hh-shipments 0.3.9 → 0.3.11

package/package.json

@@ -18,7 +18,7 @@
     "test": "node test/run-tests.js && node test/run-carriers-tests.js"
   },
   "type": "commonjs",
-  "version": "0.3.9",
+  "version": "0.3.11",
   "bin": {
     "hh-shipments": "server.js"
   },

package/public/app.js

@@ -176,8 +176,17 @@ async function handlePhotoFile(f) {
   try {
     f = await shrinkImage(f);
     const url = URL.createObjectURL(f);
-    target.innerHTML = `<div class="review"><img src="${url}" style="max-width:100%;border-radius:10px"><p class="muted">📖 Reading screenshot…</p></div>`;
-    const text = await ocrFile(f, $('ocr-status'));
+    target.innerHTML = `<div class="review"><img src="${url}" style="max-width:100%;border-radius:10px"><p class="muted" id="ocr-live-photo">📷 Scanning barcode…</p></div>`;
+    const live = $('ocr-live-photo');
+    const barcode = await decodeBarcode(f);
+    let text;
+    if (barcode) {
+      text = barcode;
+      if (live) live.textContent = '📷 Barcode read — verifying…';
+    } else {
+      if (live) live.textContent = '📖 Reading screenshot…';
+      text = await ocrFile(f, $('ocr-status'));
+    }
     // Upload original for reference, then run the quick flow on OCR text
     const fd = new FormData();
     fd.append('image', f);
@@ -235,15 +244,96 @@ async function fetchTimeout(url, opts = {}, ms = 60000) {
   try { return await fetch(url, { ...opts, signal: c.signal }); }
   finally { clearTimeout(t); }
 }
+
+// ---- Barcode-first decoding (ZXing) ----
+// A decoded barcode is checksum-verified (Code128 mod-103), so it is
+// essentially never wrong — far more reliable than OCR text for labels.
+let zxingPromise = null;
+function loadZxing() {
+  if (window.ZXing) return Promise.resolve();
+  if (!zxingPromise) {
+    zxingPromise = new Promise((res, rej) => {
+      const s = document.createElement('script');
+      s.src = 'https://cdn.jsdelivr.net/npm/@zxing/library@0.21.3/umd/index.min.js';
+      s.onload = res;
+      s.onerror = () => rej(new Error('barcode reader failed to load'));
+      document.head.appendChild(s);
+    });
+  }
+  return zxingPromise;
+}
+
+// Strict client-side check: only barcode reads that look exactly like a
+// tracking number are trusted (the server re-validates anyway).
+function strictTracking(s) {
+  s = String(s || '').toUpperCase().replace(/[\s-]/g, '');
+  if (/^1Z[0-9A-Z]{16}$/.test(s)) return s;          // UPS
+  if (/^9\d{19,21}$/.test(s)) return s;               // USPS
+  if (/^(\d{12}|\d{15}|\d{20})$/.test(s)) return s;    // FedEx
+  if (/^\d{10,11}$/.test(s)) return s;                // DHL
+  return null;
+}
+
+async function decodeOne(reader, url, ms = 9000) {
+  try {
+    const img = new Image();
+    img.src = url;
+    await img.decode();
+    const p = reader.decodeFromImageElement(img);
+    const r = await Promise.race([p, new Promise((_, rej) => setTimeout(() => rej(new Error('decode timeout')), ms))]);
+    return r.getText();
+  } catch (e) { return null; }
+}
+
+// Whole-label decode usually fails (barcodes too small in frame), so try
+// horizontal region strips, upscaled — the 1Z barcode lives low on UPS labels.
+async function decodeBarcode(file) {
+  try { await loadZxing(); } catch (e) { return null; }
+  let bmp;
+  try { bmp = await createImageBitmap(file); } catch (e) { return null; }
+  const hints = new Map([[ZXing.DecodeHintType.TRY_HARDER, true]]);
+  const reader = new ZXing.BrowserMultiFormatReader(hints);
+  const regions = [
+    [0, 0, 1, 1],        // full image
+    [0, 0.35, 1, 0.35],  // middle strip
+    [0, 0.55, 1, 0.40],  // lower strip (UPS 1Z barcode zone)
+  ];
+  try {
+    for (const [fx, fy, fw, fh] of regions) {
+      const sw = Math.round(bmp.width * fw), sh = Math.round(bmp.height * fh);
+      if (sw < 60 || sh < 60) continue;
+      const scale = Math.max(1, 1200 / sw);
+      const canvas = document.createElement('canvas');
+      canvas.width = Math.round(sw * scale);
+      canvas.height = Math.round(sh * scale);
+      canvas.getContext('2d').drawImage(bmp, bmp.width * fx, bmp.height * fy, sw, sh, 0, 0, canvas.width, canvas.height);
+      const url = canvas.toDataURL('image/png');
+      const text = await decodeOne(reader, url);
+      const num = strictTracking(text);
+      if (num) return num;
+    }
+  } finally { bmp.close(); }
+  return null;
+}
 async function handlePastedImageFile(f) {
   const target = $('quick-result');
   try {
     // Instant feedback: show a thumbnail + progress the moment the image lands
     f = await shrinkImage(f);
     const url = URL.createObjectURL(f);
-    target.innerHTML = `<div class="review"><img src="${url}" style="max-width:100%;border-radius:10px"><p class="muted" id="ocr-live">📖 Reading screenshot…</p></div>`;
-    const text = await ocrFile(f, $('ocr-status-paste'));
+    target.innerHTML = `<div class="review"><img src="${url}" style="max-width:100%;border-radius:10px"><p class="muted" id="ocr-live">📷 Scanning barcode…</p></div>`;
     const live = $('ocr-live');
+    // 1) Barcode first: checksum-verified, essentially never wrong
+    const barcode = await decodeBarcode(f);
+    let text;
+    if (barcode) {
+      text = barcode;
+      if (live) live.textContent = '📷 Barcode read — verifying…';
+    } else {
+      // 2) Fallback: OCR the printed text
+      if (live) live.textContent = '📖 Reading screenshot…';
+      text = await ocrFile(f, $('ocr-status-paste'));
+    }
     if (live) live.textContent = '⬆️ Uploading…';
     const fd = new FormData();
     fd.append('image', f);
@@ -253,6 +343,7 @@ async function handlePastedImageFile(f) {
     if (!up.ok) throw new Error('upload failed');
     if (live) live.textContent = '🔍 Finding tracking number…';
     const r = await runQuick(text, target);
+    if (barcode && 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);
   } catch (err) { $('ocr-status-paste').textContent = ''; target.innerHTML = `<p class="error">${esc(err.message)}</p>`; }
 }

