2024-06-12 Agrisemantics Committee Meeting
Participants
@Names here
Meeting Logistics
Insert meeting info here or use:
See time and GoToMeeting details on the AgGateway Calendar.
Agenda and Minutes
Main Topics
Use decide tag for any decisions taken
Background & History
Topic Leader:
The desired outcome is that …
Pre-meeting notes and references:
Minutes:
Infrastructure
Topic Leader:
The desired outcome is that …
Pre-meeting notes and references:
Minutes:
Idea that JSON file contains simple lists like ADAPT UoM, when a pull request is made a GitHub action sends that new file to update a database
Need for the database use cases:
Once lists get big, getting only changes since last download
Being able to search a code/API call for a specific code with out having to download and search the whole JSON
Having ability to map relationships likely works better in a database/via API than a file on GitHub
Need for grain codes the more complex list of quality tests done for dockage compared to what is documented in GIPSA
Question on if there is a business model for charging for asserting/using relationships as a way to fund maintenance of the infrastructure
How does this fit in ADAPT
Could Modus codes be in the same data type definitions
How do the observation codes are kind of a superset of data type definition objects
Could have a generic data type definition object
Need to do some more work on the crisp delineations between concepts in the crop definition
e.g. is a different variety a different sub crop zinfandel vs Merlot grape are likely different than corn hybrid 1402 planted in a field
Many vocabularies will likely fit in the general observation model, however there will be some that require a different model like Crop Definition
If it model is too generic it becomes quite difficult to implement
Potential to have a very abstracted database, then create “schema on the query”
Might be best to get through a couple worked through to see what patterns emerge
Tech stack likely has implications on how things work best as well
DMN - Data Modeling and Notation standard
Has the ability to define anything but keep it structured so it makes sense
Object Management Group
Have made data type definitions its own standard
Has an active community that is good to work with and seems it might work well for what we might need
Controlled vocabularies we really need and support today are key value pairs where the value is a simple type
Do not want to start creating abstract, complex data types
Some concern that using DMN in ADAPT may be an extra intermediate step that is not needed
Within the infrastructure the codes are complex types
Need to start talking about observation code definitions, data type definitions, and context items all together - possibility to have one
Process submissions
Topic Leader:
The desired outcome is that …
Pre-meeting notes and references:
Minutes: