// DWDS Collocate Extractor
// Usage: Copy this code, create a bookmarklet, or run in browser console

(function() {
  // Get the search term from the URL
  const urlParams = new URLSearchParams(window.location.search);
  const searchTerm = urlParams.get('q');
  
  if (!searchTerm) {
    alert('Could not find search term in URL');
    return;
  }
  
  // Find all collocate rows in the table
  const rows = document.querySelectorAll('table tbody tr');
  const results = [];
  
  // Build a map of categories by finding all section headers
  const categoryMap = new Map();
  const tables = document.querySelectorAll('table');
  
  tables.forEach(table => {
    // Find the category in the table header
    const categoryHeader = table.querySelector('thead tr th.no-wrap span.rel-desc');
    let currentCategory = '';
    
    if (categoryHeader) {
      currentCategory = categoryHeader.textContent.trim();
    }
    
    // Process all rows in this table with the current category
    const tableRows = table.querySelectorAll('tbody tr');
    tableRows.forEach(row => {
      const cells = row.querySelectorAll('td');
      if (cells.length < 2) return;
      
      // Get the collocate from the first cell
      const collocateCell = cells[0];
      const collocateSpan = collocateCell.querySelector('.wp-collocation span[data-toggle="tooltip"]');
      
      if (!collocateSpan) return;
      
      const collocate = collocateSpan.textContent.trim();
      if (!collocate) return;
      
      // Analyze word order and format accordingly
      let formatted = formatCollocation(collocate, searchTerm);
      
      // Add category if found
      if (currentCategory) {
        formatted += ` /cat: ${currentCategory}`;
      }
      
      results.push(formatted);
    });
  });
  
  function formatCollocation(collocate, baseTerm) {
    // Remove any extra whitespace
    collocate = collocate.replace(/\s+/g, ' ').trim();
    
    // Check if collocate already contains the base term
    if (collocate.toLowerCase().includes(baseTerm.toLowerCase())) {
      return collocate;
    }
    
    // Determine position based on grammatical patterns
    
    // Adjectives typically come before nouns
    if (isAdjective(collocate) && isNoun(baseTerm)) {
      return `${collocate} ${baseTerm}`;
    }
    
    // Past participles often come before nouns
    if (isPastParticiple(collocate) && isNoun(baseTerm)) {
      return `${collocate} ${baseTerm}`;
    }
    
    // Verbs with separable prefixes
    if (collocate.includes('_')) {
      const parts = collocate.split('_');
      return `${baseTerm} ${parts.join(' ')}`;
    }
    
    // Prepositions come before the term
    if (isPreposition(collocate)) {
      return `${collocate} ${baseTerm}`;
    }
    
    // Coordinating conjunctions (und, oder, etc.)
    if (isConjunction(collocate)) {
      return `${baseTerm} ${collocate}`;
    }
    
    // Default: adjective/modifier before noun, verb patterns after
    if (isNoun(baseTerm)) {
      return `${collocate} ${baseTerm}`;
    } else {
      return `${baseTerm} ${collocate}`;
    }
  }
  
  function isPastParticiple(word) {
    return word.startsWith('ge') || 
           word.match(/t$|en$/) ||
           word.includes('umgestürzt') ||
           word.includes('gefällt');
  }
  
  function isAdjective(word) {
    // Common adjective patterns
    return word.match(/lich$|ig$|bar$|sam$|haft$|los$|voll$|isch$/) ||
           word === 'schnell' ||
           word === 'groß' ||
           word === 'klein';
  }
  
  function isNoun(word) {
    // Nouns start with capital letter in German
    return word.charAt(0) === word.charAt(0).toUpperCase();
  }
  
  function isPreposition(word) {
    const preps = ['an', 'auf', 'aus', 'bei', 'durch', 'für', 'gegen', 
                   'hinter', 'in', 'mit', 'nach', 'neben', 'ohne', 
                   'über', 'um', 'unter', 'von', 'vor', 'zu', 'zwischen'];
    return preps.includes(word.toLowerCase());
  }
  
  function isConjunction(word) {
    const conj = ['und', 'oder', 'aber', 'sondern', 'denn'];
    return conj.includes(word.toLowerCase());
  }
  
  // Display results
  if (results.length === 0) {
    alert('No collocates found. Make sure you are on a DWDS collocation results page.');
    return;
  }
  
  const output = results.join('\n');
  
  // Create a modal to display results
  const modal = document.createElement('div');
  modal.style.cssText = `
    position: fixed;
    top: 50%;
    left: 50%;
    transform: translate(-50%, -50%);
    background: white;
    padding: 20px;
    border: 2px solid #333;
    border-radius: 8px;
    z-index: 10000;
    max-width: 600px;
    max-height: 80vh;
    overflow: auto;
    box-shadow: 0 4px 6px rgba(0,0,0,0.3);
  `;
  
  const textarea = document.createElement('textarea');
  textarea.value = output;
  textarea.style.cssText = `
    width: 100%;
    height: 400px;
    font-family: monospace;
    padding: 10px;
    border: 1px solid #ccc;
    border-radius: 4px;
  `;
  
  const closeBtn = document.createElement('button');
  closeBtn.textContent = 'Close';
  closeBtn.style.cssText = `
    margin-top: 10px;
    padding: 8px 16px;
    background: #007bff;
    color: white;
    border: none;
    border-radius: 4px;
    cursor: pointer;
  `;
  closeBtn.onclick = () => modal.remove();
  
  const copyBtn = document.createElement('button');
  copyBtn.textContent = 'Copy to Clipboard';
  copyBtn.style.cssText = `
    margin-top: 10px;
    margin-left: 10px;
    padding: 8px 16px;
    background: #28a745;
    color: white;
    border: none;
    border-radius: 4px;
    cursor: pointer;
  `;
  copyBtn.onclick = () => {
    textarea.select();
    document.execCommand('copy');
    copyBtn.textContent = 'Copied!';
    setTimeout(() => copyBtn.textContent = 'Copy to Clipboard', 2000);
  };
  
  modal.appendChild(textarea);
  modal.appendChild(copyBtn);
  modal.appendChild(closeBtn);
  document.body.appendChild(modal);
  
  // Auto-select text for easy copying
  textarea.select();
})();