[WIP] Complete task-schema.json

[annakrystalli]

Pull request resolves:

• Create JSON schema for hub tasks.json #1
• Add round name information #5
• Add round_id_from_variable property to tasks-schema.json #7
• Handle mean and median type_id specifications more efficiently? #9 (partially, still waiting for agreement on best approach)

Have also added additional field described in hubDocs and the googledoc. It did raise a question (hubverse-org/hubDocs#13) regarding the target task id which was not present in the docs but is in all the examples and two variables (outcome_variable and outcome_modifier) which were in the docs but not in any of the examples. Would be great to get some clarity on that question :)

Not sure I've captured the submission due description information correctly!

And any feedback on the waffly and perhaps imprecise description fields very welcome!

[LucieContamin]

Is this "target_date" information required? My understanding is that it would be redundant with "origin_date" and "horizon".

[annakrystalli]

My understanding from the documentation on the Task ID variables in the Data Formats Overview page where target_date is listed as a task_id is that there are a few ways to specify target date and the option to include it directly is acceptable? The docs also include the following note:

We encourage hubs to avoid redundancy in the model task columns. For example, Hubs should not include all three of target_date, origin_date, and horizon as task id columns because if any two are specified, the third can be calculated directly. Similarly, if a variable is constant, it should not be included. For example, if a Hub doesn’t include multiple outcome variables, an outcome variable should not be included among the task id columns.

Perhaps a decision was made that isn't reflected in the docs?

[nickreich]

I think target_date should not be required but should be one of several optional, valid ways to specify dates.

[nickreich]

Suggested change

	"description": "The minimum inclusive valid cumulative distribution function value",
	"description": "The minimum inclusive valid cumulative distribution function value.",

[elray1]

Is there ever a setting in which a hub would want to specify a minimum cdf value other than 0, or a maximum cdf value other than 1? My inclination is that this is kind of a strange thing to let hubs specify, but maybe I'm not thinking of the use case.

[annakrystalli]

Indeed. I wasn't sure whether such a use case would exist so left it as a possibility. But happy to hard code 0 & 1 into the checks

[elray1]

Is there ever a setting in which a hub would want to specify a minimum cdf value other than 0, or a maximum cdf value other than 1? My inclination is that this is kind of a strange thing to let hubs specify, but maybe I'm not thinking of the use case.

[elray1]

same comment as above

Suggested change

	"minimum": 0,
	"maximum": 1

[annakrystalli]

Thank you for the reviews and commit suggestions @LucieContamin, @nickreich & @elray1 ! And apologies for the messy commit history from accepting them one by one 🤦‍♀️. I'll go ahead and implement some of the agreed changes.

[annakrystalli]

As suggested by @nickreich , merging so we can start afresh using issues.