Typo in EAN, how to handle ?

[serpico]

Problem

Might be related to #182

I noticed/realized 3564700010546 was in fact 3564700010846 it'd be nice if the prices could be connected to the correct EAN, maybe some kind of merging process.

What would be/will be the process in such case ?

[raphael0202]

I implemented on openfoodfacts-python library a function that verifies the checksum of the EAN: https://github.com/openfoodfacts/openfoodfacts-python/blob/develop/openfoodfacts/barcode.py#L20
It could be used to display an error message when the barcode is invalid.

[raphodn]

Would you go as far as not allowing any prices for barcodes that don't validate the EAN rule ?

It could be easily implemented. Or we could run checks afterwards and flag "wrong" products.

By the way there is already a warning message in the frontend price addition form. But nothing in the mobile. I would tend to keep it on the user side, no ?

[raphael0202]

It could be easily implemented. Or we could run checks afterwards and flag "wrong" products.

Why not preventing price addition completely.

By the way there is already a warning message in the frontend price addition form. But nothing in the mobile. I would tend to keep it on the user side, no ?

I agree, digit checksum is easy to compute, it can be done client-side.

[raphodn]

Why not preventing price addition completely.

Just to be sure, is this check already done in Product Opener ?
I fear the edge cases where some (local) producers have "wrong" barcodes.

[raphael0202]

Just to be sure, is this check already done in Product Opener ?

It's not!

I fear the edge cases where some (local) producers have "wrong" barcodes.

That's a valid point, but is it possible to generate a barcode image that's recognized by barcode scanners and not valid with respect to the checksum?

ping @teolemon

[serpico]

@raphael0202 I had a real life case just yesterday , I squinted beyond reason to finally decipher 3450970151458 from a picture I took the same day, but I did a quick duckduckgo search to firm it up using a description of the product and the brand which returned 3450970151153 ( the correct one ). I was wondering how the openfoodfacts-python library would fare in that case ? and what would be displayed to the user ?