Closed dancivitarese closed 8 months ago
Hi @dancivitarese, for a quick inspection you can retrieve the fields with:
root = zarr.open_group(z)
root.tree() ---> plots the internal structure
root['field/sample=XXXXX/data'][:]
For a more in depth analysis we have written an interface to easily retrieve the data as xarrays. it is still being finalised, but you can find some examples here: https://github.com/iluise/atmorep_analysis/tree/main
if you ran in global_forecast
mode I suggest you to look at forecasting/forecasting_zarr.py
, while if you ran in BERT mode you can look at trainings/
.
The interface in itself is in utils/read_atmorep_data.py
.
Best, Ilaria
Thank you, @iluise for such quick response. The first option already solved my issue, but I will also test the other repo.
Great.
just one thing I forgot to add: if you ran with main
you might get incorrect outcomes if you plot through lat/lons coordinates. please use the branch iluise-fix-dims
for now.
best,
Ilaria
I tried to read the content as:
It creates a storage like
<zarr.storage.ZipStore at ...>
, but I don't know how to read the values, groups, etc.