@tkeron/html-parser 0.1.5 → 1.0.0

package/src/parser.ts

@@ -1,38 +1,14 @@
 import type { Token } from './tokenizer.js';
 import { TokenType } from './tokenizer.js';
-
-export interface ASTNode {
-  type: ASTNodeType;
-  tagName?: string;
-  attributes?: Record<string, string>;
-  children?: ASTNode[];
-  content?: string;
-  parent?: ASTNode;
-  isSelfClosing?: boolean;
-  position?: {
-    start: number;
-    end: number;
-    line: number;
-    column: number;
-  };
-}
-
-export enum ASTNodeType {
-  DOCUMENT = 'DOCUMENT',
-  ELEMENT = 'ELEMENT',
-  TEXT = 'TEXT',
-  COMMENT = 'COMMENT',
-  CDATA = 'CDATA',
-  DOCTYPE = 'DOCTYPE',
-  PROCESSING_INSTRUCTION = 'PROCESSING_INSTRUCTION'
-}
+import { createDocument, createElement, createTextNode, createComment, createDoctype, appendChild } from './dom-simulator.js';
 
 export interface ParserState {
   tokens: Token[];
   position: number;
   length: number;
-  stack: ASTNode[];
-  root: ASTNode;
+  stack: any[]; // DOM elements
+  root: any; // Document
+  insertionMode: InsertionMode;
   errors: ParseError[];
 }
 
@@ -44,6 +20,32 @@ export interface ParseError {
   severity: 'error' | 'warning';
 }
 
+export enum InsertionMode {
+  Initial = 'initial',
+  BeforeHtml = 'beforeHtml',
+  BeforeHead = 'beforeHead',
+  InHead = 'inHead',
+  AfterHead = 'afterHead',
+  InBody = 'inBody'
+}
+
+export enum ASTNodeType {
+  Document = 'document',
+  Element = 'element',
+  Text = 'text',
+  Comment = 'comment',
+  Doctype = 'doctype',
+  CDATA = 'cdata'
+}
+
+export interface ASTNode {
+  type: ASTNodeType;
+  tagName?: string;
+  value?: string;
+  attributes?: Record<string, string>;
+  children?: ASTNode[];
+}
+
 const VOID_ELEMENTS = new Set([
   'area', 'base', 'br', 'col', 'embed', 'hr', 'img', 'input',
   'link', 'meta', 'param', 'source', 'track', 'wbr'
@@ -57,7 +59,41 @@ const AUTO_CLOSE_RULES: Record<string, string[]> = {
   'li': ['li'],
   'dt': ['dt', 'dd'],
   'dd': ['dt', 'dd'],
+  'address': ['p'],
+  'article': ['p'],
+  'aside': ['p'],
+  'blockquote': ['p'],
+  'center': ['p'],
+  'details': ['p'],
+  'dialog': ['p'],
+  'dir': ['p'],
+  'div': ['p'],
+  'dl': ['p'],
+  'fieldset': ['p'],
+  'figcaption': ['p'],
+  'figure': ['p'],
+  'footer': ['p'],
+  'form': ['p'],
+  'h1': ['p'],
+  'h2': ['p'],
+  'h3': ['p'],
+  'h4': ['p'],
+  'h5': ['p'],
+  'h6': ['p'],
+  'header': ['p'],
+  'hgroup': ['p'],
+  'hr': ['p'],
+  'listing': ['p'],
+  'main': ['p'],
+  'menu': ['p'],
+  'nav': ['p'],
+  'ol': ['p'],
   'p': ['p'],
+  'pre': ['p'],
+  'section': ['p'],
+  'summary': ['p'],
+  'table': ['p'],
+  'ul': ['p'],
   'rt': ['rt', 'rp'],
   'rp': ['rt', 'rp'],
   'optgroup': ['optgroup'],
@@ -70,7 +106,7 @@ const AUTO_CLOSE_RULES: Record<string, string[]> = {
   'th': ['td', 'th']
 };
 
