cbizon
commented
2 years ago Notes from Brett via slack follow. I've added the mappings themselves to Babel/input_data/foodb-output.json
Hi Chris, I've attached some Food Ontology mappings of FooDB
These were made possible with a collaborator on the SPOKE NSF project, Josef Hardi in the Protege Group at Stanford, as well as Ebru Akbas in Sergio's group at UCSF.
They took a long time evaluating different food databases and their mappings and ultimately decided to write a custom mapper. If you'd like to try it or integrate it into the the node normalizer / Babel build process, it can now be found here: https://github.com/protegeteam/foodmapr If you choose to use that, you'll have to use the branch that's available as a PR. Unfortunately, the master branch doesn't build as-is at the moment.
The output mappings aren't perfect, but they're pretty good. You might want to review them by eye to see if they're sufficiently good for the NN.
The source data that I used for these can be found https://foodb.ca/downloadshttps://foodb.ca/downloads
FYI, there are also mappings to Food Standards Australia New Zealand. They recently released a version 2, which isn't in SPOKE just yet. I can provide you the version 1 mappings now or version 2 soon.
Spoke has fooddb to foodon identifer mappings - add them to nodenorm