CRS query parameter support for EDR

[petergarnaes]

Section 8.2.1.4 of the EDR specification v. 1.0.1 (which is the newest version of writing this) states that a client can specify a crs query parameter.

Currently this parameter is not passed to the provider as this code shows, so there is no way to support other CRS than WGS84, because it is the default when not specified

In fact pygeoapi seems to break requirement A.2.12 part C which states that HTTP status code 400 should be returned if CRS is unsupported. If CRS is not passed to the provider, there is no way it can fail if CRS is unsupported. In my testing the crs query parameter is completely ignored by pygeoapi for EDR.

Describe the solution you'd like
crs should be passed to the provider as a query argument and should default to either WGS84 or WGS84h depending on if the Z-axis parameter is specified in either in the bbox query argument or z query parameter.

The config should also allow specifying supported CRS for the collection, so requirement A.57 can be adhered to.

So config would look something like:

resources:
  my_collection:
    type: collection
    title: My Collection
    ...
    crs:
     - http://www.opengis.net/def/crs/OGC/1.3/CRS84
     - native

There should also be an exception that can be thrown from a provider that results in an HTTP status code 400. Don't know if this is already supported.

[github-actions]

As per RFC4, this Issue has been inactive for 90 days. In order to manage maintenance burden, it will be automatically closed in 7 days.

[github-actions]

As per RFC4, this Issue has been closed due to there being no activity for more than 90 days.

[sjordan29]

I came across this issue while addressing #1578. This will be important to support datasets in included in your STAC, @dblodgett-usgs -- there are a number of datasets whose primary coordinates are in a projected coordinate reference system rather than WGS84. @aufdenkampe

@tomkralidis, what does the future of adding support for a crs query parameter in EDR (and subsequent CRS transformations) look like? It looks like there is support for transforming features (via the crs_transform decorator) , as well as in the rasterio_.py coverage provider (using pyproj's Transformer.