yourpart3/backend/scripts/repair-bisaya-number-answer-pairs.js

#!/usr/bin/env node
/**
 * Repairs stale Bisaya number pairs where the Bisaya side is correct, but the
 * German answer side contains a currency phrase such as "20 peso".
 *
 * Default: dry-run. Use --apply to write changes.
 */

import crypto from 'crypto';
import { Op } from 'sequelize';
import { sequelize } from '../utils/sequelize.js';
import VocabCourseLesson from '../models/community/vocab_course_lesson.js';
import VocabSrsItem from '../models/community/vocab_srs_item.js';

const BISAYA_TO_GERMAN_DIGIT = new Map([
  ['usa', '1'],
  ['duha', '2'],
  ['tulo', '3'],
  ['upat', '4'],
  ['lima', '5'],
  ['unom', '6'],
  ['pito', '7'],
  ['walo', '8'],
  ['siyam', '9'],
  ['napulo', '10'],
  ['onse', '11'],
  ['dose', '12'],
  ['trese', '13'],
  ['katorse', '14'],
  ['kinse', '15'],
  ['disisais', '16'],
  ['disisiete', '17'],
  ['disiotso', '18'],
  ['disinuybe', '19'],
  ['kawhaan', '20'],
]);

const CURRENCY_WORDS = new Set(['peso', 'pesos', 'piso', 'pisos']);
const GERMAN_NUMBER_TO_DIGIT = new Map([
  ['ein', '1'],
  ['eins', '1'],
  ['zwei', '2'],
  ['drei', '3'],
  ['vier', '4'],
  ['fuenf', '5'],
  ['funf', '5'],
  ['fünf', '5'],
  ['sechs', '6'],
  ['sieben', '7'],
  ['acht', '8'],
  ['neun', '9'],
  ['zehn', '10'],
  ['elf', '11'],
  ['zwoelf', '12'],
  ['zwolf', '12'],
  ['zwölf', '12'],
  ['dreizehn', '13'],
  ['vierzehn', '14'],
  ['fuenfzehn', '15'],
  ['funfzehn', '15'],
  ['fünfzehn', '15'],
  ['sechzehn', '16'],
  ['siebzehn', '17'],
  ['achtzehn', '18'],
  ['neunzehn', '19'],
  ['zwanzig', '20'],
]);

function parseArgs(argv) {
  return {
    apply: argv.includes('--apply'),
  };
}

function normalizeText(value) {
  return String(value || '')
    .trim()
    .toLowerCase()
    .normalize('NFKC')
    .replace(/[\p{P}\p{S}]+/gu, ' ')
    .replace(/\s+/g, ' ')
    .trim();
}

function compact(value) {
  return normalizeText(value).replace(/\s+/g, '');
}

function getBisayaDigit(value) {
  return BISAYA_TO_GERMAN_DIGIT.get(compact(value)) || '';
}

function getCurrencyNumber(value) {
  const text = normalizeText(value);
  if (!text) return '';
  const parts = text.split(/\s+/).filter(Boolean);
  if (parts.length < 2) return '';
  const last = parts[parts.length - 1];
  if (!CURRENCY_WORDS.has(last)) return '';
  const numberPart = parts.slice(0, -1).join('');
  if (/^\d+$/.test(numberPart)) return String(Number(numberPart));
  return GERMAN_NUMBER_TO_DIGIT.get(numberPart) || '';
}

function fixPair(left, right) {
  const leftBisayaDigit = getBisayaDigit(left);
  const rightCurrencyDigit = getCurrencyNumber(right);
  if (leftBisayaDigit && rightCurrencyDigit && leftBisayaDigit === rightCurrencyDigit) {
    return { left, right: leftBisayaDigit, changed: right !== leftBisayaDigit };
  }

  const rightBisayaDigit = getBisayaDigit(right);
  const leftCurrencyDigit = getCurrencyNumber(left);
  if (rightBisayaDigit && leftCurrencyDigit && rightBisayaDigit === leftCurrencyDigit) {
    return { left: rightBisayaDigit, right, changed: left !== rightBisayaDigit };
  }

  return { left, right, changed: false };
}

function patchPattern(entry) {
  if (!entry || typeof entry !== 'object' || Array.isArray(entry)) {
    return { value: entry, changed: false };
  }

  const next = { ...entry };
  let changed = false;
  const pairKeys = [
    ['target', 'gloss'],
    ['learning', 'reference'],
    ['bisaya', 'native'],
  ];

  pairKeys.forEach(([leftKey, rightKey]) => {
    if (!(leftKey in next) || !(rightKey in next)) return;
    const pair = fixPair(next[leftKey], next[rightKey]);
    if (!pair.changed) return;
    next[leftKey] = pair.left;
    next[rightKey] = pair.right;
    changed = true;
  });

  return { value: next, changed };
}

function patchJson(value) {
  if (Array.isArray(value)) {
    let changed = false;
    const next = value.map((entry) => {
      const patched = patchJson(entry);
      changed = changed || patched.changed;
      return patched.value;
    });
    return { value: next, changed };
  }

  if (value && typeof value === 'object') {
    const direct = patchPattern(value);
    let next = direct.value;
    let changed = direct.changed;

