text search

utils/geo.go

@@ -83,6 +83,7 @@ func SearchTokens(label string) []string {
 	var tokens []string
 	seen := make(map[string]bool)
 	kept := 0 // multi-character tokens so far: a lone letter needs one
+	var filler []string // packaging words, kept only if nothing else survives
 	for _, raw := range strings.FieldsFunc(strings.ToLower(label), func(r rune) bool {
 		return !(r >= 'a' && r <= 'z' || r >= '0' && r <= '9')
 	}) {
@@ -92,15 +93,39 @@ func SearchTokens(label string) []string {
 		if len(raw) < 2 && (kept == 0 || isMultiplier(raw)) {
 			continue
 		}
+		seen[raw] = true
+		if isPackaging(raw) {
+			filler = append(filler, raw)
+			continue
+		}
 		if len(raw) > 1 {
 			kept++
 		}
-		seen[raw] = true
 		tokens = append(tokens, raw)
 	}
+	// "Dettol bottle pack" is a scan of Dettol. Only when the label is nothing
+	// but packaging does that packaging become the search.
+	if len(tokens) == 0 {
+		return filler
+	}
 	return tokens
 }
 
+// isPackaging is what the label says about the wrapper rather than the
+// product: "Dettol bottle pack", "Parle G biscuit pack". Lens reads these off
+// the packet and no catalogue name carries them, so every one of them used to
+// be a word the row had to contain — and "Dettol bottle pack" found no Dettol
+// at all. Treated like pack sizes: real words, just not the product's name.
+func isPackaging(tok string) bool {
+	switch tok {
+	case "pack", "packs", "packet", "packets", "bottle", "bottles",
+		"box", "boxes", "jar", "jars", "tin", "tins", "pouch", "pouches",
+		"carton", "cartons", "sachet", "sachets", "combo", "refill":
+		return true
+	}
+	return false
+}
+
 // isMultiplier is the "x" of "2 x 50gm" and the "n" of a multipack — a single
 // character that joins sizes rather than naming a product.
 func isMultiplier(tok string) bool {

utils/geo_test.go

@@ -102,3 +102,29 @@ func TestFoldAndTightenSeparators(t *testing.T) {
 		t.Fatalf("tighten: got %q", got)
 	}
 }
+
+// Lens reads the wrapper as well as the product. No catalogue name carries
+// "bottle" or "pack", so requiring them found no Dettol at all.
+func TestSearchTokensDropsPackagingWords(t *testing.T) {
+	for _, tc := range []struct {
+		label string
+		want  []string
+	}{
+		{"Dettol bottle pack", []string{"dettol"}},
+		{"Parle G biscuit pack", []string{"parle", "g", "biscuit"}},
+		{"Milk Bikis pack", []string{"milk", "bikis"}},
+		{"Nescafe jar 50g", []string{"nescafe"}},
+		{"pack", []string{"pack"}},                // nothing else: the wrapper is the search
+		{"combo pack", []string{"combo", "pack"}}, // ditto, both kept
+	} {
+		got := SearchTokens(tc.label)
+		if len(got) != len(tc.want) {
+			t.Fatalf("%q: got %v want %v", tc.label, got, tc.want)
+		}
+		for i := range tc.want {
+			if got[i] != tc.want[i] {
+				t.Fatalf("%q: got %v want %v", tc.label, got, tc.want)
+			}
+		}
+	}
+}