-export function parse(tokens: Token[]): ASTNode {
+export function parse(tokens: Token[]): any {
   const state = createParserState(tokens);
   
   while (state.position < state.length) {
@@ -84,21 +120,119 @@ export function parse(tokens: Token[]): ASTNode {
     advance(state);
   }
 
+  // Create implicit html, head, body if needed
+  if (state.root.childNodes && state.root.childNodes.length > 0) {
+    let hasHtml = false;
+    for (const child of state.root.childNodes) {
+      if (child.nodeType === 1 && child.tagName === 'HTML') {
+        hasHtml = true;
+        state.root.documentElement = child;
+        break;
+      }
+    }
+    if (!hasHtml) {
+      const html = createElement('html', {});
+      const head = createElement('head', {});
+      const body = createElement('body', {});
+      appendChild(html, head);
+      appendChild(html, body);
+      
+      const doctypes: any[] = [];
+      const children = [...state.root.childNodes];
+      for (const child of children) {
+        if (child.nodeType === 10) {
+          doctypes.push(child);
+        } else {
+          appendChild(body, child);
+        }
+      }
+      
+      state.root.childNodes = [];
+      for (const doctype of doctypes) {
+        doctype.parentNode = null;
+        appendChild(state.root, doctype);
+      }
+      appendChild(state.root, html);
+      state.root.documentElement = html;
+      state.root.head = head;
+      state.root.body = body;
+    }
+  }
+
   while (state.stack.length > 1) {
     const unclosedElement = state.stack.pop()!;
     const currentToken = getCurrentToken(state);
-    addError(state, `Unclosed tag: ${unclosedElement.tagName}`, currentToken?.position?.start || 0);
+    addError(state, `Unclosed tag: ${unclosedElement.tagName}`, currentToken?.position?.offset || 0);
   }
 
   return state.root;
 }
 
+export function domToAST(dom: any): ASTNode {
+  function convert(node: any): ASTNode | null {
+    if (!node) return null;
+    
+    if (node.nodeType === 9) {
+      const children: ASTNode[] = [];
+      if (node.childNodes) {
+        for (const child of node.childNodes) {
+          const converted = convert(child);
+          if (converted) children.push(converted);
+        }
+      }
+      return {
+        type: ASTNodeType.Document,
+        children
+      };
+    }
+    
+    if (node.nodeType === 1) {
+      const children: ASTNode[] = [];
+      if (node.childNodes) {
+        for (const child of node.childNodes) {
+          const converted = convert(child);
+          if (converted) children.push(converted);
+        }
+      }
+      const tagName = node.tagName?.toLowerCase();
+      return {
+        type: ASTNodeType.Element,
+        tagName,
+        attributes: node.attributes || {},
+        children,
+        isSelfClosing: VOID_ELEMENTS.has(tagName)
+      } as ASTNode & { isSelfClosing: boolean };
+    }
+    
+    if (node.nodeType === 3) {
+      return {
+        type: ASTNodeType.Text,
+        content: node.nodeValue || ''
+      } as ASTNode & { content: string };
+    }
+    
+    if (node.nodeType === 8) {
+      return {
+        type: ASTNodeType.Comment,
+        content: node.nodeValue || ''
+      } as ASTNode & { content: string };
+    }
+    
+    if (node.nodeType === 10) {
+      return {
+        type: ASTNodeType.Doctype,
+        content: node.name || 'html'
+      } as ASTNode & { content: string };
+    }
+    
+    return null;
+  }
+  
+  return convert(dom) || { type: ASTNodeType.Document, children: [] };
+}
+
 function createParserState(tokens: Token[]): ParserState {
-  const root: ASTNode = {
-    type: ASTNodeType.DOCUMENT,
-    children: [],
-    tagName: '#document'
-  };
+  const root = createDocument();
 
   return {
     tokens,
@@ -106,81 +240,223 @@ function createParserState(tokens: Token[]): ParserState {
     length: tokens.length,
     stack: [root],
     root,
+    insertionMode: InsertionMode.Initial,
     errors: []
   };
 }
 
