MediaWiki:Gadget-LabelScan.js: Unterschied zwischen den Versionen

Version vom 5. November 2025, 22:35 Uhr

/* global mw, Tesseract */
(function () {
  'use strict';

  // ========== ADOS: Kategorien & Distillery-Tokens ==========
  const ADOS_CATEGORIES = [
    'Alle A Dream of Scotland Abfüllungen',
    'Alle A Dream of Ireland Abfüllungen',
    'Alle A Dream of... – Der Rest der Welt Abfüllungen',
    'Friendly Mr. Z Whiskytainment Abfüllungen',
    'Die Whisky Elfen Abfüllungen',
    'The Fine Art of Whisky Abfüllungen',
    'Alle Rumbastic Abfüllungen'
  ];

  const KNOWN_TOKENS = [
    'Ardbeg','Ardmore','Arran','Auchroisk','Ben Nevis','Blair Athol','Bowmore',
    'Caol Ila','Clynelish','Glenallachie','Glenrothes','Longmorn','Lagavulin',
    'Tullibardine','Dalmore','Benrinnes','Mortlach','Glenlivet','Inchgower',
    'Islay','Speyside','Highland','Lowland','Campbeltown','Ireland'
  ];

  // ========== Utility: Normalisierung & Fuzzy ==========
  function norm(s){
    return String(s||'')
      .toLowerCase()
      .normalize('NFD').replace(/[\u0300-\u036f]/g,'')
      .replace(/[^a-z0-9\s\-]/g,' ')
      .replace(/\s+/g,' ')
      .trim();
  }

  function levenshtein(a,b){
    a = norm(a); b = norm(b);
    const m = a.length, n = b.length;
    if (!m) return n; if (!n) return m;
    const dp = new Array(n+1);
    for (let j=0;j<=n;j++) dp[j]=j;
    for (let i=1;i<=m;i++){
      let prev = dp[0], tmp; dp[0]=i;
      for (let j=1;j<=n;j++){
        tmp = dp[j];
        dp[j] = (a[i-1]===b[j-1]) ? prev : 1 + Math.min(prev, dp[j-1], dp[j]);
        prev = tmp;
      }
    }
    return dp[n];
  }

  function tokenSet(str){ return new Set(norm(str).split(' ').filter(Boolean)); }
  function overlapScore(a,b){
    const A = tokenSet(a), B = tokenSet(b);
    if (!A.size || !B.size) return 0;
    let inter=0; A.forEach(t=>{ if (B.has(t)) inter++; });
    return inter / Math.max(A.size, B.size);
  }

  function escHTML (s) {
    return String(s ?? '')
      .replace(/&/g,'&amp;')
      .replace(/</g,'&lt;')
      .replace(/>/g,'&gt;')
      .replace(/"/g,'&quot;')
      .replace(/'/g,'&#39;');
  }

  // ========== UI Präsenz & Helfer ==========
  function hasUI () {
    return !!document.getElementById('ados-scan-run') &&
           !!document.getElementById('ados-scan-file');
  }
  function setStatus (t) {
    var el = document.getElementById('ados-scan-status');
    if (el) el.textContent = t || '';
  }
  function setProgress (p) {
    var bar = document.getElementById('ados-scan-progress');
    if (!bar) return;
    if (p == null) { bar.hidden = true; bar.value = 0; }
    else { bar.hidden = false; bar.value = Math.max(0, Math.min(1, p)); }
  }
  function showPreview (file) {
    var url = URL.createObjectURL(file);
    var prev = document.getElementById('ados-scan-preview');
    if (prev) {
      prev.innerHTML = '<img alt="Vorschau" src="' + url + '">';
      prev.setAttribute('aria-hidden', 'false');
    }
  }

  // ========== Tesseract bei Bedarf laden ==========
  var tesseractReady;
  function ensureTesseract () {
    if (tesseractReady) return tesseractReady;
    tesseractReady = new Promise(function (resolve, reject) {
      if (window.Tesseract) return resolve();
      var s = document.createElement('script');
      s.src = 'https://cdn.jsdelivr.net/npm/tesseract.js@5/dist/tesseract.min.js';
      s.async = true;
      s.onload = resolve;
      s.onerror = function () {
        var s2 = document.createElement('script');
        s2.src = 'https://unpkg.com/tesseract.js@5/dist/tesseract.min.js';
        s2.async = true;
        s2.onload = resolve;
        s2.onerror = function () { reject(new Error('Tesseract konnte nicht geladen werden')); };
        document.head.appendChild(s2);
      };
      document.head.appendChild(s);
    });
    return tesseractReady;
  }

