Thanks! This seems fine to me, but deferring to @nvkelso if this is cool to add in.

Mind adding a quick test for this? You should be able to copy an existing one from the integration-test directory.

Thanks for tackling this! Please update it as detailed below and we can get this merged ASAP :)

Because direction is sometimes not provided as a numerical degree (eg a value of 90 meaning 90°), like E for "east" = 90 and so on as detailed #598 (comment), we'll also need to add a new vectordatasource transform in Python to ensure values are in numerical range of 0-359 after also converting valid string input to numerical values. If the input isn't valid we should drop the property.

An existing transform to model off is vectordatasource.transform.height_to_meters or vectordatasource.transform.pois_capacity_int.

Note there's also a helper function to _to_float_meters that we'd also need to duplicate (and call something like _to_int_degrees (we could do float, but if so we'd only want to the tenths place?).

Once the transform Python code is added, we also need to trigger it by adding new line after:

• https://github.com/tilezen/vector-datasource/blob/v1.8.0/queries.yaml#L277

So every feature in the POIs layer has the transform applied.

Then we'd need to add a test like Rob suggested... and make sure the test file tested for valid numerical, out of range numerical, and E > 90 conversions.