fix: allow hooking into script content

Author: harlan-zw

CLAUDE.md

@@ -70,7 +70,7 @@ The plugin system allows you to customize HTML to Markdown conversion by hooking
 
 - `beforeNodeProcess`: Called before any node processing, can skip nodes
 - `onNodeEnter`: Called when entering an element node
-- `onNodeExit`: Called when exiting an element node  
+- `onNodeExit`: Called when exiting an element node
 - `processTextNode`: Called for each text node
 - `processAttributes`: Called to process element attributes
 
@@ -88,7 +88,7 @@ export function myPlugin() {
         return '**Custom content:** '
       }
     },
-    
+
     processTextNode(textNode) {
       // Transform text content
       if (textNode.value.includes('TODO')) {
@@ -104,14 +104,14 @@ export function myPlugin() {
 ```typescript
 export function headerExtractPlugin() {
   const headers: string[] = []
-  
+
   return createPlugin({
     onNodeEnter(element) {
       if (['h1', 'h2', 'h3', 'h4', 'h5', 'h6'].includes(element.tagName)) {
         // Will collect text in processTextNode
       }
     },
-    
+
     processTextNode(textNode) {
       const parent = textNode.parent
       if (parent && parent.tagName?.match(/^h[1-6]$/)) {
@@ -129,13 +129,13 @@ export function adBlockPlugin() {
   return createPlugin({
     beforeNodeProcess(event) {
       const { node } = event
-      
+
       if (node.type === ELEMENT_NODE) {
         const element = node as ElementNode
-        
+
         // Skip ads and promotional content
-        if (element.attributes?.class?.includes('ad') || 
-            element.attributes?.id?.includes('promo')) {
+        if (element.attributes?.class?.includes('ad')
+          || element.attributes?.id?.includes('promo')) {
           return { skip: true }
         }
       }
@@ -180,4 +180,3 @@ export function adBlockPlugin() {
 - Real-world test fixtures in `test/fixtures/` (GitHub, Wikipedia HTML)
 - Template tests for complex HTML structures (navigation, tables, etc.)
 - Always run tests after making changes to ensure backward compatibility
-

README.md

@@ -143,7 +143,7 @@ The plugin system allows you to customize HTML to Markdown conversion by hooking
 #### Plugin Hooks
 
 - `beforeNodeProcess`: Called before any node processing, can skip nodes
-- `onNodeEnter`: Called when entering an element node  
+- `onNodeEnter`: Called when entering an element node
 - `onNodeExit`: Called when exiting an element node
 - `processTextNode`: Called for each text node
 - `processAttributes`: Called to process element attributes
@@ -153,23 +153,23 @@ The plugin system allows you to customize HTML to Markdown conversion by hooking
 Use `createPlugin()` to create a plugin with type safety:
 
 ```ts
+import type { ElementNode, TextNode } from 'mdream'
 import { htmlToMarkdown } from 'mdream'
 import { createPlugin } from 'mdream/plugins'
-import type { ElementNode, TextNode } from 'mdream'
 
 const myPlugin = createPlugin({
   onNodeEnter(node: ElementNode): string | undefined {
     if (node.name === 'h1') {
       return '🔥 '
     }
   },
-  
-  processTextNode(textNode: TextNode): { content: string; skip: boolean } | undefined {
+
+  processTextNode(textNode: TextNode): { content: string, skip: boolean } | undefined {
     // Transform text content
     if (textNode.parent?.attributes?.id === 'highlight') {
-      return { 
-        content: `**${textNode.value}**`, 
-        skip: false 
+      return {
+        content: `**${textNode.value}**`,
+        skip: false
       }
     }
   }
@@ -183,19 +183,19 @@ const markdown: string = htmlToMarkdown(html, { plugins: [myPlugin] })
 #### Example: Content Filter Plugin
 
 ```ts
+import type { ElementNode, NodeEvent } from 'mdream'
 import { ELEMENT_NODE } from 'mdream'
 import { createPlugin } from 'mdream/plugins'
-import type { NodeEvent, ElementNode } from 'mdream'
 
 const adBlockPlugin = createPlugin({
   beforeNodeProcess(event: NodeEvent): { skip: boolean } | undefined {
     const { node } = event
-    
+
     if (node.type === ELEMENT_NODE && node.name === 'div') {
       const element = node as ElementNode
       // Skip ads and promotional content
-      if (element.attributes?.class?.includes('ad') || 
-          element.attributes?.id?.includes('promo')) {
+      if (element.attributes?.class?.includes('ad')
+        || element.attributes?.id?.includes('promo')) {
         return { skip: true }
       }
     }
@@ -208,8 +208,8 @@ const adBlockPlugin = createPlugin({
 Extract specific elements and their content during HTML processing for data analysis or content discovery:
 
 ```ts
-import { extractionPlugin, htmlToMarkdown } from 'mdream'
 import type { ExtractedElement } from 'mdream/plugins'
+import { extractionPlugin, htmlToMarkdown } from 'mdream'
 
 const html: string = `
   <article>

src/markdown.ts

@@ -93,6 +93,12 @@ export function processHtmlEventToMarkdown(
           textNode.value = pluginResult.content
         }
       }
+
+      // Skip text nodes that are excluded from markdown output (e.g., script/style content)
+      if (textNode.excludedFromMarkdown) {
+        return
+      }
+
       // Skip leading spaces after newlines
       if (textNode.value === ' ' && lastChar === '\n') {
         return

src/parser.ts

@@ -280,10 +280,13 @@ function processTextBuffer(textBuffer: string, state: MdreamProcessingState, han
   const containsWhitespace = state.textBufferContainsWhitespace
   state.textBufferContainsNonWhitespace = false
   state.textBufferContainsWhitespace = false
-  if (!state.currentNode || state.currentNode?.tagHandler?.excludesTextNodes) {
+  if (!state.currentNode) {
     return
   }
 
+  // // Check if this element excludes text nodes from markdown output
+  const excludesTextNodes = state.currentNode?.tagHandler?.excludesTextNodes
+
   // Check if parent is a <pre> tag to handle whitespace properly
   const inPreTag = state.depthMap[TAG_PRE] > 0
 
@@ -325,6 +328,7 @@ function processTextBuffer(textBuffer: string, state: MdreamProcessingState, han
     index: state.currentNode.currentWalkIndex!++,
     depth: state.depth,
     containsWhitespace,
+    excludedFromMarkdown: excludesTextNodes,
   }
 
   for (const parent of parentsToIncrement) {

src/types.ts

@@ -93,6 +93,8 @@ export interface TextNode extends Node {
   value: string
   /** Custom data added by plugins */
   context?: Record<string, any>
+  /** Whether this text node should be excluded from markdown output (for script/style elements) */
+  excludedFromMarkdown?: boolean
 }
 
 /**