CentralBackend/scripts/compareAboContractDocxVersions.js

const fs = require('fs');
const path = require('path');
const yauzl = require('yauzl');

const ABO_TEMPLATE_DIR = path.join(__dirname, '..', 'templates', 'abo');

const COMPARISONS = [
  {
    label: 'DE',
    previous: path.join(ABO_TEMPLATE_DIR, 'abo-contract-DE.docx'),
    next: path.join(ABO_TEMPLATE_DIR, 'new', 'abo-contract-DE-NEW.docx'),
  },
  {
    label: 'SL',
    previous: path.join(ABO_TEMPLATE_DIR, 'abo-contract-SL.docx'),
    next: path.join(ABO_TEMPLATE_DIR, 'new', 'abo-contract-SL-NEW.docx'),
  },
];

function decodeXml(value) {
  return String(value || '')
    .replace(/&/g, '&')
    .replace(/&lt;/g, '<')
    .replace(/&gt;/g, '>')
    .replace(/&quot;/g, '"')
    .replace(/&apos;/g, "'");
}

function openZip(zipPath) {
  return new Promise((resolve, reject) => {
    yauzl.open(zipPath, { lazyEntries: true }, (error, zipFile) => {
      if (error) {
        reject(error);
        return;
      }

      resolve(zipFile);
    });
  });
}

function readEntry(zipFile, entry) {
  return new Promise((resolve, reject) => {
    zipFile.openReadStream(entry, (error, stream) => {
      if (error) {
        reject(error);
        return;
      }

      const chunks = [];
      stream.on('data', (chunk) => chunks.push(chunk));
      stream.on('end', () => resolve(Buffer.concat(chunks).toString('utf8')));
      stream.on('error', reject);
    });
  });
}

async function readWordXmlEntries(zipPath) {
  const zipFile = await openZip(zipPath);
  const documents = [];

  return new Promise((resolve, reject) => {
    zipFile.readEntry();

    zipFile.on('entry', (entry) => {
      if (!/^word\/(document|header\d+|footer\d+|footnotes|endnotes)\.xml$/i.test(entry.fileName)) {
        zipFile.readEntry();
        return;
      }

      readEntry(zipFile, entry)
        .then((xml) => {
          documents.push({ fileName: entry.fileName, xml });
          zipFile.readEntry();
        })
        .catch(reject);
    });

    zipFile.on('end', () => {
      zipFile.close();
      resolve(documents);
    });

    zipFile.on('error', reject);
  });
}

function xmlToLines(xml) {
  const withStructuralBreaks = String(xml || '')
    .replace(/<w:tab\b[^>]*\/>/gi, '\t')
    .replace(/<w:(?:br|cr)\b[^>]*\/>/gi, '\n')
    .replace(/<\/w:p>/gi, '\n')
    .replace(/<\/w:tr>/gi, '\n')
    .replace(/<\/w:tc>/gi, '\t')
    .replace(/<w:t\b[^>]*>([\s\S]*?)<\/w:t>/gi, (_, value) => decodeXml(value))
    .replace(/<w:delText\b[^>]*>([\s\S]*?)<\/w:delText>/gi, (_, value) => decodeXml(value));

  const textOnly = decodeXml(withStructuralBreaks.replace(/<[^>]+>/g, ' '));

  return textOnly
    .split(/\r?\n/)
    .map((line) => line.replace(/[ \t]+/g, ' ').trim())
    .filter((line) => line.length > 0);
}

async function extractDocxLines(docxPath) {
  const entries = await readWordXmlEntries(docxPath);
  const ordered = entries.sort((left, right) => left.fileName.localeCompare(right.fileName));

  return ordered.flatMap((entry) => xmlToLines(entry.xml));
}

function buildLcsMatrix(left, right) {
  const matrix = Array.from({ length: left.length + 1 }, () => Array(right.length + 1).fill(0));

  for (let leftIndex = left.length - 1; leftIndex >= 0; leftIndex -= 1) {
    for (let rightIndex = right.length - 1; rightIndex >= 0; rightIndex -= 1) {
      if (left[leftIndex] === right[rightIndex]) {
        matrix[leftIndex][rightIndex] = matrix[leftIndex + 1][rightIndex + 1] + 1;
      } else {
        matrix[leftIndex][rightIndex] = Math.max(
          matrix[leftIndex + 1][rightIndex],
          matrix[leftIndex][rightIndex + 1],
        );
      }
    }
  }

  return matrix;
}

function diffLines(previousLines, nextLines) {
  const matrix = buildLcsMatrix(previousLines, nextLines);
  const changes = [];

  let previousIndex = 0;
  let nextIndex = 0;

  while (previousIndex < previousLines.length && nextIndex < nextLines.length) {
    if (previousLines[previousIndex] === nextLines[nextIndex]) {
      previousIndex += 1;
      nextIndex += 1;
      continue;
    }

    if (matrix[previousIndex + 1][nextIndex] >= matrix[previousIndex][nextIndex + 1]) {
      changes.push({ type: 'removed', line: previousIndex + 1, text: previousLines[previousIndex] });
      previousIndex += 1;
    } else {
      changes.push({ type: 'added', line: nextIndex + 1, text: nextLines[nextIndex] });
      nextIndex += 1;
    }
  }

  while (previousIndex < previousLines.length) {
    changes.push({ type: 'removed', line: previousIndex + 1, text: previousLines[previousIndex] });
    previousIndex += 1;
  }

  while (nextIndex < nextLines.length) {
    changes.push({ type: 'added', line: nextIndex + 1, text: nextLines[nextIndex] });
    nextIndex += 1;
  }

  return changes;
}

async function comparePair({ label, previous, next }) {
  const previousExists = fs.existsSync(previous);
  const nextExists = fs.existsSync(next);

  if (!previousExists || !nextExists) {
    throw new Error(`${label}: missing input file(s)`);
  }

  const [previousLines, nextLines] = await Promise.all([
    extractDocxLines(previous),
    extractDocxLines(next),
  ]);

  const changes = diffLines(previousLines, nextLines);

  return {
    label,
    previous,
    next,
    previousLineCount: previousLines.length,
    nextLineCount: nextLines.length,
    changes,
  };
}

function printComparison(result) {
  console.log(`\n=== ${result.label} ===`);
  console.log(`Old: ${path.basename(result.previous)} (${result.previousLineCount} lines)`);
  console.log(`New: ${path.basename(result.next)} (${result.nextLineCount} lines)`);

  if (!result.changes.length) {
    console.log('No textual differences detected.');
    return;
  }

  console.log(`Detected ${result.changes.length} text-level changes:`);
  result.changes.forEach((change) => {
    const marker = change.type === 'added' ? '+' : '-';
    console.log(`${marker} [${change.line}] ${change.text}`);
  });
}

async function main() {
  const results = [];

  for (const comparison of COMPARISONS) {
    results.push(await comparePair(comparison));
  }

  results.forEach(printComparison);
}

main().catch((error) => {
  console.error('[compareAboContractDocxVersions] failed:', error?.message || error);
  process.exitCode = 1;
});