JupyterLab plugin says Kite Engine is missing when it's installed

[shivamsingha]

I've installed Kite desktop app. I use jupyter in a conda environment. Installed the pip package and jupyter extension inside the env. I get this message in Jupyter lab

The jupyterlab-kite extension will not work because you are missing the Kite Engine desktop application. To fix this, please install the Kite Engine desktop application.

Application Log

~$ pip list | grep kite
jupyter-kite (1.1.0)

~$ jupyter labextension list
JupyterLab v2.2.6
Known labextensions:
   app dir: /home/shivam/miniconda3/envs/datasc/share/jupyter/lab
        @kiteco/jupyterlab-kite v1.1.0  enabled  OK

OS: Ubuntu 20.04

[frankfletcher]

Like the error says - you need to also install the desktop app. You can't do that from pip or jupyter.

https://help.kite.com/article/143-how-to-install-the-jupyterlab-plugin

[shivamsingha]

I already have installed it. That was the first thing I did and then I followed the instructions from the link you mentioned.

[frankfletcher]

Sorry I see that in your first message now. :) Was the kite desktop application running at the time?

[shivamsingha]

Yup it was running.

[NyanKiyoshi]

It is possible that you have multiple python environments in your PATH and your shell is using a given one, jupyter another one from the PATH.

1. Check PATH

2. Check the Python executable used by Jupyter is the same that the one you install the dependencies with

   # 2.1. In a notebook:
   import sys
   sys.executable
   
   # 2.2. In your shell:
   which python
   which pip

3. If they don't match, it means Jupyter is using the wrong Python environment. More information here.

   A quick fix could also be instead of doing:

   jupyter lab

   Do:

   /path/to/my/env/bin/jupyter lab

[shivamsingha]

I'm using conda as I said. And conda does it however it has to do. The python in PATH is the one from the conda env being used and Jupyter is using the right one.

[baggiponte]

Hello, I am experiencing the same problem. OS is macOS 10.15, I am using both miniconda and pyenv.

I tried what @NyanKiyoshi suggested but found out that both executables match:

>>> import sys
>>> sys.executable
'/Users/luca/.pyenv/versions/miniconda3-latest/envs/transports/bin/python'

while which python in the shell returns /Users/luca/.pyenv/shims/python. [This is what pyenv is supposed to do and the active environment is indeed transports (the project I am working on).]

I am attaching my environment as Kite documentation suggests here. Can't attach the .yml file, though, so here you are the .txt.

environment.txt

EDIT 17/11

Yesterday it briefly worked on a new project with the same environment I was using before. But today it is blocked at the initialising phase.

[shivamsingha]

I haven't tried using Kite anymore tbh. I've just been using VS Code. It opens Jupyter notebooks just fine, a bit glitchy sometimes with conda. And it uses VS's intellisense or whatever they call it. Does the same job as Kite.

I might give it another try when they figure virtual environments out.

[baggiponte]

I haven't tried using Kite anymore tbh. I've just been using VS Code. It opens Jupyter notebooks just fine, a bit glitchy sometimes with conda. And it uses VS's intellisense or whatever they call it. Does the same job as Kite.

I do not like how notebooks are rendered neither on VSCode nor on Pycharm unfortunately, that's why I went with the vanilla one. Unfortunately, syntax highlighting and autosuggestions are not really customisable. Perhaps it was better on plain notebooks, but I am quite a newby.

[1ut]

I found the solution on reddit for jupyterLab on Anaconda. All I had to do was to upgrade jupyterlab by conda install -c conda-forge jupyterlab and uninstall & install jupyter-kite and labextention "@kiteco/jupyterlab-kite"

https://www.reddit.com/r/datascience/comments/idxbob/kite_on_jupyter_lab/

[andrew-son]

I've had the same issue (when kite is installed and JupyterLab doesn't see it). I've tired redoing the setup (reinstalling jupyter jupyter-kite and redoing the setup in Kite desktop) but that didn't work. However, running jupyter serverextension enable jupyter_kite fixed it. Hope that it helps.