 function parseToken(state: ParserState, token: Token): void {
-  switch (token.type) {
-    case TokenType.TAG_OPEN:
-      parseOpenTag(state, token);
+  switch (state.insertionMode) {
+    case InsertionMode.Initial:
+      parseTokenInInitialMode(state, token);
       break;
-    case TokenType.TAG_CLOSE:
-      parseCloseTag(state, token);
+    case InsertionMode.BeforeHtml:
+      parseTokenInBeforeHtmlMode(state, token);
       break;
-    case TokenType.TEXT:
-      parseText(state, token);
-      break;
-    case TokenType.COMMENT:
-      parseComment(state, token);
+    case InsertionMode.BeforeHead:
+      parseTokenInBeforeHeadMode(state, token);
       break;
-    case TokenType.CDATA:
-      parseCDATA(state, token);
+    case InsertionMode.InHead:
+      parseTokenInInHeadMode(state, token);
       break;
-    case TokenType.DOCTYPE:
-      parseDoctype(state, token);
+    case InsertionMode.AfterHead:
+      parseTokenInAfterHeadMode(state, token);
       break;
-    case TokenType.PROCESSING_INSTRUCTION:
-      parseProcessingInstruction(state, token);
+    case InsertionMode.InBody:
+      parseTokenInInBodyMode(state, token);
       break;
+    default:
+      parseTokenInInBodyMode(state, token); // fallback
   }
 }
 
-function parseOpenTag(state: ParserState, token: Token): void {
-  const tagName = token.value.toLowerCase();
-
-  handleAutoClosing(state, tagName);
-
-  const currentParent = getCurrentParent(state);
+function parseTokenInInitialMode(state: ParserState, token: Token): void {
+  if (token.type === TokenType.DOCTYPE) {
+    // TODO: Create DOCTYPE node
+    parseDoctype(state, token);
+    state.insertionMode = InsertionMode.BeforeHtml;
+  } else if (token.type === TokenType.COMMENT) {
+    parseComment(state, token);
+  } else if (token.type === TokenType.TEXT && token.value.trim() === '') {
+    // Ignore whitespace
+  } else {
+    // No DOCTYPE, create implicit DOCTYPE and switch to BeforeHtml
+    const doctype = createDoctype('html');
+    appendChild(state.root, doctype);
+    state.insertionMode = InsertionMode.BeforeHtml;
+    parseToken(state, token); // Re-parse in new mode
+  }
+}
 
-  const element: ASTNode = {
-    type: ASTNodeType.ELEMENT,
-    tagName,
-    attributes: token.attributes || {},
-    children: [],
-    parent: currentParent,
-    isSelfClosing: token.isSelfClosing || VOID_ELEMENTS.has(tagName),
-    position: token.position
-  };
+function parseTokenInBeforeHtmlMode(state: ParserState, token: Token): void {
+  if (token.type === TokenType.TAG_OPEN && token.value.toLowerCase() === 'html') {
+    const html = createElement('html', token.attributes || {});
+    appendChild(state.root, html);
+    state.root.documentElement = html;
+    state.stack.push(html);
+    state.insertionMode = InsertionMode.BeforeHead;
+  } else if (token.type === TokenType.COMMENT) {
+    parseComment(state, token);
+  } else if (token.type === TokenType.DOCTYPE) {
+    // Ignore
+  } else if (token.type === TokenType.TEXT && token.value.trim() === '') {
+    // Ignore whitespace
+  } else {
+    const html = createElement('html', {});
+    appendChild(state.root, html);
+    state.root.documentElement = html;
+    state.stack.push(html);
+    state.insertionMode = InsertionMode.BeforeHead;
+    parseToken(state, token);
+  }
+}
 
