Add observation attributes

[jframi]

In the analytics discussion we have been discussing about how to handle summary statistics for a given variable in a study (e.g. heritability, CV, Wald p-value, etc.)
One option would be to define an observation at the study level for the same variable having the study general mean as value.
If we can define observation attributes, as a combination of a variableDbId (each statistic being defined as a variable) and a value, summary statistics could be handled this way.

[daveneti]

We have generalised 'attributes' in our BrAPI extensions. Now we can apply attrubutes to other enitites as well as Germplasm, for example InventoryLot (SeedLot) and ObservationUnit. Would you like me to share our extensions?

[jframi]

Below is an example of how this could be implemented

POST /observationunits

[
  {
    "locationDbId": "0e208b20",
    "locationName": "Field Station Alpha",
    "observationUnitName": "Purple_Tomato_1",
    "observationUnitPUI": "http://pui.per/plot/1a9afc14",
    "observationUnitPosition": {
      "entryType": "TEST",
      "geoCoordinates": {
        "geometry": {
          "coordinates": [
            -76.506042,
            42.417373,
            123
          ],
          "type": "Point"
        },
        "type": "Feature"
      },
      "observationLevel": {
        "levelCode": "study_1",
        "levelName": "study", <-- observation level is study
        "levelOrder": 0
      },
    },
    "programDbId": "2d763a7a",
    "programName": "The Perfect Breeding Program",
    "studyDbId": "9865addc",
    "studyName": "Purple_Tomato_1",
    "trialDbId": "776a609c",
    "trialName": "Purple Tomato"
  }
]

POST /observations

[
  {
    "collector": "917d3ae0",
    "observationTimeStamp": "2024-04-17T12:47:42.359Z",
    "observationUnitDbId": "598111d4",
    "observationUnitName": "Study name",
    "observationVariableDbId": "c403d107",
    "observationVariableName": "Plant Height in meters",
    "studyDbId": "ef2829db",
    "uploadedBy": "a2f7f60b",
    "value": "350", <-- this is the study average of plant height
    "observationAttributes": [
       {
	    "observationAttributeAttributeDbId": "101290",
	    "observationAttributeAttributeName": "Heritability", <-- Heritability as an observation attribute
	    "value": "0.9"
      },
      {
	    "observationAttributeAttributeDbId": "101291",
	    "observationAttributeAttributeName": "CV", <-- Any other summary statistics as observation attributes
	    "value": "13.2"
      }   	
    ]
  }
]

[BrapiCoordinatorSelby]

I'm not including this in V2.2 because the analytics working group has just been restarted and might have opinions on how these sorts of averages and output metadata should be done. I'm a little concerned about adding observationAttributes to the observations model directly, since 99% of all observations won't need this field. This is a push in the right direction for handling averages and similar combination analysis results, but I think we can do better.

@daveneti I would be curious about your generalized attribute system if you are able to share how you structured it.