hh-shipments 0.3.20 → 0.3.22

package/package.json

@@ -18,7 +18,7 @@
     "test": "node test/run-tests.js && node test/run-carriers-tests.js"
   },
   "type": "commonjs",
-  "version": "0.3.20",
+  "version": "0.3.22",
   "bin": {
     "hh-shipments": "server.js"
   },

package/public/app.js

@@ -120,12 +120,19 @@ async function runQuick(text, targetEl, preset) {
       new Promise((_, rej) => setTimeout(() => rej(new Error('extract timed out — check your connection and try again')), 45000)),
     ]);
     let html = '';
-    r.added.forEach((a) => { html += `<div class="review">✅ Added <b>${esc(a.number)}</b> (${esc(a.carrier.toUpperCase())})${r.simulated ? ' <span class="sim-badge">SIMULATED</span>' : ''}</div>`; });
+    r.added.forEach((a) => { html += `<div class="review">✅ Added <b>${esc(a.number)}</b> (${esc(a.carrier.toUpperCase())})${a.ocr_fixed ? ' <span class="muted" title="The scan misread a character; the number was repaired against the strict carrier format">· OCR-corrected</span>' : ''}${r.simulated ? ' <span class="sim-badge">SIMULATED</span>' : ''}</div>`; });
     r.duplicates.forEach((n) => { html += `<div class="review">⛔ <b>${esc(n)}</b> is already tracked — skipped.</div>`; });
     (r.updated || []).forEach((u) => { html += `<div class="review">🔄 Updated <b>${esc(u.number)}</b> — seller: <b>${esc(u.seller_name || '—')}</b>${u.device ? ` · ${esc(u.device)}` : ''}</div>`; });
     r.needs_review.forEach((c, i) => {
-      html += `<div class="review" id="rev-${i}"><span class="warn">Needs review</span> — <b>${esc(c.number)}</b> (${esc(c.carrier.toUpperCase())}, ${esc(c.confidence)} confidence)${c.error ? ' — ' + esc(c.error) : ''}
-        <div class="row"><button class="secondary" data-add="${i}">Add anyway</button></div></div>`;
+      // An 'unknown'-carrier candidate can never pass server validation, so
+      // "Add anyway" on it is a dead button that only ever answers
+      // "invalid unknown tracking number format". Show the raw read for a
+      // manual fix instead.
+      const addable = c.carrier && c.carrier !== 'unknown';
+      html += `<div class="review" id="rev-${i}"><span class="warn">Needs review</span> — <b>${esc(c.number)}</b> (${esc(String(c.carrier).toUpperCase())}, ${esc(c.confidence)} confidence)${c.error ? ' — ' + esc(c.error) : ''}`;
+      html += addable
+        ? `<div class="row"><button class="secondary" data-add="${i}">Add anyway</button></div></div>`
+        : `<div class="muted">Couldn't match a carrier — the raw read is below; type the number into the box above.</div></div>`;
     });
     if (r.context && (r.context.seller_name || r.context.device)) {
       html = `<div class="review">📝 Detected: ${esc(r.context.seller_name || '—')} · ${esc(r.context.device || '—')}</div>` + html;
@@ -168,9 +175,50 @@ function loadTesseract() {
   return tesseractLoading;
 }
 
