hh-shipments 0.3.23 → 0.3.24

package/package.json

@@ -18,7 +18,7 @@
     "test": "node test/run-tests.js && node test/run-carriers-tests.js"
   },
   "type": "commonjs",
-  "version": "0.3.23",
+  "version": "0.3.24",
   "bin": {
     "hh-shipments": "server.js"
   },

package/public/app.js

@@ -382,24 +382,41 @@ async function decodeBarcode(file) {
   try { bmp = await createImageBitmap(file); } catch (e) { return null; }
   const hints = new Map([[ZXing.DecodeHintType.TRY_HARDER, true]]);
   const reader = new ZXing.BrowserMultiFormatReader(hints);
+  // Aggressive region/scale sweep: labels photographed at angles need multiple
+  // crops. The UPS 1Z barcode is typically in the lower half.
   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)
+    [0, 0, 1, 1],
+    [0, 0.30, 1, 0.40],
+    [0, 0.50, 1, 0.45],
+    [0, 0.60, 1, 0.35],
+    [0.05, 0.55, 0.90, 0.40],
   ];
   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;
+      for (const scale of [1, 1.5, 2]) {
+        const sw = Math.round(bmp.width * fw), sh = Math.round(bmp.height * fh);
+        if (sw < 80 || sh < 40) continue;
+        const canvas = document.createElement('canvas');
+        canvas.width = Math.round(sw * scale);
+        canvas.height = Math.round(sh * scale);
+        const ctx = canvas.getContext('2d');
+        ctx.drawImage(bmp, bmp.width * fx, bmp.height * fy, sw, sh, 0, 0, canvas.width, canvas.height);
+        // Boost contrast: barcodes need sharp black/white separation
+        try {
+          const img = ctx.getImageData(0, 0, canvas.width, canvas.height);
+          const d = img.data;
+          for (let i = 0; i < d.length; i += 4) {
+            const v = 0.299 * d[i] + 0.587 * d[i + 1] + 0.114 * d[i + 2];
+            const nv = v < 128 ? 0 : 255;
+            d[i] = d[i + 1] = d[i + 2] = nv;
+          }
+          ctx.putImageData(img, 0, 0);
+        } catch (e) {}
+        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;

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="0.3.23">
+<meta name="app-version" content="0.3.24">
 <title>HH Shipments</title>
 <link rel="manifest" href="/manifest.webmanifest">
 <link rel="apple-touch-icon" href="/icon-192.png">

package/src/extract.js

@@ -152,20 +152,51 @@
   }
 
   // UPS number split by spaces/dashes: "1Z 9999 99 ..." (or "12 ..."/"IZ ..." misreads)
+  // Hot-zone: text right after "TRACKING" often has a dropped character
+  // (e.g. "17 1) 539 031342 3923" = 17 chars, missing the J in 1XJ).
+  // Try expanding punct confusions that may hide two chars.
+  function expandPunctHotZone(text, found, seen) {
+    var m = /TRACKING\s*#?\s*:?\s*([0-9A-Z)\]}(\[{|!\s-]{15,25})/i.exec(text);
+    if (!m) return;
+    var raw = m[1].replace(/[\s-]/g, '');
+    // Don't swallow trailing English words ("thanks", "you") as number chars
+    raw = raw.replace(/(THANK|THANKS|YOU|PLEASE)$/i, '');
+    if (/^(1Z|12|IZ|17)/i.test(raw) && (raw.length === 17 || raw.length === 16 || raw.length === 18)) {
+      var anchored = raw.replace(/^(12|IZ|17)/i, '1Z');
+      // ")" may be "XJ" collapsed (X+J look like ) when smudged)
+      var expanded = anchored.replace(/\)/g, 'XJ').replace(/\(/g, 'CC');
+      if (/^1Z[0-9A-Z]{16}$/.test(expanded)) {
+        var num = repairUpsTail(expanded.slice(2)) || expanded;
+        if (num && UPS_STRICT.test(num) && /[A-Z]/.test(num.slice(2, 8)) && !seen[num]) {
+          seen[num] = true;
+          found.push({ number: num, carrier: 'ups', confidence: 'high', ocr_fixed: true });
+        }
+      }
+    }
+  }
+
   function spacedUpsIn(text, found, seen, conf) {
-    var m, re = /\b(?:1Z|12|IZ)[\s-]?(?:[0-9A-Z][\s-]?){16}\b/gi;
+    // Anchor covers common 1Z misreads: 12, IZ, 17 (7→Z confusion)
+    var m, re = /\b(?:1Z|12|IZ|17)[\s-]?(?:[0-9A-Z)\]}(\[{|!][\s-]?){16}\b/gi;
     while ((m = re.exec(text)) !== null) {
       var ups = normalize(m[0]);
       var num = null;
-      if (/^1Z[0-9A-Z]{6}[0-9]{10}$/.test(ups)) num = ups; // fully strict
-      else if (/^1Z[0-9A-Z]{16}$/.test(ups)) {
-        // letters sitting in the digit zone are OCR confusions — repair them
-        num = repairUpsTail(ups.slice(2)) || ups;
+      // Normalize the anchor: 12/IZ/17 → 1Z
+      var anchored = ups.replace(/^(12|IZ|17)/i, '1Z');
+      if (/^1Z[0-9A-Z]{6}[0-9]{10}$/.test(anchored)) num = anchored; // fully strict
+      else if (/^1Z[0-9A-Z]{16}$/.test(anchored)) {
+        // Repair tail; if repair fails the candidate is invalid — never accept raw
+        num = repairUpsTail(anchored.slice(2));
+      }
+      else num = repairUpsZ(anchored) || repairUpsTail(anchored.slice(2));
+      // ")" can be X or J — try X variant if J mapping failed validation
+      if (!num && /\)/.test(anchored)) {
+        var xVar = anchored.replace(/\)/g, 'X');
+        if (/^1Z[0-9A-Z]{16}$/.test(xVar)) num = repairUpsTail(xVar.slice(2));
       }
-      else num = repairUpsZ(ups) || repairUpsTail(ups.slice(2));
       if (num && !seen[num]) {
         seen[num] = true;
-        found.push({ number: num, carrier: 'ups', confidence: (num !== ups && conf === 'medium') ? 'high' : conf, ocr_fixed: num !== ups });
+        found.push({ number: num, carrier: 'ups', confidence: 'high', ocr_fixed: num !== ups });
       }
     }
   }
@@ -316,6 +347,7 @@
     // pass already tried the hot zones at high confidence). Repairs are
     // strict-structure-validated, so they stay high confidence.
     spacedUpsIn(text, found, seen, 'high');
+    expandPunctHotZone(text, found, seen);
     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;