add support for feature/record domain queries

[tomkralidis]

Overview

This PR adds support for OGC API - Features and OGC API - Records domains within /collections/{collectionId}/queryables), as per OGC API - Features - Part 5: Schemas, and as discussed during last week's OGC Standards Code Sprint. The feature is driven by the OGC API - Records SWG discussion, but also applies to OGC API - Features.

Related Issue / discussion

Additional information

See discussion in OGC API - Records / OGC API - Features SWGs:

• GetDomain operation of OGC APIs opengeospatial/ogcapi-records#461
• GetDomain operation for OGC API Features opengeospatial/ogcapi-features#977 (comment)

Note that functionality may be updated in the future.

Dependency policy (RFC2)

• I have ensured that this PR meets RFC2 requirements

Updates to public demo

• I have ensured that breaking changes to the pygeoapi master demo server have been addressed
  • https://github.com/geopython/demo.pygeoapi.io/blob/master/services/pygeoapi_master/local.config.yml

Contributions and licensing

(as per https://github.com/geopython/pygeoapi/blob/master/CONTRIBUTING.md#contributions-and-licensing)

• I'd like to contribute [feature X|bugfix Y|docs|something else] to pygeoapi. I confirm that my contributions to pygeoapi will be compatible with the pygeoapi license guidelines at the time of contribution
• I have already previously agreed to the pygeoapi Contributions and Licensing Guidelines

[pvgenuchten]

do you have a link to that discussion, from the code i understand the domains method returns the unique domain values for each of the properties of records in the records collection.

• Does it make sense for fields like title/abstract?
• How does this functionality compare to the facets extension, which returns the same information for the current search (and including count)

[tomkralidis]

do you have a link to that discussion, from the code i understand the domains method returns the unique domain values for each of the properties of records in the records collection.

There will be an issue in OGC API - Records shortly, and I will link it here.

Note that this PR, given updated discussion, now provides this functionality in .../querables (and not .../domains), and applies to BOTH OGC API - Records and OGC API - Features (though OGC API - Records is a more direct use case IMO, we still get it "for free" given OGC API - Features - Part 5, where this work is targeted).

* Does it make sense for fields like title/abstract?

In theory it shouldn't matter. From a pygeoapi perspective, the default providers with this support are simply operating on the underlying data. A custom provider plugin can exercise more control as required.

* How does this functionality compare to the facets extension, which returns the same information for the current search (and including count)

OGC API - Records - Part 2: Facets is more about aggregations/computes based on search results, vs. describing a collection's queryables/live enums.

[pvgenuchten]

current?

[tomkralidis]

Thanks, fixed.

[doublebyte1]

Thanks for this really cutting edge feature! 👍🏽

• I am not sure where this is documented, but if there is a place maybe it is worth leaving a link to it in the pygeoapi docs.
• This is kind of a low level thing, being very implementation specific. I realised the success of this functionality is also deployment specific; in the case of elasticsearch, it will fail to get the domains if the mappings are not done in a certain way. But I guess this is a best effort from the server.

[doublebyte1]

I noticed that elastic fails to do the aggregation of strings, when fields are not mapped as raw, but I guess this is too specific of the index, not much we can do. Something to keep in mind, though.

This fails:

            "namealt": {
              "type": "text",
              "fields": {
                "keyword": {
                  "type": "keyword",
                  "ignore_above": 256
                }
              }
            },

this works:

            "nameascii": {
              "type": "text",
              "fields": {
                "raw": {
                  "type": "keyword"
                }
              } 

[doublebyte1]

[tomkralidis]

do you have a link to that discussion, from the code i understand the domains method returns the unique domain values for each of the properties of records in the records collection.

There will be an issue in OGC API - Records shortly, and I will link it here.

FYI the associated GitHub issues can now be found per below:

• GetDomain operation of OGC APIs opengeospatial/ogcapi-records#461
• GetDomain operation for OGC API Features opengeospatial/ogcapi-features#977 (comment)