Process Graph Testing: Local backend for testing
Closed this issue · 3 comments
| Title | Process Graph Testing: Local backend for testing |
|---|---|
| Date | 2021-11-11 |
| Issue | #88 |
| Category | Usability, Debugging |
| Description | In the python client their are thoughts to create a "local mode": Load data locally and apply a process graph to it. Whereas the local python functions mimic the functions in the backend. This is valuable for debugging and tracing the dimensionality of a datacube in a process graph at a given point. For the R-Client this would mean to implement all openEO functions in R (probably using stars). |
| Dependencies | stars |
| Links | Local Prototyping |
| Priority | Low |
| Impact | High |
In the python client their are thoughts to create a "local mode": Load data locally and apply a process graph to it.
I guess since many things rely on xarray in python (and many process are already written in python at the backend) it will not be a very long stretch.
I don't know how much effort this would be to implement in R?
Load collection - save result -> NetCDF -> Load to stars object
Test openEO processes locally (map open processes to R functions)
Effort for implementing all the functions in R would be too high
Possibility to use python local mode via dedicated API as soon as python has implemented it
General question about usefulness
--> Highly optional that we will implement this in the R4openEO
Out of Scope