  // ========== Bild-Vorverarbeitung (Canvas) ==========
  async function preprocessImage (file) {
    const img = await new Promise((res, rej) => {
      const o = new Image();
      o.onload = () => res(o);
      o.onerror = rej;
      o.src = URL.createObjectURL(file);
    });

    // Längste Kante auf ~1800px skalieren (besser für OCR)
    const MAX = 1800;
    const scale = (img.width > img.height) ? (MAX / img.width) : (MAX / img.height);
    const w = Math.round(img.width * scale);
    const h = Math.round(img.height * scale);

    const c = document.createElement('canvas');
    c.width = w; c.height = h;
    const ctx = c.getContext('2d');
    ctx.imageSmoothingEnabled = true;
    ctx.drawImage(img, 0, 0, w, h);

    // Graustufen + leichter Kontrastboost
    const id = ctx.getImageData(0, 0, w, h);
    const d = id.data;
    for (let i = 0; i < d.length; i += 4) {
      const g = 0.2126 * d[i] + 0.7152 * d[i + 1] + 0.0722 * d[i + 2];
      const v = Math.max(0, Math.min(255, (g - 128) * 1.15 + 128));
      d[i] = d[i + 1] = d[i + 2] = v;
    }
    ctx.putImageData(id, 0, 0);

    return c; // Canvas an Tesseract übergeben
  }

  // ========== OCR ==========
  async function runOCR (file) {
    await ensureTesseract();
    setProgress(0);
    const canvas = await preprocessImage(file);

    const res = await Tesseract.recognize(canvas, 'deu+eng', {
      // psm 6: „ein Block Text“ – robust für Label
      tessedit_pageseg_mode: 6,
      preserve_interword_spaces: 1,
      logger: function (m) {
        if (m && m.status === 'recognizing text' && typeof m.progress === 'number') {
          setProgress(m.progress);
        }
      }
    });

    setProgress(null);
    return (res && res.data && res.data.text) || '';
  }

  // ========== Hinweise aus OCR extrahieren ==========
  function extractHints (text) {
    const raw = String(text || '').replace(/\s+/g, ' ').trim();

    // Distillery-/Marken-Token
    const foundNames = [];
    KNOWN_TOKENS.forEach(t => {
      const re = new RegExp('\\b' + t.replace(/[.*+?^${}()|[\]\\]/g, '\\$&') + '\\b', 'i');
      if (re.test(raw)) foundNames.push(t);
    });

    // Alter
    const ageRe = /\b([1-9]\d?)\s?(?:years?|yo|jahr(?:e)?)\b/gi;
    const ages = []; let m;
    while ((m = ageRe.exec(raw)) !== null){ const n = m[1]; if (!ages.includes(n)) ages.push(n); }

    // Jahrgänge
    const yearRe = /\b(19|20)\d{2}\b/g;
    const years=[]; while ((m = yearRe.exec(raw)) !== null){ if(!years.includes(m[0])) years.push(m[0]); }

    // ABV
    const abvRe = /\b(\d{2}(?:[.,]\d)?)\s?%\b/g;
    const abvs=[]; while ((m = abvRe.exec(raw)) !== null){ abvs.push(m[1].replace(',','.')); }

    // Serien-Keywords (leichtgewichtig)
    const seriesMap = [
      ['A Dream of Scotland','dream of scotland'],
      ['A Dream of Ireland','dream of ireland'],
      ['A Dream of... – Der Rest der Welt','rest der welt','dream of','rest of the world'],
      ['Friendly Mr. Z Whiskytainment Abfüllungen','friendly mr z','mr z'],
      ['Die Whisky Elfen Abfüllungen','whisky elfen','elfen'],
      ['The Fine Art of Whisky Abfüllungen','fine art of whisky','fine art'],
      ['Alle Rumbastic Abfüllungen','rumbastic']
    ];
    const low = norm(raw); const series=[];
    seriesMap.forEach(([label, hint])=>{
      if (low.includes(norm(hint))) series.push(label);
    });

    // „Promi-Wörter“
    const wordRe = /\b[A-ZÄÖÜ][A-Za-zÄÖÜäöüß\-]{3,}\b/g;
    const uniq = new Set(); let w; const words=[];
    while ((w = wordRe.exec(raw)) !== null){
      const s = w[0]; if(!uniq.has(s)){ uniq.add(s); words.push(s); if(words.length>=8) break; }
    }

    return { names: foundNames, ages, years, words, abvs, series, raw };
  }

  // ========== Smarte Wiki-Suche (3 Pässe) ==========
  async function searchWikiSmart (hints, limit) {
    await mw.loader.using('mediawiki.api');
    const api = new mw.Api();
    const ns0 = 0;
    const MAX = limit || 12;

    function incatStr () {
      return ADOS_CATEGORIES.map(c => 'incategory:"' + c + '"').join(' ');
    }

