From 6250c9f16369e156d12d5fa943e7d4ca8ab3809c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?St=C3=A9phane=20Gigandet?= <stephane@openfoodfacts.org>
Date: Fri, 5 Jan 2024 16:25:57 +0100
Subject: [PATCH] fix: remove warnings for empty/undef quantities (#9573)

---
 lib/ProductOpener/Units.pm | 44 ++++++++++++++++++++++++--------------
 tests/unit/units.t         | 21 ++++++++++--------
 2 files changed, 40 insertions(+), 25 deletions(-)

diff --git a/lib/ProductOpener/Units.pm b/lib/ProductOpener/Units.pm
index ee3d042e2fe79..b2ee561a16add 100644
--- a/lib/ProductOpener/Units.pm
+++ b/lib/ProductOpener/Units.pm
@@ -248,17 +248,19 @@ sub parse_quantity_unit ($quantity, $standard_unit_bool = undef) {
 	# 6 bricks de 1 l
 	# 10 unités, 170 g
 	# 4 bouteilles en verre de 20cl
-	if ($quantity
-		=~ /(?<number>\d+)(\s(\p{Letter}| )+)?(\s)?( de | of |x|\*)(\s)?(?<quantity>$number_regexp)(\s)?(?<unit>$units_regexp)\b/i
-		)
-	{
-		$m = $+{number};
-		$q = lc($+{quantity});
-		$u = $+{unit};
-	}
-	elsif ($quantity =~ /(?<quantity>$number_regexp)(\s)?(?<unit>$units_regexp)\s*\b/i) {
-		$q = lc($+{quantity});
-		$u = $+{unit};
+	if (defined $quantity) {
+		if ($quantity
+			=~ /(?<number>\d+)(\s(\p{Letter}| )+)?(\s)?( de | of |x|\*)(\s)?(?<quantity>$number_regexp)(\s)?(?<unit>$units_regexp)\b/i
+			)
+		{
+			$m = $+{number};
+			$q = lc($+{quantity});
+			$u = $+{unit};
+		}
+		elsif ($quantity =~ /(?<quantity>$number_regexp)(\s)?(?<unit>$units_regexp)\s*\b/i) {
+			$q = lc($+{quantity});
+			$u = $+{unit};
+		}
 	}
 
 	return ($q, $m, $u);
@@ -303,13 +305,23 @@ Returns undef if no unit was detected.
 =cut
 
 sub extract_standard_unit ($quantity_field) {
+
+	my $standard_unit = undef;
+
 	my (undef, undef, $unit) = parse_quantity_unit($quantity_field);
 
-	# search in the map of all synonyms in all languages ($units_names)
-	$unit = lc($unit);
-	my $unit_id = $units_names{$unit};    # $unit_id can be undefined
+	if (defined $unit) {
+
+		# search in the map of all synonyms in all languages ($units_names)
+		$unit = lc($unit);
+		my $unit_id = $units_names{$unit};    # $unit_id can be undefined
+		if (defined $unit_id) {
+
+			$standard_unit = $units{$unit_id}{standard_unit};    # standard_unit can be undefined
+		}
+	}
 
-	return $units{$unit_id}{standard_unit};    # standard_unit can be undefined
+	return $standard_unit;
 }
 
 =head2 normalize_serving_size($serving)
@@ -324,7 +336,7 @@ sub normalize_serving_size ($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>$number_regexp)( )?(?<unit>$units_regexp)\b/i) {
+	if ((defined $serving) and ($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
index bc835523f6df0..94437c28ee71c 100644
--- a/tests/unit/units.t
+++ b/tests/unit/units.t
@@ -151,19 +151,22 @@ 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"],
-	["1.5kg", "1500"],
-	["2,5g", "2.5"],
-	["1 plate (25g)", "25"],
-	["1 grilled link (82g)", "82"],
-	["2 buns = 20g", "20"],
-	["43 someinvalidunit (430g)", "430"],
-	["1500ml", "1500"],
+	[undef, undef, undef],
+	["", undef, undef],
+	["100g", "100", "g"],
+	["250 g", "250", "g"],
+	["1.5kg", "1500", "g"],
+	["2,5g", "2.5", "g"],
+	["1 plate (25g)", "25", "g"],
+	["1 grilled link (82g)", "82", "g"],
+	["2 buns = 20g", "20", "g"],
+	["43 someinvalidunit (430g)", "430", "g"],
+	["1500ml", "1500", "ml"],
 );
 
 foreach my $test_ref (@serving_sizes) {
 	is(normalize_serving_size($test_ref->[0]), $test_ref->[1]) or diag explain $test_ref;
+	is(extract_standard_unit($test_ref->[0]), $test_ref->[2]) or diag explain $test_ref;
 }
 
 done_testing();