Rounding question on the average Nutri-Score of white rice (B vs. A)

[aleene]

What

• I looked at the category white rices and saw that the average nutriscore value was -0.71. This is however rounded to Nutri-Score B.
• I would expect Nutri-Score A, with rounding to -1 or for all negative raw values.

Part of

• 🚦Nutri-Score (tracker) #5517

[hangy]

That seems to be correct, according to the specs:

openfoodfacts-server/lib/ProductOpener/Food.pm

Lines 4397 to 4421 in 25d4175

	# New grades from HCSP avis 20150602 hcspa20150625_infoqualnutprodalim.pdf
	# Tableau 1 : Seuils du score FSA retenus pour le cas général
	# Classe du 5-C
	# Bornes du score FSA
	# A/Vert Min - -1
	# B/Jaune 0 – 2
	# C/Orange 3 – 10
	# D/Rose 11 – 18
	# E/Rouge 19 – Max
	if ($fr_score <= -1) {
	$grade = 'a';
	}
	elsif ($fr_score <= 2) {
	$grade = 'b';
	}
	elsif ($fr_score <= 10) {
	$grade = 'c';
	}
	elsif ($fr_score <= 18) {
	$grade = 'd';
	}
	else {
	$grade = 'e';
	}

[stephanegigandet]

It could be interpreted either way. -0.71 is above -1 (and thus not A) and less than 0 (and thus not B). As it is an average score, I guess we could round the score to the nearest integer. Or keep the current behaviour.

[teolemon]

Is it fixed by the recode ? @stephanegigandet