Commercial food translation and image scraping

[bglusman]

@komizutama and I talked about this some this morning and this is mostly a placeholder to possibly attract ideas and/or collaborators (ping @davidbella discussed a few days ago! and @dmlicht in case interested! Also @brittlewis12 any chance you or plated can help with this?)

We currently rely on USDA database, but its not very consumer-friendly in its name, doesnt have images, and would be a massive task to translate (see #1) whereas retailers like amazon or freshdirect probably stock most or all of the foods on offer, have images associated, consumer friendly food titles, and translations of those foods into many languages.

Ideally we might integrate with the USDA db and or the nutes repo (https://github.com/openpantry/nutes) but whoever is doing should also feel free to just focus on creating a new food db another way if easier/more practical... theres a lot to do here and many questions to answer on how best it would work, and outside any of the maintainers core skills, so appreciate ideas and feedback!

[bglusman]

Oh snap! just found this with a google, dunno how I missed it before? looks perfect at a glance, wish I'd started out using this instead of USDA data!
https://world.openfoodfacts.org/
https://us.openfoodfacts.org/contribute
https://github.com/openfoodfacts
https://us.openfoodfacts.org/data

[teolemon]

Hello :-) Open Food Facts here. We're catching up on email. :-P

[bglusman]

Yay! I haven't had a chance to play around with integrating/importing your downloaded data, and not totally sure the state of it yet, but advice/help/and or collaboration welcome! We're particularly interested in connecting your names (in English and other languages were available) and pictures to existing USDA IDs/entries if possible, or replacing our food table with something imported/adapted from you guys. Barcodes would be amazing also if possible but no concrete use today until we have up scanner integration or something.

[bglusman]

There's a slightly old DB dump of ours linked in readme by the way, currently images are on a different table (stocks not foods) and not in that dump but that's a minor point

[teolemon]

We have already imported the USDA branded database.
If you're talking about the USDA generic profiles, they are connected to Wikidata, to which our category system is also connected.
Also, Open Food Facts is all about crowdsourcing, so you can get your users to take pictures of missing products, and send them our way. Or you can use the official OFF app to do that

Let's continue the discussion on Slack. :-)

[teolemon]

https://us.openfoodfacts.org/contributor/usda-ndb-import

[bglusman]

Awesome! We have a slack but happy to join yours (I don't actually have control of ours anyway). Thanks!