refactor(Markdown): remove rehype-sanitize and implement custom element filtering

- Removed rehype-sanitize dependency and its related configuration.
- Introduced ALLOWED_ELEMENTS and DISALLOWED_ELEMENTS for custom HTML element filtering.
- Updated rehypePlugins logic to conditionally apply plugins based on message content.
- Added encodeHTML utility function for HTML entity encoding.

package.json

@@ -190,7 +190,6 @@
     "rehype-katex": "^7.0.1",
     "rehype-mathjax": "^7.0.0",
     "rehype-raw": "^7.0.0",
-    "rehype-sanitize": "^6.0.0",
     "remark-cjk-friendly": "^1.1.0",
     "remark-gfm": "^4.0.0",
     "remark-math": "^6.0.0",

src/renderer/src/pages/home/Markdown/Markdown.tsx

@@ -8,16 +8,14 @@ import type { Message } from '@renderer/types'
 import { parseJSON } from '@renderer/utils'
 import { escapeBrackets, removeSvgEmptyLines, withGeminiGrounding } from '@renderer/utils/formats'
 import { findCitationInChildren } from '@renderer/utils/markdown'
-import { sanitizeSchema } from '@renderer/utils/markdown'
 import { isEmpty } from 'lodash'
 import { type FC, useMemo } from 'react'
 import { useTranslation } from 'react-i18next'
 import ReactMarkdown, { type Components } from 'react-markdown'
 import rehypeKatex from 'rehype-katex'
-// @ts-ignore next-line
+// @ts-ignore rehype-mathjax is not typed
 import rehypeMathjax from 'rehype-mathjax'
 import rehypeRaw from 'rehype-raw'
-import rehypeSanitize from 'rehype-sanitize'
 import remarkCjkFriendly from 'remark-cjk-friendly'
 import remarkGfm from 'remark-gfm'
 import remarkMath from 'remark-math'
@@ -26,6 +24,10 @@ import CodeBlock from './CodeBlock'
 import ImagePreview from './ImagePreview'
 import Link from './Link'
 
+const ALLOWED_ELEMENTS =
+  /<(style|p|div|span|b|i|strong|em|ul|ol|li|table|tr|td|th|thead|tbody|h[1-6]|blockquote|pre|code|br|hr|svg|path|circle|rect|line|polyline|polygon|text|g|defs|title|desc|tspan|sub|sup)/i
+const DISALLOWED_ELEMENTS = ['iframe']
+
 interface Props {
   message: Message
 }
@@ -43,9 +45,12 @@ const Markdown: FC<Props> = ({ message }) => {
     return removeSvgEmptyLines(escapeBrackets(content))
   }, [message, t])
 
+  const rehypeMath = useMemo(() => (mathEngine === 'KaTeX' ? rehypeKatex : rehypeMathjax), [mathEngine])
+
   const rehypePlugins = useMemo(() => {
-    return [rehypeRaw, [rehypeSanitize, sanitizeSchema], mathEngine === 'KaTeX' ? rehypeKatex : rehypeMathjax]
-  }, [mathEngine])
+    const hasElements = ALLOWED_ELEMENTS.test(messageContent)
+    return hasElements ? [rehypeRaw, rehypeMath] : [rehypeMath]
+  }, [messageContent, rehypeMath])
 
   const components = useMemo(() => {
     const baseComponents = {
@@ -71,6 +76,7 @@ const Markdown: FC<Props> = ({ message }) => {
       remarkPlugins={remarkPlugins}
       className="markdown"
       components={components}
+      disallowedElements={DISALLOWED_ELEMENTS}
       remarkRehypeOptions={{
         footnoteLabel: t('common.footnotes'),
         footnoteLabelTagName: 'h4',

src/renderer/src/utils/__tests__/markdown.test.ts

@@ -1,12 +1,6 @@
 import { describe, expect, it } from 'vitest'
 
-import {
-  convertMathFormula,
-  findCitationInChildren,
-  MARKDOWN_ALLOWED_TAGS,
-  removeTrailingDoubleSpaces,
-  sanitizeSchema
-} from '../markdown'
+import { convertMathFormula, findCitationInChildren, removeTrailingDoubleSpaces } from '../markdown'
 
 describe('markdown', () => {
   describe('findCitationInChildren', () => {
@@ -72,27 +66,6 @@ describe('markdown', () => {
     })
   })
 
-  describe('markdown configuration constants', () => {
-    it('MARKDOWN_ALLOWED_TAGS contains expected tags', () => {
-      expect(MARKDOWN_ALLOWED_TAGS).toContain('p')
-      expect(MARKDOWN_ALLOWED_TAGS).toContain('div')
-      expect(MARKDOWN_ALLOWED_TAGS).toContain('code')
-      expect(MARKDOWN_ALLOWED_TAGS).toContain('svg')
-      expect(MARKDOWN_ALLOWED_TAGS.length).toBeGreaterThan(10)
-    })
-
-    it('sanitizeSchema contains proper configuration', () => {
-      expect(sanitizeSchema.tagNames).toBe(MARKDOWN_ALLOWED_TAGS)
-      expect(sanitizeSchema.attributes).toHaveProperty('*')
-      expect(sanitizeSchema.attributes).toHaveProperty('svg')
-      expect(sanitizeSchema.attributes).toHaveProperty('a')
-    })
-
-    it('sanitizeSchema matches snapshot', () => {
-      expect(sanitizeSchema).toMatchSnapshot()
-    })
-  })
-
   describe('convertMathFormula', () => {
     it('should convert LaTeX block delimiters to $$$$', () => {
       // 验证将 LaTeX 块分隔符转换为 $$$$