-  if (currentParent.children) {
-    currentParent.children.push(element);
+function parseTokenInBeforeHeadMode(state: ParserState, token: Token): void {
+  if (token.type === TokenType.TAG_OPEN && token.value.toLowerCase() === 'head') {
+    const head = createElement('head', token.attributes || {});
+    appendChild(getCurrentParent(state), head);
+    state.root.head = head;
+    state.stack.push(head);
+    state.insertionMode = InsertionMode.InHead;
+  } else if (token.type === TokenType.COMMENT) {
+    parseComment(state, token);
+  } else if (token.type === TokenType.TEXT && token.value.trim() === '') {
+    // Ignore whitespace
+  } else {
+    const head = createElement('head', {});
+    appendChild(getCurrentParent(state), head);
+    state.root.head = head;
+    state.stack.push(head);
+    state.insertionMode = InsertionMode.InHead;
+    parseToken(state, token);
   }
+}
 
-  if (!element.isSelfClosing) {
+function parseOpenTag(state: ParserState, token: Token): void {
+  const tagName = token.value.toLowerCase();
+  const currentParent = getCurrentParent(state);
+  const element = createElement(tagName, token.attributes || {});
+  appendChild(currentParent, element);
+  
+  if (!token.isSelfClosing && !VOID_ELEMENTS.has(tagName)) {
     state.stack.push(element);
   }
 }
 
-function parseCloseTag(state: ParserState, token: Token): void {
-  const tagName = token.value.toLowerCase();
+function parseTokenInInHeadMode(state: ParserState, token: Token): void {
+  const currentElement = getCurrentElement(state);
+  const currentTagName = currentElement?.tagName?.toLowerCase();
   
-  let found = false;
-  for (let i = state.stack.length - 1; i >= 0; i--) {
-    const element = state.stack[i]!;
-    if (element.tagName === tagName) {
-      while (state.stack.length > i + 1) {
-        const unclosedElement = state.stack.pop()!;
-        addError(state, `Unclosed tag: ${unclosedElement.tagName}`, token.position?.start || 0);
-      }
+  if (RAW_TEXT_ELEMENTS.has(currentTagName)) {
+    if (token.type === TokenType.TEXT) {
+      parseText(state, token);
+      return;
+    } else if (token.type === TokenType.TAG_CLOSE && token.value.toLowerCase() === currentTagName) {
       state.stack.pop();
-      found = true;
-      break;
+      return;
+    }
+  }
+  
+  if (token.type === TokenType.TAG_OPEN) {
+    const tagName = token.value.toLowerCase();
+    if (tagName === 'title' || tagName === 'style' || tagName === 'script' || tagName === 'noscript') {
+      parseOpenTag(state, token);
+    } else if (tagName === 'meta' || tagName === 'link' || tagName === 'base') {
+      parseOpenTag(state, token);
+    } else if (tagName === 'head') {
+    } else {
+      state.stack.pop();
+      state.insertionMode = InsertionMode.AfterHead;
+      parseToken(state, token);
+    }
+  } else if (token.type === TokenType.TAG_CLOSE) {
+    const tagName = token.value.toLowerCase();
+    if (tagName === 'head') {
+      state.stack.pop();
+      state.insertionMode = InsertionMode.AfterHead;
+    } else if (tagName === 'title' || tagName === 'style' || tagName === 'script' || tagName === 'noscript') {
+      if (currentTagName === tagName) {
+        state.stack.pop();
+      }
     }
+  } else if (token.type === TokenType.COMMENT) {
+    parseComment(state, token);
+  } else if (token.type === TokenType.TEXT && token.value.trim() === '') {
+  } else {
+    state.stack.pop();
+    state.insertionMode = InsertionMode.AfterHead;
+    parseToken(state, token);
   }
+}
 
