Encoding for time series data

INSPIRE-MIF / gp-ogc-api-features

Good Practice document for INSPIRE download services based on OGC API - Features

12 stars 12 forks source link

Encoding for time series data #53

Closed tervo closed 3 years ago

tervo commented 4 years ago

Currently recommendation /rec/geojson/geojson-inspire recommends GeoJSON as encoding. Pure GeoJSON does not, however, work well for time series data. Alternative encoding proposal is defined in OGC OMSF-profile and discussed in https://ies-svn.jrc.ec.europa.eu/projects/2017-2/wiki

The O&M guideline is currently under review in OGC. Proposed OMSF-profile is most probably not the final version but it's still a good baseline for time series data. Can we already propose that as encoding for time series data as pure GeoJSON is not feasible?

cportele commented 4 years ago

I do not think this is necessary:

Following the INSPIRE-GeoJSON-encoding rule is just a recommendation, so there is no conformance issue, if one deviates from the encoding rule. We do that, too. I plan to document the reasons in the deployment issues. Since the JSON variant of OMSF is a GeoJSON profile, an API that uses OMSF-GeoJSON would still conform to the "INSPIRE-OAPIF-GeoJSON" requirements class.
OMSF is a good starting point, but maybe there has not been enough discussion/review/uptake to explicitly recommend it at this point?

alexanderkotsev commented 4 years ago

I agree with @cportele that we probably do not need that at least at this stage. In addition to what is already said, there is work ongoing for proposing a good practice on SensorThings API as an INSPIRE Download service. The default encoding is JSON, but there is discussion on also having a GeoJSON encoding for SensorThings. See this GitHub issue. This approach would provide an approach which is dedicated explicitly to time series.

sgrellet commented 4 years ago

I agree with both @alexanderkotsev and @cportele. Indeed we work hard on having SensorThings API as a good practice for INSPIRE download service. In addition, with regards to GeoJSON as a resultFormat, it's more than a discussion there are already implementations work going on : see https://github.com/opengeospatial/sensorthings/issues/70

alexanderkotsev commented 3 years ago

The proposed line of action of the sub-group is to discuss this issue at the forthcoming MIG and MIG-T meetings and also for the communities dealing with multidimensional data structures to consider the possibility to create a INSPIRE good practice that can handle generation of feature data on demand through e.g. the OGC Environmental Data Retrieval API.