refactor(web): adopt modular Vue architecture and isolate bearer sessions

2026-10-25 20:47:32 +08:00
1 parent 1b2b2b12ff
commit 3cd641a380
@@ -0,0 +1,84 @@
+/* eslint-disable no-control-regex -- 粘贴链接必须识别并移除协议中的控制字符。 */
+import TurndownService from 'turndown'
+
+export async function htmlToMarkdown(html: string, importImage: (url: string) => Promise<string>, onError: (error: unknown) => void): Promise<string> {
+  // template 中的外部资源和脚本保持惰性，不让剪贴板 HTML 进入活动页面。
+  const template = document.createElement('template')
+  template.innerHTML = html
+  const fragment = template.content
+  const base = fragment.querySelector('base[href]')?.getAttribute('href')
+  fragment.querySelectorAll('script,style,iframe,object,embed,template,meta,link,base').forEach(node => node.remove())
+  if (base) {
+    for (const node of fragment.querySelectorAll('a[href],img[src]')) {
+      const attribute = node.tagName === 'A' ? 'href' : 'src'
+      try {
+        node.setAttribute(attribute, new URL(node.getAttribute(attribute)!, base).href)
+      } catch (error) {
+        onError(new Error('来源基址无效，已保留原链接', { cause: error }))
+      }
+    }
+  }
+
+  const images = new Map<string, HTMLImageElement[]>()
+  for (const image of fragment.querySelectorAll('img')) {
+    let src = image.getAttribute('src') ?? ''
+    if (src.startsWith('//')) {
+      src = `https:${src}`; image.setAttribute('src', src)
+    }
+    if (!/^https?:\/\//i.test(src)) continue
+    const group = images.get(src) ?? []
+    group.push(image)
+    images.set(src, group)
+  }
+  const sources = [...images]
+  // 网页常有多张图片；限制并行数量，避免一次粘贴占满浏览器连接和服务端下载内存。
+  for (let offset = 0; offset < sources.length; offset += 3) {
+    await Promise.all(sources.slice(offset, offset + 3).map(async ([src, nodes]) => {
+      try {
+        const link = await importImage(src)
+        for (const node of nodes) node.setAttribute('src', link)
+      } catch (error) {
+        onError(error)
+      }
+    }))
+  }
+
+  const service = new TurndownService({ headingStyle: 'atx', bulletListMarker: '-', codeBlockStyle: 'fenced' })
+  service.addRule('strikethrough', { filter: node => ['DEL', 'S', 'STRIKE'].includes(node.tagName), replacement: content => `~~${content}~~` })
+  service.addRule('table', {
+    filter: 'table',
+    replacement: (_content, node) => {
+      const rows = [...node.querySelectorAll('tr')].filter(row => row.closest('table') === node).map(row =>
+        [...row.cells].map(cell =>
+          service.turndown(cell).replace(/\|/g, '\\|').replace(/\n+/g, ' '),
+        ),
+      )
+      if (!rows.length) return ''
+      const columns = Math.max(...rows.map(row => row.length))
+      if (!columns) return ''
+      const line = (cells: string[]) => `| ${Array.from({ length: columns }, (_, index) => cells[index] ?? '').join(' | ')} |`
+      return `\n\n${[line(rows[0]), line(Array.from<string>({ length: columns }).fill('---')), ...rows.slice(1).map(line)].join('\n')}\n\n`
+    },
+  })
+  // 链接目标用尖括号包裹，避免 URL 中的括号截断 Markdown；不保留可执行协议。
+  service.addRule('link', {
+    filter: 'a',
+    replacement: (content, node) => {
+      const href = node.getAttribute('href') ?? ''
+      return href && !/^[\u0000-\u0020\u00A0\u1680\u2000-\u200A\u2028\u2029\u202F\u205F\u3000\uFEFF]*(?:javascript|vbscript|data|file):/i.test(href.replace(/[\u0000-\u0020]/g, ''))
+        ? `[${content}](<${href.replace(/</g, '%3C').replace(/>/g, '%3E')}>)`
+        : content
+    },
+  })
+  service.addRule('image', {
+    filter: 'img',
+    replacement: (_content, node) => {
+      const src = node.getAttribute('src') ?? ''
+      const alt = (node.getAttribute('alt') ?? '').replace(/[[\]\\]/g, '\\$&').replace(/\n/g, ' ')
+      return src && !/^(?:javascript|vbscript|file):/i.test(src.replace(/[\u0000-\u0020]/g, ''))
+        ? `![${alt}](<${src.replace(/</g, '%3C').replace(/>/g, '%3E')}>)`
+        : alt
+    },
+  })
+  return service.turndown(fragment)
+}