-  if (!found) {
-    addError(state, `Unexpected closing tag: ${tagName}`, token.position?.start || 0);
+function parseTokenInAfterHeadMode(state: ParserState, token: Token): void {
+  if (token.type === TokenType.TAG_OPEN && token.value.toLowerCase() === 'body') {
+    const body = createElement('body', token.attributes || {});
+    appendChild(getCurrentParent(state), body);
+    state.root.body = body;
+    state.stack.push(body);
+    state.insertionMode = InsertionMode.InBody;
+  } else if (token.type === TokenType.COMMENT) {
+    parseComment(state, token);
+  } else if (token.type === TokenType.TEXT && token.value.trim() === '') {
+    // Ignore whitespace
+  } else {
+    const body = createElement('body', {});
+    appendChild(getCurrentParent(state), body);
+    state.root.body = body;
+    state.stack.push(body);
+    state.insertionMode = InsertionMode.InBody;
+    parseToken(state, token);
+  }
+}
+
+function parseTokenInInBodyMode(state: ParserState, token: Token): void {
+  if (token.type === TokenType.TAG_OPEN) {
+    const tagName = token.value.toLowerCase();
+
+    handleAutoClosing(state, tagName);
+
+    const currentParent = getCurrentParent(state);
+
+    const element = createElement(tagName, token.attributes || {});
+
+    appendChild(currentParent, element);
+
+    if (!token.isSelfClosing && !VOID_ELEMENTS.has(tagName)) {
+      state.stack.push(element);
+    }
+  } else if (token.type === TokenType.TAG_CLOSE) {
+    const tagName = token.value.toLowerCase();
+    
+    // Generate implied end tags
+    const impliedEndTags = ['dd', 'dt', 'li', 'option', 'optgroup', 'p', 'rb', 'rp', 'rt', 'rtc'];
+    while (state.stack.length > 1) { // Don't pop document
+      const currentElement = getCurrentElement(state);
+      if (!currentElement || !impliedEndTags.includes(currentElement.tagName.toLowerCase()) || currentElement.tagName.toLowerCase() === tagName) {
+        break;
+      }
+      state.stack.pop();
+      addError(state, `Implied end tag: ${currentElement.tagName}`, token.position?.offset || 0);
+    }
+    
+    const currentElement = getCurrentElement(state);
+    if (currentElement && currentElement.tagName.toLowerCase() === tagName) {
+      state.stack.pop();
+    } else {
+      // For now, just ignore unmatched closing tags
+      // TODO: Implement full adoption agency algorithm
+      addError(state, `Unmatched closing tag: ${tagName}`, token.position?.offset || 0);
+    }
+  } else if (token.type === TokenType.TEXT) {
+    parseText(state, token);
+  } else if (token.type === TokenType.COMMENT) {
+    parseComment(state, token);
+  } else if (token.type === TokenType.CDATA) {
+    parseCDATA(state, token);
+  } else if (token.type === TokenType.DOCTYPE) {
+    // Ignore
+  } else if (token.type === TokenType.PROCESSING_INSTRUCTION) {
+    parseProcessingInstruction(state, token);
   }
 }
 
@@ -192,76 +468,134 @@ function parseText(state: ParserState, token: Token): void {
     return;
   }
 
-  const textNode: ASTNode = {
-    type: ASTNodeType.TEXT,
-    content,
-    parent: currentParent,
-    position: token.position
-  };
-
-  if (currentParent.children) {
-    currentParent.children.push(textNode);
-  }
+  const textNode = createTextNode(content);
+  appendChild(currentParent, textNode);
 }
 
 function parseComment(state: ParserState, token: Token): void {
   const currentParent = getCurrentParent(state);
 
-  const commentNode: ASTNode = {
-    type: ASTNodeType.COMMENT,
-    content: token.value,
-    parent: currentParent,
-    position: token.position
-  };
-
-  if (currentParent.children) {
-    currentParent.children.push(commentNode);
-  }
+  const commentNode = createComment(token.value);
+  appendChild(currentParent, commentNode);
 }
 
 function parseCDATA(state: ParserState, token: Token): void {
-  const currentParent = getCurrentParent(state);
-
-  const cdataNode: ASTNode = {
-    type: ASTNodeType.CDATA,
-    content: token.value,
-    parent: currentParent,
-    position: token.position
-  };
-
-  if (currentParent.children) {
-    currentParent.children.push(cdataNode);
-  }
+  // TODO: implement CDATA
 }
 