package/src/extract.js

@@ -41,14 +41,43 @@
   function extractContext(text) {
     var out = { seller_name: null, device: null };
     if (!text) return out;
+    // A plausible person name: no digits (addresses have them), not ALL CAPS
+    // (shipping-label blocks are), proper name shape. Shipping labels like
+    // "ORANGE NJ 07050 DWT: 11,9,7" must never land in the seller field.
+    function looksLikeName(s) {
+      s = String(s).trim();
+      if (!s || s.length > 40) return false;
+      if (/\d/.test(s)) return false;
+      if (/^[A-Z\s.,'-]+$/.test(s)) return false;
+      if (!/^[A-Za-z][A-Za-z .,'-]*$/.test(s)) return false;
+      return true;
+    }
     // WhatsApp export style: "[10/12/26, 9:41 AM] John Smith: message"
     var m = text.match(/^\[?[^\]\n]*\]?\s*([^:\[\]\n]{2,40}):\s/m);
-    if (m && !/^(you|me|http)/i.test(m[1].trim())) out.seller_name = m[1].trim();
+    if (m && !/^(you|me|http)/i.test(m[1].trim()) && looksLikeName(m[1])) out.seller_name = m[1].trim();
     // "from John", "seller: John"
     if (!out.seller_name) {
       var m2 = text.match(/(?:from|seller)\s*:\s*([A-Z][a-z]+(?:\s+[A-Z][a-z]+)?)/);
       if (m2) out.seller_name = m2[1].trim();
     }
+    // Shipping labels: the sender name is the first or second line
+    // ("MOHAMMED SALEM"). Only take it when it's clearly a name — never
+    // digits, never address fragments, never OCR garbage like "EE SS".
+    if (!out.seller_name) {
+      var labelWords = function (s) { return String(s).trim().split(/\s+/); };
+      var looksLikeLabelName = function (s) {
+        var words = labelWords(s);
+        if (words.length < 2 || words.length > 4) return false;
+        for (var i = 0; i < words.length; i++) if (!/^[A-Za-z]{3,}$/.test(words[i])) return false;
+        if (/\b(UPS|FEDEX|USPS|DHL|GROUND|EXPRESS|PRIORITY|BILLING|SHIP\b|FROM|TRACKING|POSTAGE|LABEL)\b/i.test(s)) return false;
+        return true;
+      };
+      var lines = String(text).split(/\r?\n/);
+      for (var li = 0; li < Math.min(2, lines.length); li++) {
+        var cleaned = lines[li].replace(/[^A-Za-z ]/g, ' ').replace(/\s+/g, ' ').trim();
+        if (looksLikeLabelName(cleaned)) { out.seller_name = cleaned; break; }
+      }
+    }
     // Device: iPhone 15 Pro Max 256GB, iPad Pro 11, etc.
     var dm = text.match(/\b(iPhone\s+\d{1,2}\s*(?:Pro\s*Max|Pro|Plus|Mini|e)?|iPad\s*(?:Pro|Air|Mini)?\s*\d*(?:\.\d)?)\b/i);
     var sm = text.match(/\b(64|128|256|512)\s?GB\b/i) || text.match(/\b1\s?TB\b/i);