    // PASS 1: intitle-Kombis (präzise)
    const pass1 = [];
    if (hints.names.length) {
      hints.names.forEach(n => {
        if (hints.ages.length) hints.ages.forEach(a => pass1.push(`intitle:"${n}" intitle:${a} ${incatStr()}`));
        if (hints.years.length) hints.years.forEach(y => pass1.push(`intitle:"${n}" "${y}" ${incatStr()}`));
        pass1.push(`intitle:"${n}" ${incatStr()}`);
      });
    }

    // PASS 2: gewichtete Volltextsuche
    const key = []
      .concat(hints.names.slice(0, 2), hints.ages.slice(0, 1), hints.years.slice(0, 1), hints.words.slice(0, 3))
      .map(x => `"${x}"`).join(' ');
    const pass2 = key ? [ `${key} ${incatStr()}` ] : [];

    // PASS 3: Prefix auf Titel
    const pass3 = [];
    if (hints.names.length) pass3.push(hints.names[0]);
    if (!pass3.length && hints.words.length) pass3.push(hints.words[0]);

    const seen = new Set(); const out = [];

    async function runSr (q) {
      const r = await api.get({ action: 'query', list: 'search', srsearch: q, srnamespace: ns0, srlimit: MAX, formatversion: 2 });
      (r.query?.search || []).forEach(it => {
        const k = it.title;
        if (seen.has(k)) return;
        seen.add(k);
        out.push(it);
      });
    }

    for (const q of pass1) { await runSr(q); if (out.length >= MAX) return out.slice(0, MAX); }
    for (const q of pass2) { await runSr(q); if (out.length >= MAX) return out.slice(0, MAX); }

    // Prefix (list=prefixsearch)
    for (const p of pass3) {
      const r = await api.get({ action: 'query', list: 'prefixsearch', pssearch: p, psnamespace: ns0, pslimit: MAX });
      (r.query?.prefixsearch || []).forEach(it=>{
        const title = it.title || it['*'];
        const k = title; if (seen.has(k)) return; seen.add(k);
        out.push({ title, snippet: '' });
      });
      if (out.length >= MAX) break;
    }

    return out.slice(0, MAX);
  }

  // ========== Titel aus Kategorien + Fuzzy-Fallback ==========
  async function fetchTitlesFromCategories(){
    await mw.loader.using('mediawiki.api');
    const api = new mw.Api();
    const titles = new Set();

    for (const cat of ADOS_CATEGORIES){
      let cmcontinue = undefined;
      do {
        const r = await api.get({
          action:'query',
          list:'categorymembers',
          cmtitle:'Kategorie:' + cat,
          cmtype:'page',
          cmlimit:'max',
          cmcontinue
        });
        (r.query?.categorymembers||[]).forEach(it=> titles.add(it.title));
        cmcontinue = r.continue && r.continue.cmcontinue;
      } while (cmcontinue);
    }
    return Array.from(titles);
  }

  function scoreTitleAgainstHints(title, hints){
    const normTitle = norm(title);

    // 1) Token-Overlap
    const base = overlapScore(title, hints.raw);

    // 2) Distillery-Boost
    let nameBoost = 0;
    hints.names.forEach(n=>{
      const t = norm(n);
      const d = levenshtein(normTitle, t);
      if (normTitle.includes(t)) nameBoost = Math.max(nameBoost, 0.35);
      else if (d <= 4) nameBoost = Math.max(nameBoost, 0.2);
    });

    // 3) Alter/Jahr/ABV im Titel
    let numBoost = 0;
    hints.ages.forEach(a=>{
      if (new RegExp('\\b'+a+'\\b').test(normTitle)) numBoost = Math.max(numBoost, 0.15);
    });
    hints.years.forEach(y=>{
      if (normTitle.includes(y)) numBoost = Math.max(numBoost, 0.15);
    });
    hints.abvs.forEach(p=>{
      if (normTitle.includes(p.replace('.',''))) numBoost = Math.max(numBoost, 0.1);
    });

    // 4) Serien-Bonus
    let seriesBoost = 0;
    hints.series.forEach(s=>{
      const key = norm(s.split(' Abfüllungen')[0]); // Kern
      if (normTitle.includes(key)) seriesBoost = Math.max(seriesBoost, 0.15);
    });

