basic Dask_jupyerlab extension demo¶
see https://www.youtube.com/watch?feature=player_embedded&v=EX_voquHdk0
If not working, see the how-to: https://github.com/dask/dask-labextension
You may have, if not working, to re-run;
jupyter serverextension enable dask_labextension
Step 1: launch a Client¶
- warning; Dask-Jupyterlab doesn't work for Python-3.9.1, even with Distributed-2020.12.0
- check here for progress on the issue: https://github.com/dask/distributed/issues/4168)
In [ ]:
from dask.distributed import Client
import multiprocessing.popen_spawn_win32
client = Client()
client
step2¶
- click on the DASK Extension (the red pictures) on the left left pannel
- copy the Dashoard url (in blue in the cell above this one) in the DASK DASHBOARD URL ot the Dask Extension panel that appears
- click on the refresh circualr arrow on the left of the bottom CLUSTERS line of the Dask Extension panel
- .... this make the buttons of the DASK extension to go from GREY to ORANGE
- .....clicking on each of them open a dedicated pan with related information: for example "GRAPH"
- ..... with jupyterlan you can show it side-to-side with current notebook by moving tabs
In [ ]:
import dask.array as da
x = da.random.random(((20000 , 20000)), chunks="16 Mib")
x
In [ ]:
y = (x + x.T) - x.mean(axis=0)
y.sum().compute()
In [ ]: