dsh-plugin-office-markdown 1.2.1

package/lib/fallback-node.js

@@ -0,0 +1,334 @@
+/*!
+ * dsh-plugin-office-markdown — 纯 Node 兜底转换器
+ *
+ * 不依赖 Python、不联网、不调用任何 API：用 Node 内置的 zlib 解开 OOXML
+ * （.docx/.xlsx/.pptx 本质上是 zip + xml），再按 XML 结构抽取文字与表格，
+ * 生成 Markdown。保真度有限（无版式/图表/批注/图片/公式/OCR），但保证在
+ * “目标电脑既没有 Python 也没有网络”时仍能把文件读成 Markdown。
+ *
+ * 既可被 lib/convert.js 直接 import 调用，也可作为 CLI 使用：
+ *   node fallback-node.js --input a.xlsx --output out.md
+ */
+import fs from 'node:fs'
+import path from 'node:path'
+import zlib from 'node:zlib'
+import { pathToFileURL } from 'node:url'
+
+export const NODE_SUPPORTED = new Set(['.docx', '.docm', '.xlsx', '.xlsm', '.pptx', '.pptm'])
+
+const UNSUPPORTED = {
+  '.pdf': 'Node 兜底不支持 PDF（PDF 需要 MarkItDown 或本机 Python 兜底）。',
+  '.doc': '旧版二进制 .doc 需要 MarkItDown 转换，Node 兜底不支持。',
+  '.xls': '旧版二进制 .xls 需要 MarkItDown 转换，Node 兜底不支持。',
+  '.ppt': '旧版二进制 .ppt 需要 MarkItDown 转换，Node 兜底不支持。',
+  '.msg': 'Outlook .msg 需要 MarkItDown 转换，Node 兜底不支持。',
+  '.epub': 'EPUB 需要 MarkItDown 转换，Node 兜底不支持。',
+  '.odt': 'ODF/OpenDocument 格式需要 MarkItDown 转换，Node 兜底不支持。',
+  '.ods': 'ODF/OpenDocument 格式需要 MarkItDown 转换，Node 兜底不支持。',
+  '.odp': 'ODF/OpenDocument 格式需要 MarkItDown 转换，Node 兜底不支持。'
+}
+
+/* ---------- zip ---------- */
+
+const SIG_EOCD = 0x06054b50
+const SIG_CENTRAL = 0x02014b50
+
+function readZip(file) {
+  const buf = fs.readFileSync(file)
+  let eocd = -1
+  const floor = Math.max(0, buf.length - 66000)
+  for (let i = buf.length - 22; i >= floor; i--) {
+    if (buf.readUInt32LE(i) === SIG_EOCD) { eocd = i; break }
+  }
+  if (eocd < 0) throw new Error('无法解析为 zip 容器（文件可能已损坏或不是真正的 OOXML 文件）')
+  const count = buf.readUInt16LE(eocd + 10)
+  let off = buf.readUInt32LE(eocd + 16)
+  const entries = new Map()
+  for (let i = 0; i < count; i++) {
+    if (off + 46 > buf.length || buf.readUInt32LE(off) !== SIG_CENTRAL) break
+    const method = buf.readUInt16LE(off + 10)
+    const csize = buf.readUInt32LE(off + 20)
+    const nlen = buf.readUInt16LE(off + 28)
+    const elen = buf.readUInt16LE(off + 30)
+    const clen = buf.readUInt16LE(off + 32)
+    const lho = buf.readUInt32LE(off + 42)
+    const name = buf.toString('utf8', off + 46, off + 46 + nlen)
+    entries.set(name, { method, csize, lho })
+    off += 46 + nlen + elen + clen
+  }
+  return { buf, entries }
+}
+
+function readEntry(buf, entry) {
+  const nlen = buf.readUInt16LE(entry.lho + 26)
+  const elen = buf.readUInt16LE(entry.lho + 28)
+  const start = entry.lho + 30 + nlen + elen
+  const raw = buf.subarray(start, start + entry.csize)
+  if (entry.method === 0) return raw
+  if (entry.method === 8) return zlib.inflateRawSync(raw)
+  throw new Error('zip 压缩方式 ' + entry.method + ' 不受支持')
+}
+
+/* ---------- xml helpers ---------- */
+
+function unescapeXml(s) {
+  return String(s)
+    .replace(/&#x([0-9a-fA-F]+);/g, (_, h) => String.fromCodePoint(parseInt(h, 16)))
+    .replace(/&#(\d+);/g, (_, d) => String.fromCodePoint(Number(d)))
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&quot;/g, '"')
+    .replace(/&apos;/g, "'")
+    .replace(/&amp;/g, '&')
+}
+
+const TEXT_RE = /<(?:[A-Za-z0-9_]+:)?t(?:\s[^>]*)?>([\s\S]*?)<\/(?:[A-Za-z0-9_]+:)?t>/g
+
+function texts(xml) {
+  let out = ''
+  for (const m of String(xml).matchAll(TEXT_RE)) out += unescapeXml(m[1])
+  return out
+}
+
+function cell(text) {
+  return String(text).replace(/\|/g, '\\|').replace(/\s+/g, ' ').trim()
+}
+
+function mdTable(rows, opt) {
+  if (!rows.length) return ''
+  const width = Math.min(opt.maxTableCols, Math.max(...rows.map((r) => r.length)))
+  if (width <= 0) return ''
+  const lines = []
+  const limit = Math.min(rows.length, opt.maxRowsPerSheet)
+  for (let i = 0; i < limit; i++) {
+    const cells = []
+    for (let c = 0; c < width; c++) cells.push(rows[i][c] === undefined ? '' : cell(rows[i][c]))
+    lines.push('| ' + cells.join(' | ') + ' |')
+    if (i === 0) lines.push('| ' + cells.map(() => '---').join(' | ') + ' |')
+  }
+  if (rows.length > limit) lines.push('', `> 表格已截断：共 ${rows.length} 行，只显示前 ${limit} 行。`)
+  return lines.join('\n')
+}
+
+/* ---------- docx ---------- */
+
+function paraToMd(chunk) {
+  const text = texts(chunk).replace(/\s+$/, '')
+  if (!text.trim()) return ''
+  const style = (chunk.match(/<w:pStyle[^>]*w:val="([^"]*)"/) || [])[1] || ''
+  const heading = style.match(/^(?:Heading|heading|标题)\s*(\d+)$/)
+  if (heading) return '#'.repeat(Math.min(6, Number(heading[1]) || 1)) + ' ' + text
+  if (/^(?:Title|标题|Subtitle)$/i.test(style)) return '# ' + text
+  if (/<w:numPr\b/.test(chunk)) return '- ' + text
+  return text
+}
+
+function docxTableToMd(chunk, opt) {
+  const rows = []
+  for (const r of chunk.matchAll(/<w:tr\b[^>]*>([\s\S]*?)<\/w:tr>/g)) {
+    const cells = []
+    for (const c of r[1].matchAll(/<w:tc\b[^>]*>([\s\S]*?)<\/w:tc>/g)) cells.push(texts(c[1]))
+    if (cells.length) rows.push(cells)
+  }
+  return mdTable(rows, opt)
+}
+
+function docxToMarkdown(buf, entries, opt) {
+  const entry = entries.get('word/document.xml')
+  if (!entry) throw new Error('缺少 word/document.xml，可能不是有效的 .docx')
+  const xml = readEntry(buf, entry).toString('utf8')
+  const blockRe = /<w:tbl\b[\s\S]*?<\/w:tbl>|<w:p\b[^>]*\/>|<w:p\b[^>]*>[\s\S]*?<\/w:p>/g
+  const parts = []
+  for (const m of xml.matchAll(blockRe)) {
+    const chunk = m[0]
+    const piece = chunk.startsWith('<w:tbl') ? docxTableToMd(chunk, opt) : paraToMd(chunk)
+    if (piece && piece.trim()) parts.push(piece)
+  }
+  return parts.join('\n\n')
+}
+
+/* ---------- xlsx ---------- */
+
+function colIndex(ref) {
+  let n = 0
+  for (const ch of ref) {
+    const c = ch.charCodeAt(0)
+    if (c < 65 || c > 90) break
+    n = n * 26 + (c - 64)
+  }
+  return n - 1
+}
+
+function sheetOrder(buf, entries) {
+  const rels = {}
+  const relEntry = entries.get('xl/_rels/workbook.xml.rels')
+  if (relEntry) {
+    const xml = readEntry(buf, relEntry).toString('utf8')
+    for (const m of xml.matchAll(/<Relationship\b[^>]*>/g)) {
+      const id = (m[0].match(/Id="([^"]*)"/) || [])[1]
+      const target = (m[0].match(/Target="([^"]*)"/) || [])[1]
+      if (id && target) rels[id] = target.replace(/^\/?xl\//, '').replace(/^\.\//, '')
+    }
+  }
+  const out = []
+  const wbEntry = entries.get('xl/workbook.xml')
+  if (wbEntry) {
+    const xml = readEntry(buf, wbEntry).toString('utf8')
+    for (const m of xml.matchAll(/<sheet\b[^>]*>/g)) {
+      const name = (m[0].match(/name="([^"]*)"/) || [])[1] || ''
+      const rid = (m[0].match(/r:id="([^"]*)"/) || [])[1] || ''
+      const target = rels[rid]
+      if (target) out.push({ name: unescapeXml(name), target: 'xl/' + target })
+    }
+  }
+  if (!out.length) {
+    const names = [...entries.keys()]
+      .filter((n) => /^xl\/worksheets\/sheet\d+\.xml$/.test(n))
+      .sort((a, b) => Number(a.match(/(\d+)/)[1]) - Number(b.match(/(\d+)/)[1]))
+    names.forEach((n, i) => out.push({ name: 'Sheet' + (i + 1), target: n }))
+  }
+  return out
+}
+
+function xlsxToMarkdown(buf, entries, opt) {
+  const shared = []
+  const sharedEntry = entries.get('xl/sharedStrings.xml')
+  if (sharedEntry) {
+    const xml = readEntry(buf, sharedEntry).toString('utf8')
+    for (const m of xml.matchAll(/<si\b[^>]*>([\s\S]*?)<\/si>/g)) shared.push(texts(m[1]))
+  }
+  const parts = []
+  let budget = opt.maxCellsPerSheet
+  for (const sheet of sheetOrder(buf, entries)) {
+    const entry = entries.get(sheet.target)
+    if (!entry) continue
+    const xml = readEntry(buf, entry).toString('utf8')
+    const rows = []
+    for (const r of xml.matchAll(/<row\b[^>]*>([\s\S]*?)<\/row>/g)) {
+      if (rows.length >= opt.maxRowsPerSheet || budget <= 0) break
+      const arr = []
+      for (const c of r[1].matchAll(/<c\b([^>]*)>([\s\S]*?)<\/c>|<c\b([^>]*)\/>/g)) {
+        const attrs = c[1] || c[3] || ''
+        const inner = c[2] || ''
+        const type = (attrs.match(/t="([^"]*)"/) || [])[1] || ''
+        const ref = (attrs.match(/r="([A-Z]+)\d+"/) || [])[1] || ''
+        let value = ''
+        if (type === 's') {
+          const idx = (inner.match(/<v>([\s\S]*?)<\/v>/) || [])[1]
+          value = shared[Number(idx)] !== undefined ? shared[Number(idx)] : ''
+        } else if (type === 'inlineStr') {
+          value = texts(inner)
+        } else {
+          const v = (inner.match(/<v>([\s\S]*?)<\/v>/) || [])[1]
+          value = v === undefined ? '' : unescapeXml(v)
+        }
+        const at = ref ? colIndex(ref) : arr.length
+        if (at < 0) continue
+        while (arr.length < at) arr.push('')
+        arr[at] = value
+        budget--
+        if (budget <= 0) break
+      }
+      if (arr.length) rows.push(arr)
+    }
+    if (!rows.length) continue
+    const table = mdTable(rows, opt)
+    if (table) parts.push('## ' + (sheet.name || 'Sheet') + '\n\n' + table)
+  }
+  if (!parts.length) throw new Error('工作簿里没有可提取的单元格文本')
+  return parts.join('\n\n')
+}
+
+/* ---------- pptx ---------- */
+
+function pptxToMarkdown(buf, entries) {
+  const names = [...entries.keys()]
+    .filter((n) => /^ppt\/slides\/slide\d+\.xml$/.test(n))
+    .sort((a, b) => Number(a.match(/(\d+)/)[1]) - Number(b.match(/(\d+)/)[1]))
+  const parts = []
+  names.forEach((n, i) => {
