separate download container

In the current setup, @geoase provided a container doing download , downscaling and PET calculation at once. @Felix.Greifeneder requested the operational download of additional variables (soil moisture on 4 layers), but without subsequent processing.

This can be met with two different approaches:

• adaption of main.py: keep current setup but filter out variable names handed over to downscaling
• split out the downloading into an individual container

I consider the second approach to be the cleaner and more transparent (considering the pipeline) solution and will start working on it. @geoase, please contact me if you have any doubts or expertise to add.

TODO:

• download container
• downscale + PET container
• update flow chart and text in readme of meta repo