Recipe model

[ddooley]

Following last week’s Recipe modelling presentation by Usashi, I think this invites a focus on modelling of recipe from an OBOFoundry / FoodOn perspective – something that has been sorely missing. I have set up a new tab called “Recipe Defs” in the Dietary terms sheet. I invite everyone to dive in and carry out the same critique as we did with diet terms, and add more terms where you see fit. (We may need to add a supporting google doc too). The definitions for recipe, ingredient, ingredient list etc. are all very raw first drafts I fashioned from scratch just to satisfy some BFO / OBI / FoodOn class hierarchy context needs, so they are all subject to revision.
Aside from mention of a "recipe process" this doesn't really touch the myriad processes (as hinted at in the EXACT2 ontology) that could be involved in recipe preparation; I foresee that being a longer, iterative curation project.

For convenience here's a copy of the tabular sheet of terms and defs.

parent class	label	definition
protocol / plan specification	recipe	A protocol which has an objective of creating one or more food products, and has a part one or more quantitative ingredient lists, as well as a plan for executing processes that are applied to the ingredients.
--	--	--
recipe	food product formula	A recipe intended for commercial reproducibility that includes reference to specific devices, settings, food process transformations and process durations.
planned process / material processing	recipe process	A process which executes a recipe. The process outputs one or more dishes (as well as waste material). The process has participants (chef, device, operator), and inputs including ingredient list.
	food preparation device
	chef
material transformation objective	food product formulation	A material transformation objective which specifies a food product’s resulting composition in a quantitative ingredient list, but does not include a preparation process. A formulation may exclude ingredients that are chemically transformed (like baking soda), or may include or exclude ingredients (e.g. additives) according to product labelling legislation. A given formulation may be achieved by more than one recipe. In theory a food product formulation could be measured analytically.
food product	ingredient	A food product which is a member of an ingredient list.
material transformation objective	ingredient list	A list of ingredients used to make a food product.
ingredient list	quantitative ingredient list	An ingredient list that includes portion information in unit quantities, or as ratios or percentages.
quantitative ingredient list	measured ingredient list	An ingredient list that includes portion information in unit quantities.
ingredient list	ranked ingredient list	A list of ingredients whose order indicates relative proportion within an associated food product.

[ddooley]

Some context about the idea of calling a recipe a "protocol", which is a child of OBI "plan specification": The def. for protocol: "A plan specification which has sufficient level of detail and quantitative information to communicate it between investigation agents, so that different investigation agents will reliably be able to independently reproduce the process." , in other words, two cooks can reproduce the same result. About the "plan specification" class: It has an "objective specification" part, i.e. a representation of a particular dish outcome, and an "action specification" part, instructions on how to make it. I guess an ingredient list would be part of the action specification.

[cyang0128]

In the Ontology for Nutritional Epidemiology (ONE), we have included a short list of terms regarding diet and food intake measurement (e.g. diet portion size, time and seasons, validity of dietary data, etc.).
I am just wondering whether it would be nice to include the list of terms for diet description? http://purl.bioontology.org/ontology/ONE?conceptid=http%3A%2F%2Fpurl.bioontology.org%2Fontology%2FONE%2FT00137

[oldskeptic]

Is there currently a mechanism for recording consumption (which should be useful when recording dietary information) of something by someone in time and space? Right now, my instance data is typed with both schema:DrinkAction and time:TemporalEntity, but I'd entertain another approach.

[ddooley]

We could move this to another ONS or ONE related thread? I see Gene Ontology has drinking behaviour and eating behaviour as biological processes; Perhaps use "'located in' to locate them in some physical context; and perhaps a "survey time" characteristic/quality that "'has quantity' some date" would work for?

@cyang0128 has ONE settled on a more specific data model for dietary survey collection of individual consumption?

[cyang0128]

@ddooley thanks Damion, it's my pleasure to say that we have included several terms relevant to this issue. They are all under the class "quality indicator of a dietary survey" (http://purl.obolibrary.org/obo/ONE_0001001), for instance:

1. time of dietary record
   IRI: http://purl.obolibrary.org/obo/ONE_0001007
   definition: an indicator that specifies when the dietary intake was recorded. (e.g. during eating occasion, immediately after eating occasion, or later).

2. administration type of a diet recall method
   IRI: http://purl.obolibrary.org/obo/ONE_0001006
   definition: an indicator that specifies the administration type of a diet recall method. A diet recall method can be proxy administered, self-administered or interview administered.

...

looking forward to a further discussion, if you are interested in using the terms, or an update of the terms is necessary before using.
Cheers!

[ChefRobertDanhi]

What about the cultural context of a recipe, did this paper with Matthew Lange and Tarini Naravane on it.... just beginning, going to be my PhD Focus.
DECOR 2019_Title - Crowdsourced approach to extracting cultural and culinary knowledge from Recipes_PUBLISHED.pdf

[ddooley]

@ChefRobertDanhi indeed cultural context brings allot of fun, "flavour," and provenance to the experience. Great to see work in that area. From FoodOn's perspective its happy to see specialized domains created from scratch or calved off from other larger ontologies, such that FoodOn can refer to them. I think the challenge for appraising cultural dimensions beyond the "laboratory" version of what a recipe entails is the challenge of identifying one or more communities of curator interested in the related areas who want to discuss the "digitization", term by term. The joint workgroup could be a good place to start with a presentation to get colleagues involved? The one cautionary note I'd say is that if working on a phd on the subject requires some level of secrecy or ownership, this becomes a tension against the essential open source discussion required for evolving ontologies.

[oldskeptic]

Can we please pull this work back into it to align the different vues together:
https://docs.google.com/document/d/1sTLHkgZ65frihBLck1h8dEaRNlJf792XEAC-_rbIUQ4/edit

[ChefRobertDanhi]

@ddooley Can you elaborate on "The one cautionary note I'd say is that if working on a phd on the subject requires some level of secrecy or ownership, this becomes a tension against the essential open source discussion required for evolving ontologies." as I hope to house the data model, methodology and resulting database created by those cultures themselves into a shared database by an "impartial organization" such as UNESCO to help provide a platform to crowdsource information and knowledge

[ddooley]

Ok, that sounds cool and comfortable with an open source iterative ontology development. I hear masters or doctoral work can sometimes have to be under wraps until publishing, or alternately the work gets published all in one go at the end but has become a single author effort ... both those don't quite fit within a paradigm of an evolving developing, multi-agency or curator vocabulary that has maintenance resilience and momentum beyond a single creator. The cultural involvement itself sounds like it would add momentum.

[ddooley]

Latest recipe model is evolving in this process ontology paper: http://www.semantic-web-journal.net/system/files/swj3096.pdf