    // 5) leichte Strafe bei sehr kleinem Overlap
    const penalty = base < 0.2 ? -0.05 : 0;

    return Math.max(0, base + nameBoost + numBoost + seriesBoost + penalty);
  }

  async function fallbackFuzzyTitles(hints, limit){
    const titles = await fetchTitlesFromCategories();
    const scored = titles.map(t => ({ title: t, _score: scoreTitleAgainstHints(t, hints) }));
    scored.sort((a,b)=> b._score - a._score);
    const top = scored.slice(0, limit||12).filter(x=> x._score >= 0.15);
    return top.map(x=> ({ title: x.title, snippet: '' }));
  }

  // ========== Treffer-Rendering ==========
  function renderResults (items) {
    var box = document.getElementById('ados-scan-results');
    if (!box) return;
    box.innerHTML = '';
    if (!items || !items.length) {
      box.innerHTML = '<div class="ados-hit">Keine klaren Treffer. Bitte anderes Foto oder manuell suchen.</div>';
      return;
    }
    items.slice(0, 12).forEach(function (it) {
      var title = it.title || '';
      var link = (mw && mw.util) ? mw.util.getUrl(title.replace(/ /g, '_')) : ('/wiki/' + encodeURIComponent(title.replace(/ /g,'_')));
      var snip = String(it.snippet || '').replace(/<\/?span[^>]*>/g, '').replace(/&quot;/g, '"');
      var div = document.createElement('div');
      div.className = 'ados-hit';
      div.innerHTML =
        '<b><a href="' + link + '">' + escHTML(title) + '</a></b>' +
        (snip ? '<div class="meta">' + snip + '</div>' : '');
      box.appendChild(div);
    });
  }

  // ========== Binding ==========
  var BOUND = false;
  function bind () {
    if (BOUND || !hasUI()) return;

    var runBtn = document.getElementById('ados-scan-run');
    var fileIn = document.getElementById('ados-scan-file');
    var bigBtn = document.getElementById('ados-scan-bigbtn');
    var form   = document.getElementById('ados-scan-form');

    if (!runBtn || !fileIn) return;
    if (runBtn.dataset.bound === '1') return;
    runBtn.dataset.bound = '1'; BOUND = true;

    if (bigBtn) bigBtn.addEventListener('click', function () { fileIn.click(); });
    fileIn.addEventListener('change', function () {
      if (this.files && this.files[0]) showPreview(this.files[0]);
    });
    if (form) {
      form.addEventListener('submit', function(ev){
        ev.preventDefault();
        runBtn.click();
      });
    }

    runBtn.addEventListener('click', async function (ev) {
      ev.preventDefault();
      if (!(fileIn.files && fileIn.files[0])) { alert('Bitte ein Foto auswählen oder aufnehmen.'); return; }
      var f = fileIn.files[0];
      try {
        runBtn.disabled = true; runBtn.textContent = 'Erkenne …';
        setStatus('Erkenne Label …');
        var text = await runOCR(f);

        setStatus('Suche im Wiki …');
        var hints = extractHints(text);
        var hits = await searchWikiSmart(hints, 12);

        if (!hits || !hits.length) {
          setStatus('Kein direkter Treffer – versuche Fuzzy-Matching …');
          hits = await fallbackFuzzyTitles(hints, 12);
        }

        renderResults(hits);
        setStatus(hits && hits.length ? 'Fertig.' : 'Keine klaren Treffer.');
      } catch (e) {
        console.error('[LabelScan]', e);
        setStatus('Fehler bei Erkennung/Suche. Bitte erneut versuchen.');
      } finally {
        runBtn.disabled = false; runBtn.textContent = 'Erkennen & suchen';
      }
    });

    // Sicherheit gegen Overlays
    var wrap = document.getElementById('ados-labelscan');
    if (wrap) wrap.style.position = 'relative';
    runBtn.style.position = 'relative';
    runBtn.style.zIndex = '9999';
    runBtn.style.pointerEvents = 'auto';
  }

  // Erstbindung + Fallbacks + Observer
  if (document.readyState === 'loading') {
    document.addEventListener('DOMContentLoaded', bind);
  } else {
    bind();
  }
  setTimeout(bind, 250);
  setTimeout(bind, 1000);
  var mo = new MutationObserver(function () { if (!BOUND) bind(); });
  mo.observe(document.documentElement || document.body, { childList: true, subtree: true });
})();