fix: quantities starting with a dot .33L (#9284)

openfoodfacts · Nov 13, 2023 · 5523a15 · 5523a15
1 parent 8638175
commit 5523a15
Show file tree

Hide file tree

Showing 3 changed files with 31 additions and 4 deletions.
diff --git a/lib/ProductOpener/Numbers.pm b/lib/ProductOpener/Numbers.pm
@@ -40,13 +40,31 @@ BEGIN {
 
 		&remove_insignificant_digits
 		&convert_string_to_number
+		$number_regexp
 
 	);    # symbols to export on request
 	%EXPORT_TAGS = (all => [@EXPORT_OK]);
 }
 
 use vars @EXPORT_OK;
 
+=head1 VARIABLES
+
+=head2 $number_regexp
+
+Regular expression to match something that looks like a number:
+32
+32.5
+0.5
+.5
+32,5
+
+=cut
+
+# dot followed by digits,
+# or digits followed by a dot or a comma, optionnaly followed by 0 or more digits
+$number_regexp = '\.\d+|\d+(?:(?:\,|\.)\d+)?';
+
 =head1 FUNCTIONS
 
 =head2 remove_insignificant_digits($)

diff --git a/lib/ProductOpener/Units.pm b/lib/ProductOpener/Units.pm
@@ -245,7 +245,7 @@ sub normalize_quantity ($quantity) {
 	# 10 unités, 170 g
 	# 4 bouteilles en verre de 20cl
 	if ($quantity
-		=~ /(?<number>\d+)(\s(\p{Letter}| )+)?(\s)?( de | of |x|\*)(\s)?(?<quantity>(\d+)(\.|,)?(\d+)?)(\s)?(?<unit>$units_regexp)\b/i
+		=~ /(?<number>\d+)(\s(\p{Letter}| )+)?(\s)?( de | of |x|\*)(\s)?(?<quantity>$number_regexp)(\s)?(?<unit>$units_regexp)\b/i
 		)
 	{
 		my $m = $+{number};
@@ -254,7 +254,7 @@ sub normalize_quantity ($quantity) {
 		$q = convert_string_to_number($q);
 		$q = unit_to_g($q * $m, $u);
 	}
-	elsif ($quantity =~ /(?<quantity>(\d+)(\.|,)?(\d+)?)(\s)?(?<unit>$units_regexp)\s*\b/i) {
+	elsif ($quantity =~ /(?<quantity>$number_regexp)(\s)?(?<unit>$units_regexp)\s*\b/i) {
 		$q = lc($+{quantity});
 		$u = $+{unit};
 		$q = convert_string_to_number($q);
@@ -276,7 +276,7 @@ sub normalize_serving_size ($serving) {
 
 	# Regex captures any <number>( )?<unit-identifier> group, but leaves allowances for a preceding
 	# token to allow for patterns like "One bag (32g)", "1 small bottle (180ml)" etc
-	if ($serving =~ /^(.*[ \(])?(?<quantity>(\d+)(\.|,)?(\d+)?)( )?(?<unit>$units_regexp)\b/i) {
+	if ($serving =~ /^(.*[ \(])?(?<quantity>$number_regexp)( )?(?<unit>$units_regexp)\b/i) {
 		my $q = $+{quantity};
 		my $u = $+{unit};
 		$q = convert_string_to_number($q);

diff --git a/tests/unit/units.t b/tests/unit/units.t
@@ -141,6 +141,15 @@ is(normalize_quantity("2 kgr"), 2000);
 is(normalize_quantity("2 kilogramme"), 2000);
 is(normalize_quantity("2 kilogrammes"), 2000);
 
+# . without a 0 before
+is(normalize_quantity(".33L"), 330);
+is(normalize_quantity(".33 l"), 330);
+is(normalize_serving_size(".33L"), 330);
+is(normalize_serving_size(".33 l"), 330);
+is(normalize_serving_size("5 bottles (.33L)"), 330);
+is(normalize_serving_size("5 bottles .33L"), 330);
+is(normalize_serving_size("5 bottles2.33L"), undef);    # Broken string, missing word separator before number
+
 my @serving_sizes = (
 	["100g", "100"],
 	["250 g", "250"],
@@ -154,7 +163,7 @@ my @serving_sizes = (
 );
 
 foreach my $test_ref (@serving_sizes) {
-	is(normalize_serving_size($test_ref->[0]), $test_ref->[1]);
+	is(normalize_serving_size($test_ref->[0]), $test_ref->[1]) or diag explain $test_ref;
 }
 
 done_testing();