GeoPySpark in the long term?

[soxofaan]

from https://github.com/locationtech-labs/geopyspark:

GeoPySpark is not currently under active development

last official release was 0.4.3 in Dec 2018

how deal with this dead end?

[soxofaan]

In a GeoPySpark issue I recently reported (locationtech-labs/geopyspark#712), a developer mentions https://rasterframes.io/ as alternative:

As noted, we're not maintaining this package any longer, ...
I would, however, point you in the direction of pyrasterframes if you're looking for a replacement that is likely to be maintained for a while. It's basically GPS, but for Spark SQL, so it's more efficient, and generally more featureful (even if it may lack support for focal operations, though some of those ops have been locationtech/rasterframes#563).
RF also targets much more modern versions of Spark, so you don't have to be stuck on a three year old release.

[jdries]

Note that we're constantly decreasing our dependency on GeoPyspark, the most important bits and algorithms depend only on Geotrellis. So in a way, this is already being dealt with.
If we wanted to invest the time, we could probably move over the bits that we still depend on, to be able to remove the dependency completely.

[soxofaan]

ok, good to know that we should aim at minimizing GeoPyspark usage.

The annoying thing however is that even with the tiniest bit of GeoPyspark usage, it still weights a quite messily on the bootstrap phase of the app (e.g. a proper spark context should be available before you can import most openeogeotrellis submodules). But I guess that will be the same with any other geotrellis python wrapper/adapter.