feat: Initial support for specific ingredients parsing #6243
Conversation
stephanegigandet
added
🥗 Ingredients
🚦Nutri-Score
stephanegigandet
changed the title
lib/ProductOpener/Ingredients.pm
Outdated
| # We might have an ingredient specified multiple times (e.g. once for percent, another for origins or labels) | ||
| defined $product_ref->{specific_ingredients}{$ingredient_id} or $product_ref->{specific_ingredients}{$ingredient_id} = {}; | ||
| $product_ref->{specific_ingredients}{$ingredient_id}{ingredient} = $ingredient; | ||
| $product_ref->{specific_ingredients}{$ingredient_id}{text} = $matched_text; |
There was a problem hiding this comment.
Is it ok to smash previous values ?
There was a problem hiding this comment.
I guess we could concatenate them. I didn't do it because the case is largely hypothetical and the smashed values ("ingredient" and "text") are kept mostly for debugging purposes. But it's cleaner if we concatenate, I'll change it.
t/ingredients_parsing.t
Outdated
| @@ -168,10 +168,10 @@ my @lists =( | |||
|
|
|||
| ["fr","p\x{e2}te de cacao* de Madagascar 75%, sucre de canne*, beurre de cacao*. * issus du commerce \x{e9}quitable et de l'agriculture biologique (100% du poids total).","pâte de cacao Commerce équitable Bio de Madagascar 75%, sucre de canne Commerce équitable Bio, beurre de cacao Commerce équitable Bio."], | |||
|
|
|||
| ["fr","Céleri - rave 21% - Eau, légumes 33,6% (carottes, céleri - rave, poivrons rouges 5,8% - haricots - petits pois bio - haricots verts - courge - radis, pommes de terre - patates - fenouil - cerfeuil tubéreux - persil plat)","Céleri-rave 21% - Eau, légumes 33,6% (carottes, céleri-rave, poivrons rouges 5,8% - haricots - petits pois bio - haricots verts - courge - radis, pommes de terre - patates - fenouil - cerfeuil tubéreux - persil plat)"], | |||
| ["fr","Céleri - rave 21% - Eau, légumes 33,6% (carottes, céleri - rave, poivrons rouges 5,8% - haricots - petits pois bio - haricots verts - courge - radis, pommes de terre - patates - fenouil - cerfeuil tubéreux - persil plat)","Céleri-rave 21% - Eau, légumes 33.6% (carottes, céleri-rave, poivrons rouges 5.8% - haricots - petits pois bio - haricots verts - courge - radis, pommes de terre - patates - fenouil - cerfeuil tubéreux - persil plat)"], | |||
There was a problem hiding this comment.
wow, that would be clearer with a line between entry / expected !
I did not understand you where changing the expected result at first sight.
There was a problem hiding this comment.
Agreed, I'll change it.
|
The sentences parsed with the new function do not appear in Details of the analysis of the ingredients, so it's not easy to test manually, and I also think it's a pity as it participates to nutri-score. |
|
Applied changes from code review, thank you! Also added initial support to parse things like "Origin of the milk: UK". |
I added a sub section to list the specific ingredients in the analysis. |
|
Kudos, SonarCloud Quality Gate passed!
|
teolemon
added
🚦 Nutri-Score
and removed
🚦Nutri-Score
This is in particular to resolve Nutri-Score computation errors when we fail to parse the % of fruits/vegetables or milk content indicated at the end of the ingredients list.
The resulting specific_ingredients structure is now used when computing the Nutri-Score.
It is also potentially useful to later parse other types of mentions like "Origin of something: some country" at the end of ingredients list.
Part of