Aggregation is responsive to units.yaml

[jmcook1186]

Story

As a user I want to be able to aggregate values with shared units. I want to be able to define my units one time in one file. This is so that I can get the aggregate data that matters for my specific use-case, rather than relying on the defaults.

Rationale

We have a file called units.yaml that contains information about the unit each field in the impl is expressed in. This information should be sufficient to configure the aggregation for each impact metric. It is possible hat some additional field(s) might be required in units.yaml in order to make the aggregation work.

Implementation details

In the aggregation model plugin, the method used to aggregate a particular impact will be defined in units.yml. The model should check there to determine whether to sum or average the values as it traverses the graph.

Priority

4/5
This is next on the roadmap and must be shipped before Carbon hack

Scope

This will affect the framework itself, but not the model plugins.

Size

2 d (will be part of aggregation implementation)

What does "done" look like?

Aggregation model pulls data from units.yml correctly as demonstrated by unit tests.

Does this require updates to documentation or other materials??

Yes, there should eb amodel documentation in the model's README and information added to the main docs website.

What testing is required?

Unit tests and example impls

Is this a known/expected update?

Discussed at IEF weekly on 30/11/23