unify variable input/output for data extarction
The extraction of different variables should be more easier generalize over different datasets. Therefor the cli-syntax as well as the output format should be unified.
For the cli-syntax the following JSON structure should be used:
[
{"name": "var1", "lvl":[...], "interpolation":method},
...
]
where method
is either "p"
(pressure) or "z"
(height) and lvl
contains numerical values of the levels (either hPa or m)
The expected output format (which will be assumed in the data preprocessing/training) is as following:
- the output should be monthly netCDF files with name
<dataset>_<YYYY>-<MM>.nc
- the variables should be indexed by the axis
lat
,lon
,time
- each lvl should be a separate variable, the interpolation method and the lvl information should be appended to the variable name (i.e.
temp_z850
/temp_p850
)
Edited by Simon Grasse