Open sadielbartholomew opened 2 months ago
I am thinking of this format for the JSON schema (written and validated as valid JSON with https://jsoneditoronline.org), where <>
markers show what information is intended, with the data type first then a description:
{
"metadata": "<string, timestamp (standard format e.g. ISO8601)>",
"repository": {
"display name": "<string, name for vis. label, e.g. GH repo name>",
"URL": "<string, GH repo link>"
},
"status": {
"<numeric metric name>": {
"valid range": ["<minimum>", "<maximum>"],
"closed interval": "<Bool, True if range is closed, False if open or half open>",
"direction of health": "<Bool, True means increasing is better i.e. maximum is healthiest>"
},
"<Boolean metric name>": "<Bool, corresponds to value that is healthy>"
}
}
In short, we have some metadata to record when the information was collected, then basic info about the repo in question, then finally a dictionary recording all of the metrics, which includes the important context of which values are possible and which values represent the 'healthiest' state:
Drafted a basic schema, but need to converge on this layout in time!
The data we pull from GitHub and FAIR will be used to create JSON to capture relevant information should then be consolidated into a final output data schema which will provide all of the information needed by the visualisation tool/script to display and represent.