fix: letter A at end of string is not a stopword (#11095)

I think this may be the reason for
#11014

E120 and E124 have synonyms "Cochineal Red" and "Cochineal Red A".
In the ingredients taxonomy, "a" is a stopword.
E124 becomes a synonym of E120.

Solution: we don't consider A / a to be a stopword when it is at the end
of a string, as in English, "a" comes before a noun.

lib/ProductOpener/Tags.pm

@@ -936,6 +936,13 @@ sub remove_stopwords ($tagtype, $lc, $tagid) {
 
 		my $uppercased_stopwords_overrides = 0;
 
+		if ($lc eq 'en') {
+			# in English, "a" is a stopwords for ingredients, but we do not want to remove it at the end of a tag
+			# e.g. "Cochineal Red A" -> "cochineal-red-a" --> "a" should not be a stopword
+			$tagid =~ s/a$/A/;
+			$uppercased_stopwords_overrides = 1;
+		}
+
 		if ($lc eq 'fr') {
 			# "Dés de tomates" -> "des-de-tomates" --> "dés" should not be a stopword
 			$tagid =~ s/\bdes-de\b/DES-DE/g;
@@ -1099,7 +1106,7 @@ sub get_file_from_cache ($source, $target) {
 # e.g. if the taxonomy building algorithm or configuration has changed
 # This needs to be done also when the unaccenting parameters for languages set in Config.pm are changed
 
-my $BUILD_TAGS_VERSION = "20240828 - new [tagtype].extended.json format with normalized extended synonyms";
+my $BUILD_TAGS_VERSION = "20241206 - the letter A at the end of an entry should not be a stopword in English";
 
 sub get_from_cache ($tagtype, @files) {
 	# If the full set of cached files can't be found then returns the hash to be used

tests/unit/expected_test_results/ingredients/en-ingredient-ending-with-a.json

@@ -0,0 +1,119 @@
+{
+   "ingredients" : [
+      {
+         "id" : "en:e124",
+         "is_in_taxonomy" : 1,
+         "percent_estimate" : 60,
+         "percent_max" : 100,
+         "percent_min" : 20,
+         "text" : "e124",
+         "vegan" : "yes",
+         "vegetarian" : "yes"
+      },
+      {
+         "id" : "en:e124",
+         "is_in_taxonomy" : 1,
+         "percent_estimate" : 20,
+         "percent_max" : 50,
+         "percent_min" : 0,
+         "text" : "Ponceau 4R",
+         "vegan" : "yes",
+         "vegetarian" : "yes"
+      },
+      {
+         "id" : "en:e124",
+         "is_in_taxonomy" : 1,
+         "percent_estimate" : 10,
+         "percent_max" : 33.3333333333333,
+         "percent_min" : 0,
+         "text" : "Cochineal Red A",
+         "vegan" : "yes",
+         "vegetarian" : "yes"
+      },
+      {
+         "id" : "en:e120",
+         "is_in_taxonomy" : 1,
+         "percent_estimate" : 5,
+         "percent_max" : 25,
+         "percent_min" : 0,
+         "text" : "Cochineal Red",
+         "vegan" : "no",
+         "vegetarian" : "no"
+      },
+      {
+         "ecobalyse_code" : "pear-eu",
+         "id" : "en:pear",
+         "is_in_taxonomy" : 1,
+         "percent_estimate" : 5,
+         "percent_max" : 20,
+         "percent_min" : 0,
+         "text" : "a pear",
+         "vegan" : "yes",
+         "vegetarian" : "yes"
+      }
+   ],
+   "ingredients_analysis" : {
+      "en:non-vegan" : [
+         "en:e120"
+      ],
+      "en:non-vegetarian" : [
+         "en:e120"
+      ]
+   },
+   "ingredients_analysis_tags" : [
+      "en:palm-oil-free",
+      "en:non-vegan",
+      "en:non-vegetarian"
+   ],
+   "ingredients_hierarchy" : [
+      "en:e124",
+      "en:e120",
+      "en:pear",
+      "en:fruit"
+   ],
+   "ingredients_lc" : "en",
+   "ingredients_n" : 5,
+   "ingredients_n_tags" : [
+      "5",
+      "1-10"
+   ],
+   "ingredients_original_tags" : [
+      "en:e124",
+      "en:e124",
+      "en:e124",
+      "en:e120",
+      "en:pear"
+   ],
+   "ingredients_percent_analysis" : 1,
+   "ingredients_tags" : [
+      "en:e124",
+      "en:e120",
+      "en:pear",
+      "en:fruit"
+   ],
+   "ingredients_text" : "E124, Ponceau 4R, Cochineal Red A, Cochineal Red, a pear",
+   "ingredients_with_specified_percent_n" : 0,
+   "ingredients_with_specified_percent_sum" : 0,
+   "ingredients_with_unspecified_percent_n" : 5,
+   "ingredients_with_unspecified_percent_sum" : 100,
+   "ingredients_without_ciqual_codes" : [
+      "en:e120",
+      "en:e124",
+      "en:pear"
+   ],
+   "ingredients_without_ciqual_codes_n" : 3,
+   "ingredients_without_ecobalyse_ids" : [
+      "en:e120",
+      "en:e124"
+   ],
+   "ingredients_without_ecobalyse_ids_n" : 2,
+   "known_ingredients_n" : 4,
+   "lc" : "en",
+   "nutriments" : {
+      "fruits-vegetables-legumes-estimate-from-ingredients_100g" : 5,
+      "fruits-vegetables-legumes-estimate-from-ingredients_serving" : 5,
+      "fruits-vegetables-nuts-estimate-from-ingredients_100g" : 5,
+      "fruits-vegetables-nuts-estimate-from-ingredients_serving" : 5
+   },
+   "unknown_ingredients_n" : 0
+}

tests/unit/ingredients.t

@@ -859,6 +859,17 @@ puffed orange and caramelized unknown_fruit4.",
 			ingredients_text => "some unknown ingredient and salt",
 		}
 	],
+
+	# Do not consider A at the end of the string to be a stopword
+	# https://github.com/openfoodfacts/openfoodfacts-server/pull/11095
+	[
+		"en-ingredient-ending-with-a",
+		{
+			lc => "en",
+			ingredients_text => "E124, Ponceau 4R, Cochineal Red A, Cochineal Red, a pear",
+		}
+	],
+
 );
 
 foreach my $test_ref (@tests) {