openfoodfacts / open-prices-frontend

A vue.js front-end for Open Prices
25 stars 13 forks source link

Same EAN different weight how to handle ? #913

Open serpico opened 2 months ago

serpico commented 2 months ago

Description

Pretty sure it came up already, 2860050040672 , there's a slight variation in weight ( 1337g in the OFF product's picture, 1242g in the picture I took today ) which would skew the weight per kilo, it feels like it should be handled like a "category" product by inputting the price per quantity when entered in OP but they use the same EAN...it's great for price tracking but I think also ripe for data quality issues, especially with casual users.

P_20241002_085220

raphodn commented 2 months ago

I'm pretty sure barcodes that start with a 2 are "shop-weight specific". So we'll probably have to nudge users to add them via category instead indeed

serpico commented 2 months ago

And it's gonna be a double whammy since the parent category is not yet included in #262.

Image

Should the script https://github.com/openfoodfacts/open-prices-frontend/blob/master/data/categories/filter_categories.py be in the wiki ?

Based on based on what you said about EAN starting with a 2 should all of them ( for example https://fr.openfoodfacts.org/categorie/foie-de-poulet ) be reprocessed ?

Would you still favor tweaking the script/filter/list on a one by one basis, essentially waiting for someone to request a product which doesn't show up when typed in the category box ? or should we add more in one go and run the risk of having too long a list ?

raphodn commented 2 months ago

I'll let @aleene answer that one ^^

aleene commented 2 months ago

You can always add a product with a barcode. Note that these are internal barcodes which can represent different products for different shops. Barcodes with enocoded weights are not yet supported. I assume that there is an issue already open for this.

serpico commented 1 month ago

see also https://prices.openfoodfacts.org/products/2644188023291 identical EAN weight varying wildly