src/renderer/src/utils/markdown.ts

@@ -1,67 +1,3 @@
-export const MARKDOWN_ALLOWED_TAGS = [
-  'style',
-  'p',
-  'div',
-  'span',
-  'b',
-  'i',
-  'strong',
-  'em',
-  'ul',
-  'ol',
-  'li',
-  'table',
-  'tr',
-  'td',
-  'th',
-  'thead',
-  'tbody',
-  'h1',
-  'h2',
-  'h3',
-  'h4',
-  'h5',
-  'h6',
-  'blockquote',
-  'pre',
-  'code',
-  'br',
-  'hr',
-  'svg',
-  'path',
-  'circle',
-  'rect',
-  'line',
-  'polyline',
-  'polygon',
-  'text',
-  'g',
-  'defs',
-  'title',
-  'desc',
-  'tspan',
-  'sub',
-  'sup'
-]
-
-// rehype-sanitize配置
-export const sanitizeSchema = {
-  tagNames: MARKDOWN_ALLOWED_TAGS,
-  attributes: {
-    '*': ['className', 'style', 'id', 'title'],
-    svg: ['viewBox', 'width', 'height', 'xmlns', 'fill', 'stroke'],
-    path: ['d', 'fill', 'stroke', 'strokeWidth', 'strokeLinecap', 'strokeLinejoin'],
-    circle: ['cx', 'cy', 'r', 'fill', 'stroke'],
-    rect: ['x', 'y', 'width', 'height', 'fill', 'stroke'],
-    line: ['x1', 'y1', 'x2', 'y2', 'stroke'],
-    polyline: ['points', 'fill', 'stroke'],
-    polygon: ['points', 'fill', 'stroke'],
-    text: ['x', 'y', 'fill', 'textAnchor', 'dominantBaseline'],
-    g: ['transform', 'fill', 'stroke'],
-    a: ['href', 'target', 'rel']
-  }
-}
-
 // 更彻底的查找方法，递归搜索所有子元素
 export const findCitationInChildren = (children) => {
   if (!children) return null
@@ -107,3 +43,21 @@ export function removeTrailingDoubleSpaces(markdown: string): string {
   // 使用正则表达式匹配末尾的两个空格，并替换为空字符串
   return markdown.replace(/ {2}$/gm, '')
 }
+
+/**
+ * HTML实体编码辅助函数
+ * @param str 输入字符串
+ * @returns string 编码后的字符串
+ */
+export const encodeHTML = (str: string) => {
+  return str.replace(/[&<>"']/g, (match) => {
+    const entities: { [key: string]: string } = {
+      '&': '&amp;',
+      '<': '&lt;',
+      '>': '&gt;',
+      '"': '&quot;',
+      "'": '&apos;'
+    }
+    return entities[match]
+  })
+}

yarn.lock

@@ -4338,7 +4338,6 @@ __metadata:
     rehype-katex: "npm:^7.0.1"
     rehype-mathjax: "npm:^7.0.0"
     rehype-raw: "npm:^7.0.0"
-    rehype-sanitize: "npm:^6.0.0"
     remark-cjk-friendly: "npm:^1.1.0"
     remark-gfm: "npm:^4.0.0"
     remark-math: "npm:^6.0.0"
@@ -9104,17 +9103,6 @@ __metadata:
   languageName: node
   linkType: hard
 
-"hast-util-sanitize@npm:^5.0.0":
-  version: 5.0.2
-  resolution: "hast-util-sanitize@npm:5.0.2"
-  dependencies:
-    "@types/hast": "npm:^3.0.0"
-    "@ungap/structured-clone": "npm:^1.0.0"
-    unist-util-position: "npm:^5.0.0"
-  checksum: 10c0/20951652078a8c21341c1c9a84f90015b2ba01cc41fa16772f122c65cda26a7adb0501fdeba5c8e37e40e2632447e8fe455d0dd2dc27d39663baacca76f2ecb6
-  languageName: node
-  linkType: hard
-
 "hast-util-to-html@npm:^9.0.5":
   version: 9.0.5
   resolution: "hast-util-to-html@npm:9.0.5"
@@ -14735,16 +14723,6 @@ __metadata:
   languageName: node
   linkType: hard
 
-"rehype-sanitize@npm:^6.0.0":
-  version: 6.0.0
-  resolution: "rehype-sanitize@npm:6.0.0"
-  dependencies:
-    "@types/hast": "npm:^3.0.0"
-    hast-util-sanitize: "npm:^5.0.0"
-  checksum: 10c0/43d6c056e63c994cf56e5ee0e157052d2030dc5ac160845ee494af9a26e5906bf5ec5af56c7d90c99f9c4dc0091e45a48a168618135fb6c64a76481ad3c449e9
-  languageName: node
-  linkType: hard
-
 "remark-cjk-friendly@npm:^1.1.0":
   version: 1.1.0
   resolution: "remark-cjk-friendly@npm:1.1.0"