+// Tesseract reads dark-text-on-light far better than the reverse. Chat
+// screenshots (dark mode) come out as white-on-black, so detect dark images
+// and invert them before OCR. Light images (labels) pass through untouched.
+async function maybeInvertForOcr(file) {
+  try {
+    const bmp = await createImageBitmap(file);
+    const w = bmp.width, h = bmp.height;
+    const canvas = document.createElement('canvas');
+    // sample at reduced size for speed
+    const sw = Math.min(200, w), sh = Math.min(200, h);
+    canvas.width = sw; canvas.height = sh;
+    const ctx = canvas.getContext('2d', { willReadFrequently: true });
+    ctx.drawImage(bmp, 0, 0, sw, sh);
+    const px = ctx.getImageData(0, 0, sw, sh).data;
+    let sum = 0, n = 0;
+    for (let i = 0; i < px.length; i += 16) {
+      sum += 0.299 * px[i] + 0.587 * px[i + 1] + 0.114 * px[i + 2];
+      n++;
+    }
+    const avg = sum / n;
+    bmp.close();
+    if (avg > 110) return file; // already light — leave it alone
+    // dark image: invert to dark-on-light
+    const c2 = document.createElement('canvas');
+    c2.width = w; c2.height = h;
+    const x2 = c2.getContext('2d');
+    const full = await createImageBitmap(file);
+    x2.drawImage(full, 0, 0);
+    full.close();
+    const img = x2.getImageData(0, 0, w, h);
+    const d = img.data;
+    for (let i = 0; i < d.length; i += 4) {
+      d[i] = 255 - d[i]; d[i + 1] = 255 - d[i + 1]; d[i + 2] = 255 - d[i + 2];
+    }
+    x2.putImageData(img, 0, 0);
+    const blob = await new Promise((res) => c2.toBlob(res, 'image/jpeg', 0.92));
+    return blob ? new File([blob], 'inv.jpg', { type: 'image/jpeg' }) : file;
+  } catch (e) { return file; }
+}
+
 async function ocrFile(file, statusEl) {
   await loadTesseract();
   statusEl.textContent = 'Reading screenshot…';
+  file = await maybeInvertForOcr(file);
   const { data } = await Tesseract.recognize(file, 'eng', { logger: (m) => { if (m.status === 'recognizing text') statusEl.textContent = `Reading… ${Math.round(m.progress * 100)}%`; } });
   statusEl.textContent = '';
   return data.text || '';
@@ -211,7 +259,7 @@ async function handlePhotoFile(f) {
     const r = await runQuick(text, $('photo-result'), preset);
     if (looksLikeAppScreenshot(topText || text)) $('photo-result').insertAdjacentHTML('afterbegin', '<p class="warn">⚠️ This looks like a screenshot of the app itself — choose the label photo directly for best results.</p>');
     if (barcode && r && r.added.length) $('photo-result').insertAdjacentHTML('afterbegin', '<p class="muted">📷 Read from the barcode — checksum verified.</p>');
-    if (r && !r.added.length && !r.needs_review.length && !r.duplicates.length) showOcrRead($('photo-result'), text);
+    if (r && !r.added.length && !r.duplicates.length && !r.needs_review.some((c) => c.carrier && c.carrier !== 'unknown')) showOcrRead($('photo-result'), text);
     loadPending();
   } catch (err) { $('ocr-status').textContent = ''; $('photo-result').innerHTML = `<p class="error">${esc(err.message)}</p>`; }
 }
@@ -388,7 +436,7 @@ async function handlePastedImageFile(f) {
       const r = await runQuick(text, target, topText ? { labelText: topText } : undefined);
       if (looksLikeAppScreenshot(topText)) target.insertAdjacentHTML('afterbegin', '<p class="warn">⚠️ This looks like a screenshot of the app itself — paste the label photo directly for best results.</p>');
       if (r && r.added.length) target.insertAdjacentHTML('afterbegin', '<p class="muted">📷 Read from the barcode — checksum verified.</p>');
-      if (r && !r.added.length && !r.needs_review.length && !r.duplicates.length) showOcrRead(target, text);
+      if (r && !r.added.length && !r.duplicates.length && !r.needs_review.some((c) => c.carrier && c.carrier !== 'unknown')) showOcrRead(target, text);
       return;
     }
     // 2) Fallback: OCR the printed text
@@ -404,7 +452,7 @@ async function handlePastedImageFile(f) {
     if (live) live.textContent = '🔍 Finding tracking number…';
     const r = await runQuick(text, target);
     if (looksLikeAppScreenshot(text)) target.insertAdjacentHTML('afterbegin', '<p class="warn">⚠️ This looks like a screenshot of the app itself — paste the label photo directly for best results.</p>');
-    if (r && !r.added.length && !r.needs_review.length && !r.duplicates.length) showOcrRead(target, text);
+    if (r && !r.added.length && !r.duplicates.length && !r.needs_review.some((c) => c.carrier && c.carrier !== 'unknown')) showOcrRead(target, text);
   } catch (err) { $('ocr-status-paste').textContent = ''; target.innerHTML = `<p class="error">${esc(err.message)}</p>`; }
 }
 
