VS Code does not find Python kernel

Had to update the Jupyter extension, even though I had auto-update on.

I had a lot of problems myself in getting the updated list of jupyter kernel servers in old versions of Visual Studio code.

However, in the newer version of Visual Studio code (I am using version 1.76.2 on my ubuntu 18.04 machine; I hope this works also on Mac OS), you don't need to specify the jupyter kernel anymore: it is now enough to specify the python environment. This new feature and its use are described in this howto from visualstudio.

In short, once I installed the correct version of Visual Studio code, I performed the following steps:

• I first make a virtual environment and install ipykernel in it:

virtualenv --python=python3 venv
. venv/bin.activate
pip3 install ipykernel

• In Visual Studio code, while having opened a python file, I select using ctrl+shift+P (Shift+Command+P in Mac OS) the python interpreter (Python:Select Interpreter) and choose the one of the virtual environment that I have just created: /path/to/venv/bin/python. In my case, this step was necessary: indeed, I first had to choose the desired python interpreter on any python file before proceeding to the following step
• In Visual Studio code still, while having opened a notebook file (.pynb file), I click on select kernel on the top right of the IDE, then click on "Python Environments": the environment I selected before for the python file (/path/to/venv/bin) will appear here, without having to add environments to the kernel lists. Once this environment is selected, the notebook should work as expected,

I hope it helps.

For some upgrading to the current jupyter extension may still not work I had to find the correct combination of vscode and jupyter extensions.

I am on older mac OSX 10.13 vscode : 1.78.2 I had been using jupyter extension : 2023.1.2000312134 which would not recognize kernels installed via conda in the vscode workspace using :

    my_vs_workspace me$ conda create --prefix my_conda_env python=3.10
    my_vs_workspace me$ conda activate /my_vs_workspace/my_conda_env
    (my_conda_env) my_vs_workspace me$ python3.10 -m ipykernel install --name "my_conda_env" --user

now reload the vscode workspace and the kernel would still not be available in vscode. Upgrading to the latest jupyter extension v 2023.4.1011241018 would crash the extension and the kernel loading. vscode would recognize the kernel but during the attempt to load an error would be thrown ... bisect the extension and kernel (paraphrase -sorry i didn't write it down).

Loading the jupyter extension : v2023.1.2010391206 does work for me. Reload vscode was required after the terminal commands and the extension changes (just as a reminder)

thanks to those above phi

I just had the same issue when I couldn't get the newly created environment to show up in kernel selection window. It appeared there immediately after I selected my venv/bin/python3 as Python interpreter on the notebook file.

Python must be in the PATH, to at least get the prompt to install the ipykernel package (also required), which any user can do.

(Note that the python installer doesn't update the HKLM Uninstall registry when installing for all users.)

Vscode 1.102.0 was a problem for me, and version 1.105.0 fixed it.

Of course, the Jupyter extension has to be installed.

Had this same issue and in my case it was caused by my VS code operating in restricted mode: https://code.visualstudio.com/docs/editor/workspace-trust#:~:text=You%20will%20also%20see%20a,%2C%20workspace%20settings%2C%20and%20extensions. The issue was solved when disabling it.

If your Python came from the Microsoft Store, VS Code/Jupyter may not find kernels. Install Python from python.org, recreate the venv, and register the kernel. That worked for me.

I did the same in my vscode. Unfortunately, I can get the kernel directly from the list.

I suggest you manually add the virtual environment path in the settings.json file (use shortcuts "Ctrl+shift+P" and type Preference: Open user settings(JSON)):

"python.venvPath": "${fileDirname}/new_venv/Scripts/python",

I had the same problem. The steps below worked on VSCode v1.84.2 (Universal)

1. Press Command+Shift+P and run Notebook: Select Notebook Kernel
2. Click on Select Another Kernel
3. Click on Python environments
4. Then go to Create Python Environment
5. Select venv and then Enter interpreter path
6. Enter the relative or absolute path to your venv.

That should do it!

1. You need to download Jupyter extension in vs code for wsl again (also If you had downloaded it on windows). It will find the path of python kernel autometically.

2. Then Run .ipynb file..

When I moved to this OS, I was trying this as well. I remember that i had installed Python from the official site which is python.org, but I did not face this issue.

I suggest you update the python version on your system, ensure you installed the suitable version, and retry connecting to the virtual env via VS code. I suggest to follow the tip that was shared earlier in the following link, to reinstall, or update the Jupyter extension, and then relaunch your VS code software.

Also some quick checklist:

1. ensure ipykernel is installed in your system (if not, install it)

2. you can try registering the env by yourself on the kernel.

3. ensure kernel is selected correctly inside your VS code app.

4. if all above 3 points are satisfied, close the app and restart VS code.

Would love to hear if there's any more effective way to combat this. In addition to above, you can read the attached official documentation of VS code, with link to py environments section.