    Object.entries(next).forEach(([key, child]) => {
      if (!child || typeof child !== 'object') return;
      const patched = patchJson(child);
      if (!patched.changed) return;
      next = { ...next, [key]: patched.value };
      changed = true;
    });

    return { value: next, changed };
  }

  return { value, changed: false };
}

function normalizeSrsText(value) {
  return normalizeText(value);
}

function buildSrsItemKey({ courseId, lessonId = null, learning, reference, direction = 'BOTH' }) {
  const raw = [
    Number(courseId) || 0,
    lessonId == null ? 'course' : Number(lessonId) || 0,
    String(direction || 'BOTH').toUpperCase(),
    normalizeSrsText(learning),
    normalizeSrsText(reference),
  ].join('|');
  return crypto.createHash('sha1').update(raw).digest('hex');
}

async function repairLessons({ apply }) {
  const lessons = await VocabCourseLesson.findAll({
    where: {
      corePatterns: {
        [Op.ne]: null,
      },
    },
    attributes: ['id', 'courseId', 'lessonNumber', 'title', 'corePatterns'],
    order: [['courseId', 'ASC'], ['lessonNumber', 'ASC']],
  });

  let changedLessons = 0;
  for (const lesson of lessons) {
    const patched = patchJson(lesson.corePatterns);
    if (!patched.changed) continue;

    changedLessons += 1;
    console.log(
      `Lesson ${lesson.id} (course ${lesson.courseId}, #${lesson.lessonNumber}, ${lesson.title}): corePatterns korrigiert`
    );

    if (apply) {
      await lesson.update({ corePatterns: patched.value });
    }
  }

  return changedLessons;
}

async function repairSrsItems({ apply }) {
  const items = await VocabSrsItem.findAll({
    where: {
      [Op.or]: [
        { learning: { [Op.iLike]: '%peso%' } },
        { reference: { [Op.iLike]: '%peso%' } },
        { learning: { [Op.iLike]: '%piso%' } },
        { reference: { [Op.iLike]: '%piso%' } },
      ],
    },
    order: [['id', 'ASC']],
  });

  let changedItems = 0;
  let keyConflicts = 0;
  for (const item of items) {
    const pair = fixPair(item.learning, item.reference);
    if (!pair.changed) continue;

    changedItems += 1;
    const nextKey = buildSrsItemKey({
      courseId: item.courseId,
      lessonId: item.lessonId,
      learning: pair.left,
      reference: pair.right,
      direction: item.direction,
    });

    console.log(
      `SRS ${item.id}: "${item.learning}" | "${item.reference}" -> "${pair.left}" | "${pair.right}"`
    );

    if (!apply) continue;

    try {
      await item.update({
        learning: pair.left,
        reference: pair.right,
        itemKey: nextKey,
      });
    } catch (error) {
      keyConflicts += 1;
      console.warn(`  item_key-Konflikt bei SRS ${item.id}; aktualisiere Textwerte mit altem Key.`);
      await item.update({
        learning: pair.left,
        reference: pair.right,
      });
    }
  }

  return { changedItems, keyConflicts };
}

async function main() {
  const args = parseArgs(process.argv.slice(2));

  console.log(`Modus: ${args.apply ? 'APPLY (schreibt Änderungen)' : 'DRY-RUN (keine Änderungen)'}`);
  await sequelize.authenticate();

  const changedLessons = await repairLessons(args);
  const { changedItems, keyConflicts } = await repairSrsItems(args);

  console.log('');
  console.log(`Betroffene Lektionen: ${changedLessons}`);
  console.log(`Betroffene SRS-Items: ${changedItems}`);
  if (keyConflicts > 0) {
    console.log(`SRS-Items mit behaltenem altem item_key wegen Unique-Konflikt: ${keyConflicts}`);
  }
  if (!args.apply) {
    console.log('Zum Anwenden erneut mit --apply ausführen.');
  }
}

main()
  .then(async () => {
    await sequelize.close();
  })
  .catch(async (error) => {
    console.error('Fehler beim Reparieren der Bisaya-Zahlenpaare:', error);
    await sequelize.close();
    process.exit(1);
  });