Infer packaging from related products in the category

[CloCkWeRX]

What

• Perhaps a robotoff task is best, to allow human confirmation; but there are a number of categories where a product has certain physical properties that require a common style of packaging.

Examples

• Milk, will come as Bottled or in a Carton, Plastic, Glass or Cardboard
• Eggs, frequently in a cardboard or plastic carton.
• Pasta sauce is typically in a glass jar or bottle.

A lot of this can probably be guessed from looking at the most common packaging for a given (specific) category.

We'd also want a way to exclude some categories, where the category is really broad like "Fruits"

Part of

• Extract packaging-related information (Tracker) #97

[teolemon]

@CloCkWeRX this is a topic I'd like to get moving asap, can you create a list of categories with strong correlations ?
and possibly for each category, an ordered list of frequency, possibly with the country (I think packaging sometimes changes from country to country)

[teolemon]

en:eggs
associated_packagings:en en:Cardboard to recycle

or perhaps easier, a collaborative spreadsheet that we can iterate on ?

[teolemon]

https://world.openfoodfacts.org/category/eggs/packaging

[CloCkWeRX]

Pipeline I'm setting up for this:

For each category
 data = fetch https://world.openfoodfacts.org/state/packaging-code-completed/category/(categoryname).json
 packaging = data.jsonpath('$.products..packaging_tags.*') # Produces something like below


 packaging_counts = {}
 packaging.each {|pkg| packaging_counts[pkg] += 1 }

which is a bit easier than doing it all manually.

What's best, ruby/python/perl/node; assuming this would run periodically?

[CloCkWeRX]

https://github.com/CloCkWeRX/openfoodfacts-packaging-from-category - python 2.7 version, simple command line tool that spits out JSON. It only goes into the first page of results right now.

[CloCkWeRX]

Okay, few more tweaks, handles pagination (stops after 3 pages), world or country, etc.

Now just need to generate the most popular categories and generate a bunch of static JSON :)

May need to let you specific "skip anything < X occurences" or "maximum products" as arguments.

[CloCkWeRX]

@raphael0202 - python isn't my default language, if you have some time would you mind looking over https://github.com/CloCkWeRX/openfoodfacts-packaging-from-category and suggesting if there are better or more idiomatic ways to structure this?

In my head, this is somewhat of a static JSON generator (just push all the results to git/easy to expose via a static web server for openfoodfacts-server to do browser queries to; or curl into elasticsearch); though could probably also publish things into an sqlite DB or other datastore (more like robotoff)

[teolemon]

@stephanegigandet