Closed zaneselvans closed 2 years ago
Looking at these tests (and generally mucking with the Intake catalogs) it feels like the catalogs should be encapsulated in a @dataclass
or Pydantic model. They're kind of a chaotic freeform mess right now -- not just ours, but Intake catalogs in general.
The CarbonPlan data catalog repo provides some examples of tests that could apply to our catalog as well.
These include:
to_dask()
to avoid the need to download data while verifying that EPA CEMS dataframes have the right columns.