@@ -459,7 +507,10 @@ async function reviewPendingImage(id) {
   $('photo-result').innerHTML = `<img src="/api/intake/image/${id}" style="max-width:100%;border-radius:10px"><p class="muted">Extracting…</p>`;
   try {
     await loadTesseract();
-    const { data } = await Tesseract.recognize(`/api/intake/image/${id}`, 'eng');
+    const resp = await fetch(`/api/intake/image/${id}`);
+    const blob = await resp.blob();
+    const file = await maybeInvertForOcr(new File([blob], 'pending.jpg', { type: blob.type || 'image/jpeg' }));
+    const { data } = await Tesseract.recognize(file, 'eng');
     await api.patch('/api/intake/' + id, { state: 'reviewed' });
     await runQuick(data.text || '', $('photo-result'));
     loadPending();

package/public/index.html

@@ -4,7 +4,7 @@
 <meta charset="utf-8">
 <meta name="viewport" content="width=device-width, initial-scale=1, viewport-fit=cover">
 <meta name="theme-color" content="#0a84ff">
-<meta name="app-version" content="0.3.20">
+<meta name="app-version" content="0.3.22">
 <title>HH Shipments</title>
 <link rel="manifest" href="/manifest.webmanifest">
 <link rel="apple-touch-icon" href="/icon-192.png">

package/server.js

@@ -248,7 +248,7 @@ function quickAddText(text, opts = {}) {
       const r = addShipment({ number: c.number, carrier: c.carrier, seller_name: seller_name || autoCtx.seller_name, device: device || autoCtx.device, notes, source });
       if (r.error) needs_review.push({ ...c, error: r.error });
       else if (r.duplicate) duplicates.push(c.number);
-      else { added.push({ id: r.id, number: r.number, carrier: r.carrier }); existing.add(c.number); }
+      else { added.push({ id: r.id, number: r.number, carrier: r.carrier, ocr_fixed: !!c.ocr_fixed, merged: !!c.merged }); existing.add(c.number); }
     } else {
       needs_review.push(c);
     }

package/src/extract.js

@@ -165,7 +165,7 @@
       else num = repairUpsZ(ups) || repairUpsTail(ups.slice(2));
       if (num && !seen[num]) {
         seen[num] = true;
-        found.push({ number: num, carrier: 'ups', confidence: conf, ocr_fixed: num !== ups });
+        found.push({ number: num, carrier: 'ups', confidence: (num !== ups && conf === 'medium') ? 'high' : conf, ocr_fixed: num !== ups });
       }
     }
   }
@@ -198,7 +198,8 @@
       var num = repairUpsTail(tail);
       if (num && !seen[num]) {
         seen[num] = true;
-        found.push({ number: num, carrier: 'ups', confidence: conf, ocr_fixed: true });
+        // repairUpsTail only returns strict-validated numbers: trust them.
+        found.push({ number: num, carrier: 'ups', confidence: 'high', ocr_fixed: true });
       }
     }
   }
@@ -312,9 +313,10 @@
       found.push(c);
     }
     // a) UPS number split by spaces/dashes (whole-text fallback; the anchor
-    // pass already tried the hot zones at high confidence)
-    spacedUpsIn(text, found, seen, 'medium');
-    looseUpsIn(text, found, seen, 'medium');
+    // pass already tried the hot zones at high confidence). Repairs are
+    // strict-structure-validated, so they stay high confidence.
+    spacedUpsIn(text, found, seen, 'high');
+    looseUpsIn(text, found, seen, 'high');
     // b) long digit-like runs with OCR confusions -> fix -> re-match digit formats
     var runRe = /[0-9OILSBoilszb]{10,24}/g, run;
     while ((run = runRe.exec(text)) !== null) {
@@ -337,7 +339,20 @@
     tokens.slice(0, 3).forEach(function (t) {
       // UPS "1Z" misread as "12"/"IZ" in a joined (space-free) token
       var fixed = repairUpsZ(t);
-      if (fixed) { push(fixed, 'ups', 'medium'); return; }
+      if (fixed) { push(fixed, 'ups', 'high'); return; }
+      // Token merged with adjacent digits (e.g. the message timestamp right
+      // after the number in a chat bubble): try the 18-char prefix. Only
+      // accepted when the repaired result strictly validates — never invented.
+      if (/^(1Z|12|IZ)/i.test(t) && t.length > 18) {
+        var trunc = repairUpsZ(t.slice(0, 18));
+        if (!trunc && /^1Z[0-9A-Z]{16}$/.test(t.slice(0, 18))) {
+          // correct anchor, merged with trailing digits (chat timestamp):
+          // accept the 18-char prefix only if it strictly validates
+          var pre = t.slice(0, 18);
+          if (UPS_STRICT.test(pre) && /[A-Z]/.test(pre.slice(2, 8))) trunc = pre;
+        }
+        if (trunc) { push(trunc, 'ups', 'high', { merged: true }); return; }
+      }
       // OCR-fixed variant may match a carrier format (e.g. I->1 at the start)
       var digitFixed = ocrDigitFix(t);
       var gc = guessCarrier(digitFixed);