+    const xml = readEntry(buf, entries.get(n)).toString('utf8')
+    const lines = []
+    for (const m of xml.matchAll(/<a:p\b[^>]*>([\s\S]*?)<\/a:p>/g)) {
+      const t = texts(m[1]).trim()
+      if (t) lines.push(t)
+    }
+    if (!lines.length) return
+    const body = lines.slice(1).map((l) => '- ' + l).join('\n')
+    parts.push(`## 幻灯片 ${i + 1}：${lines[0]}` + (body ? '\n\n' + body : ''))
+  })
+  if (!parts.length) throw new Error('演示文稿里没有可提取的文本')
+  return parts.join('\n\n')
+}
+
+/* ---------- entry ---------- */
+
+export const NODE_DEFAULTS = Object.freeze({
+  maxRowsPerSheet: 400,
+  maxTableCols: 24,
+  maxCellsPerSheet: 20000
+})
+
+/**
+ * 把 srcPath 指向的 OOXML 文件转成 Markdown 写入 dstPath（纯本地、无依赖）。
+ * @returns {{bytes:number, chars:number}}
+ */
+export function convertFileNode(srcPath, dstPath, options = {}) {
+  const opt = { ...NODE_DEFAULTS, ...options }
+  const ext = path.extname(srcPath).toLowerCase()
+  if (!NODE_SUPPORTED.has(ext)) {
+    throw new Error(UNSUPPORTED[ext] || `内置 Node 兜底转换器不支持 ${ext || '该'} 格式`)
+  }
+  const { buf, entries } = readZip(srcPath)
+  let body
+  if (ext === '.docx' || ext === '.docm') body = docxToMarkdown(buf, entries, opt)
+  else if (ext === '.xlsx' || ext === '.xlsm') body = xlsxToMarkdown(buf, entries, opt)
+  else body = pptxToMarkdown(buf, entries)
+
+  if (!body || !body.trim()) throw new Error('未能从文件中提取到任何文本内容')
+
+  const lines = [
+    '<!-- 由 dsh-plugin-office-markdown 内置 Node 兜底转换器生成（非 MarkItDown），版式/图表/批注/图片/公式等信息可能缺失，保真度有限。 -->',
+    `> 源文件：\`${path.basename(srcPath)}\``,
+    '> 转换器：内置兜底（fallback-node.js，纯 Node，无需 Python）',
+    '',
+    body,
+    '',
+    '---',
+    '',
+    '**转换提示**',
+    '',
+    '- 本转换完全在本地完成：未联网、未调用任何 API、未修改原文件。',
+    '- 纯 Node 解析 OOXML（zip + xml），不保留版式、图表、批注、图片、公式与扫描件 OCR。',
+    '- 如需更高保真度：有网络时用 `uvx markitdown`，或安装 `pip install "markitdown[all]"`。'
+  ]
+  const text = lines.join('\n')
+  fs.mkdirSync(path.dirname(dstPath), { recursive: true })
+  fs.writeFileSync(dstPath, text, 'utf8')
+  return { bytes: Buffer.byteLength(text, 'utf8'), chars: text.length }
+}
+
+if (process.argv[1] && pathToFileURL(process.argv[1]).href === import.meta.url) {
+  const argv = process.argv.slice(2)
+  const arg = (flag, fallback) => {
+    const i = argv.indexOf(flag)
+    return i >= 0 && argv[i + 1] !== undefined ? argv[i + 1] : fallback
+  }
+  const input = arg('--input')
+  const output = arg('--output')
+  if (!input || !output) {
+    console.error('usage: node fallback-node.js --input <src> --output <dst> [--max-rows N] [--max-cols N] [--max-cells N]')
+    process.exit(2)
+  }
+  try {
+    const result = convertFileNode(input, output, {
+      maxRowsPerSheet: Number(arg('--max-rows', NODE_DEFAULTS.maxRowsPerSheet)),
+      maxTableCols: Number(arg('--max-cols', NODE_DEFAULTS.maxTableCols)),
+      maxCellsPerSheet: Number(arg('--max-cells', NODE_DEFAULTS.maxCellsPerSheet))
+    })
+    console.log(`OK ${result.bytes} bytes -> ${output}`)
+  } catch (error) {
+    console.error(String((error && error.message) || error))
+    process.exit(2)
+  }
+}