dtale show in jupyter notebook

Question

I am exploring this new Python package named dtale. It is very convenient for pandas data frames visualization.

https://pypi.org/project/dtale/

It worked once after 2 hours of loading.

Here is a reproduction of the code, where I reduce the dataset. It still takes hours to load on Jupyter Notebook.

import pandas as pd
import dtale
table = pd.DataFrame([[1, 2], [3, 4]], columns = ['a','b'])
d = dtale.show(table)
d

Comments from @cup and @AMC suggested to use it outside Jupyter notebook which I did using python console. It worked fine.

Any ideas on what is taking so long between Jupyter Notebook and Dtale?

Edit: It is not realy taking so long. The process block jupyter notebook while the data is accessible through the local link. http://LT0PAR01056937:40000/dtale/main/1 Jupyter notebook doesn't print the link and I can't kill it, I need to kill the entire kernel.

Thanks

Answer 1

I'm the main developer on D-Tale. So I wonder if the D-Tale is not showing up within your notebook because it is being hosted from HTTPS but the D-TAle process is hosted from HTTP (which will cause a CORS exception)

Also, if you want to kill a D-tale process within your notebook without having the restart the notebook's kernel you can do the following.

dtale.instances()
# using any of the data ids that are printed from the previous command
# it will also print the URL of each piece of data you've loaded into D-Tale
dtale.get_instance([data_id]).kill()

Also, if your notebook isn't being served under HTTPS you can also try forcing the host to 'localhost' by using the following:

dtale.show(df, host='localhost')

Then maybe it can be reached from localhost. You can also try just killing the cell you kicked your D-tale process off from before having to kill the entire kernel. Hope this helps.

Answer 2

If you want to use in Google COlab then this code will help.

import pandas as pd
import dtale
import dtale.app as dtale_app

dtale_app.USE_COLAB = True 
dtale.show(pd.DataFrame([1,2,3]))

Answer 3

In google colab, adding import dtale.app as dtale_app works fine.

import pandas as pd
import dtale
import dtale.app as dtale_app

table = pd.DataFrame([[1, 2], [3, 4]], columns = ['a','b'])
d = dtale.show(table)
d