 function parseDoctype(state: ParserState, token: Token): void {
-  const currentParent = getCurrentParent(state);
-
-  const doctypeNode: ASTNode = {
-    type: ASTNodeType.DOCTYPE,
-    content: token.value,
-    parent: currentParent,
-    position: token.position
-  };
-
-  if (currentParent.children) {
-    currentParent.children.push(doctypeNode);
-  }
+  const doctype = createDoctype(token.value || 'html');
+  appendChild(state.root, doctype);
+  state.root.doctype = doctype;
 }
 
 function parseProcessingInstruction(state: ParserState, token: Token): void {
-  const currentParent = getCurrentParent(state);
-
-  const piNode: ASTNode = {
-    type: ASTNodeType.PROCESSING_INSTRUCTION,
-    content: token.value,
-    parent: currentParent,
-    position: token.position
-  };
+  // TODO: implement ProcessingInstruction
+}
 
-  if (currentParent.children) {
-    currentParent.children.push(piNode);
+function runAdoptionAgencyAlgorithm(state: ParserState, tagName: string, token: Token): void {
+  // HTML5 Adoption Agency Algorithm - simplified but more correct implementation
+  
+  // 1. If the current node is an HTML element whose tag name matches the token's tag name, 
+  // then pop the current node off the stack of open elements and abort these steps.
+  const currentElement = getCurrentElement(state);
+  if (currentElement && currentElement.tagName.toLowerCase() === tagName) {
+    state.stack.pop();
+    return;
   }
+  
+  // 2. Let outer loop counter be 0
+  let outerLoopCounter = 0;
+  const formattingElements = ['a', 'b', 'big', 'code', 'em', 'font', 'i', 'nobr', 's', 'small', 'strike', 'strong', 'tt', 'u'];
+  
+  while (outerLoopCounter < 8) { // Prevent infinite loops
+    outerLoopCounter++;
+    
+    // 3. Let the formatting element be the last element in the list of active formatting elements 
+    // that is between the end of the list and the last scope marker or the start of the list, 
+    // if any, that has the same tag name as the token.
+    
+    // For simplicity, find the innermost element with matching tag name
+    let formattingElementIndex = -1;
+    for (let i = state.stack.length - 1; i >= 0; i--) {
+      const element = state.stack[i];
+      if (element.tagName && element.tagName.toLowerCase() === tagName && formattingElements.includes(tagName)) {
+        formattingElementIndex = i;
+        break;
+      }
+    }
+    
+    if (formattingElementIndex === -1) {
+      // No formatting element found, just find any element with matching tag name
+      for (let i = state.stack.length - 1; i >= 0; i--) {
+        const element = state.stack[i];
+        if (element.tagName && element.tagName.toLowerCase() === tagName) {
+          formattingElementIndex = i;
+          break;
+        }
+      }
+    }
+    
+    if (formattingElementIndex === -1) {
+      // No matching element found, ignore the token
+      addError(state, `Stray end tag: ${tagName}`, token.position?.offset || 0);
+      return;
+    }
+    
+    const formattingElement = state.stack[formattingElementIndex];
+    
+    // 4. If there is no element in the stack of open elements that has the same tag name as the 
+    // formatting element, then remove the element from the list of active formatting elements 
+    // and abort these steps.
+    let openElementIndex = -1;
+    for (let i = state.stack.length - 1; i >= 0; i--) {
+      if (state.stack[i] === formattingElement) {
+        openElementIndex = i;
+        break;
+      }
+    }
+    
+    if (openElementIndex === -1) {
+      // Element not in stack, ignore
+      return;
+    }
+    
+    // 5. If the element is not in the stack of open elements, then this is a parse error; 
+    // remove the element from the list of active formatting elements and abort these steps.
+    // (Already checked above)
+    
+    // 6. Let the furthest block be the topmost node in the stack of open elements that is lower 
+    // in the stack than the formatting element, and is an element in the special category.
+    const specialElements = ['address', 'article', 'aside', 'blockquote', 'center', 'details', 'dialog', 'dir', 'div', 'dl', 'fieldset', 'figcaption', 'figure', 'footer', 'form', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'header', 'hgroup', 'hr', 'li', 'listing', 'main', 'menu', 'nav', 'ol', 'p', 'pre', 'section', 'table', 'ul', 'xmp'];
+    
+    let furthestBlockIndex = -1;
+    for (let i = openElementIndex + 1; i < state.stack.length; i++) {
+      const element = state.stack[i];
+      if (element.tagName && specialElements.includes(element.tagName.toLowerCase())) {
+        furthestBlockIndex = i;
+        break;
+      }
+    }
+    
+    if (furthestBlockIndex === -1) {
+      // No special element found, just pop elements until we reach the formatting element
+      while (state.stack.length > openElementIndex + 1) {
+        state.stack.pop();
+      }
+      state.stack.pop(); // Pop the formatting element
+      return;
+    }
+    
+    // 7. Simplified: just pop everything until the formatting element
+    while (state.stack.length > openElementIndex + 1) {
+      state.stack.pop();
+    }
+    state.stack.pop(); // Pop the formatting element
+    return;
+  }
+  
+  // If we get here, something went wrong, ignore the token
+  addError(state, `Adoption agency gave up on: ${tagName}`, token.position?.offset || 0);
 }
 
