rows.length) return '_空表格_\n';
  const width = Math.max(...rows.map(row => row.length));
  const padded = rows.map(row => [...row, ...Array(Math.max(0, width - row.length)).fill('')]);
  return [
    `| ${padded[0].join(' | ')} |`,
    `| ${Array(width).fill('---').join(' | ')} |`,
    ...padded.slice(1).map(row => `| ${row.join(' | ')} |`),
  ].join('\n') + '\n';
}

function convertDocument(sourcePath: string, relativePath: string): { targetRelative: string; content: string } {
  const extension = path.extname(sourcePath).toLowerCase();
  const raw = fs.readFileSync(sourcePath, 'utf8');
  const title = path.basename(sourcePath, extension);
  const targetRelative = extension === '.md'
    ? relativePath
    : relativePath.slice(0, -extension.length) + '.md';

  if (extension === '.md' || extension === '.markdown') {
    return {
      targetRelative: targetRelative.replace(/\.markdown$/i, '.md'),
      content: raw,
    };
  }
  if (extension === '.txt') {
    return { targetRelative, content: `${markdownFrontmatter(title, relativePath)}# ${title}\n\n${raw}\n` };
  }
  if (extension === '.csv') {
    return { targetRelative, content: `${markdownFrontmatter(title, relativePath)}# ${title}\n\n${csvToMarkdown(raw)}` };
  }
  const language = extension === '.json' ? 'json' : 'yaml';
  return {
    targetRelative,
    content: `${markdownFrontmatter(title, relativePath)}# ${title}\n\n\`\`\`${language}\n${raw}\n\`\`\`\n`,
  };
}

function walk(root: string, current: string, files: string[]): void {
  if (files.length >= MAX_FILES) return;
  for (const entry of fs.readdirSync(current, { withFileTypes: true })) {
    if (entry.name.startsWith('._') || entry.name === '.DS_Store') continue;
    const fullPath = path.join(current, entry.name);
    if (entry.isSymbolicLink()) continue;
    if (entry.isDirectory()) {
      if (!entry.name.startsWith('.') && !IGNORED_DIRECTORIES.has(entry.name.toLowerCase())) {
        walk(root, fullPath, files);
      }
      continue;
    }
    files.push(fullPath);
    if (files.length >= MAX_FILES) return;
  }
}

export async function importKnowledgeFolder(sourcePath: string, repoPath: string): Promise<FolderImportResult> {
  const sourceRoot = fs.realpathSync(sourcePath);
  const sourceName = safeSegment(path.basename(sourceRoot));
  const docsRoot = path.join(repoPath, 'docs');
  const importRoot = path.join(docsRoot, '导入');
  let destinationName = sourceName;
  let destinationRoot = path.join(importRoot, destinationName);
  if (fs.existsSync(destinationRoot)) {
    const stamp = new Date().toISOString().replace(/[-:]/g, '').slice(0, 13);
    destinationName = `${sourceName}-${stamp}`;
    destinationRoot = path.join(importRoot, destinationName);
  }
  fs.mkdirSync(destinationRoot, { recursive: true });

  const candidates: string[] = [];
  walk(sourceRoot, sourceRoot, candidates);
  const result: FolderImportResult = {
    cancelled: false,
    sourceName,
    destination: `导入/${destinationName}`,
    imported: 0,
    assets: 0,
    skipped: 0,
    failed: [],
  };

  for (const sourceFile of candidates) {
    const relative = path.relative(sourceRoot, sourceFile);
    const safeRelative = relative.split(path.sep).map(safeSegment).join(path.sep);
    const extension = path.extname(sourceFile).toLowerCase();
    try {
      if (fs.statSync(sourceFile).size > MAX_FILE_BYTES) {
        result.skipped += 1;
        continue;
      }
      if (DOCUMENT_EXTENSIONS.has(extension)) {
        const converted = convertDocument(sourceFile, safeRelative);
        const target = path.join(destinationRoot, converted.targetRelative);
        fs.mkdirSync(path.dirname(target), { recursive: true });
        fs.writeFileSync(target, converted.content, 'utf8');
        const docPath = path.posix.join('导入', destinationName, converted.targetRelative.split(path.sep).join('/'));
        result.firstDocument ||= docPath;
        result.imported += 1;
      } else if (ASSET_EXTENSIONS.has(extension)) {
        const target = path.join(destinationRoot, safeRelative);
        fs.mkdirSync(path.dirname(target), { recursive: true });
        fs.copyFileSync(sourceFile, target);
        result.assets += 1;
      } else {
        result.skipped += 1;
      }
    } catch (error) {
      result.failed.push(`${relative}: ${error instanceof Error ? error.message : String(error)}`);
    }
  }

  if (result.imported === 0 && result.assets === 0) {
    fs.rmSync(destinationRoot, { recursive: true, force: true });
    return result;
  }

  const git = simpleGit(repoPath);
  await git.add(path.relative(repoPath, destinationRoot));
  await git.commit(`import: 导入本地文件夹 ${sourceName}`);
  return result;
}