feat: dq_detect_url_in_ingredients (#9586)

openfoodfacts · Jan 30, 2024 · 5695fb4 · 5695fb4
1 parent 0708361
commit 5695fb4
Show file tree

Hide file tree

Showing 2 changed files with 33 additions and 0 deletions.
diff --git a/lib/ProductOpener/DataQualityFood.pm b/lib/ProductOpener/DataQualityFood.pm
@@ -1619,6 +1619,10 @@ sub check_ingredients ($product_ref) {
 						"en:ingredients-" . $display_lc . "-unexpected-chars-question-mark";
 				}
 
+				if ($product_ref->{$ingredients_text_lc} =~ /http/i) {
+					add_tag($product_ref, "data_quality_errors", "en:ingredients-" . $display_lc . "-unexpected-url");
+				}
+
 				# French specific
 				#if ($display_lc eq 'fr') {
 

diff --git a/tests/unit/dataqualityfood.t b/tests/unit/dataqualityfood.t
@@ -1394,6 +1394,35 @@ ok(!has_tag($product_ref, 'data_quality', 'en:nutrition-sugars-plus-starch-great
 	'sum of sugars and starch greater carbohydrates')
 	or diag explain $product_ref;
 
+# unexpected character in ingredients
+$product_ref = {
+	languages_codes => {
+		en => 1
+	},
+	lc => 'en',
+	ingredients_text_en => 'AaaAAa, BbbBBB, $, @, !, ?, Https://,',
+};
+ProductOpener::DataQuality::check_quality($product_ref);
+
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-5-vowels'), '5 vowel in a row')
+	or diag explain $product_ref;
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-5-consonants'), '5 consonants in a row')
+	or diag explain $product_ref;
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-4-repeated-chars'), '4 repeated characters')
+	or diag explain $product_ref;
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-unexpected-chars-currencies'), '$')
+	or diag explain $product_ref;
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-unexpected-chars-arobase'), '@')
+	or diag explain $product_ref;
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-unexpected-chars-exclamation-mark'), '!')
+	or diag explain $product_ref;
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-unexpected-chars-question-mark'), '?')
+	or diag explain $product_ref;
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-ending-comma'), ',')
+	or diag explain $product_ref;
+ok(has_tag($product_ref, 'data_quality', 'en:ingredients-en-unexpected-url'), 'detected url')
+	or diag explain $product_ref;
+
 # jam and related categories and fruit (specific ingredients) content
 ## missing specific ingredients
 $product_ref = {