 function handleAutoClosing(state: ParserState, tagName: string): void {
@@ -269,19 +603,19 @@ function handleAutoClosing(state: ParserState, tagName: string): void {
   if (!autoCloseList) return;
 
   const currentElement = getCurrentElement(state);
-  if (currentElement && currentElement.tagName && autoCloseList.includes(currentElement.tagName)) {
+  if (currentElement && currentElement.tagName && autoCloseList.includes(currentElement.tagName.toLowerCase())) {
     state.stack.pop();
   }
 }
 
-function getCurrentParent(state: ParserState): ASTNode {
-  return state.stack[state.stack.length - 1]!;
+function getCurrentParent(state: ParserState): any {
+  return state.stack[state.stack.length - 1];
 }
 
-function getCurrentElement(state: ParserState): ASTNode | null {
+function getCurrentElement(state: ParserState): any {
   for (let i = state.stack.length - 1; i >= 0; i--) {
-    const element = state.stack[i]!;
-    if (element.type === ASTNodeType.ELEMENT) {
+    const element = state.stack[i];
+    if (element.nodeType === 1) { // ELEMENT_NODE
       return element;
     }
   }
@@ -306,7 +640,7 @@ function addError(state: ParserState, message: string, position: number): void {
   });
 }
 
-function shouldSkipWhitespace(parent: ASTNode): boolean {
+function shouldSkipWhitespace(parent: any): boolean {
   const skipWhitespaceIn = new Set([
     'html', 'head', 'body', 'table', 'tbody', 'thead', 'tfoot', 'tr',
     'ul', 'ol', 'dl', 'select', 'optgroup'
@@ -314,42 +648,3 @@ function shouldSkipWhitespace(parent: ASTNode): boolean {
 
   return parent.tagName ? skipWhitespaceIn.has(parent.tagName) : false;
 }
-
-export function traverseAST(node: ASTNode, callback: (node: ASTNode) => void): void {
-  callback(node);
-  
-  if (node.children) {
-    for (const child of node.children) {
-      traverseAST(child, callback);
-    }
-  }
-}
-
-export function findNodesByTagName(root: ASTNode, tagName: string): ASTNode[] {
-  const results: ASTNode[] = [];
-  
-  traverseAST(root, (node) => {
-    if (node.type === ASTNodeType.ELEMENT && node.tagName === tagName.toLowerCase()) {
-      results.push(node);
-    }
-  });
-
-  return results;
-}
-
-export function findNodesByAttribute(root: ASTNode, attrName: string, attrValue?: string): ASTNode[] {
-  const results: ASTNode[] = [];
-  
-  traverseAST(root, (node) => {
-    if (node.type === ASTNodeType.ELEMENT && node.attributes) {
-      const hasAttr = attrName in node.attributes;
-      const valueMatches = attrValue === undefined || node.attributes[attrName] === attrValue;
-      
-      if (hasAttr && valueMatches) {
-        results.push(node);
-      }
-    }
-  });
-
-  return results;
-}