I want to learn Data Science and so have used some really popular Python modules likes Pandas, Matplotlib, Numpy, etc. So I clean installed Anaconda and am now using it as my default Python interpreter and also using Conda for installing packages and making virtual environments. I use VS Code as my daily text editor. But I have run into some issues when using the integrated Git terminal in VS Code with the Anaconda Python interpreter.
There are a couple of issues that I am facing. One of the first issues that I see is when I am using CMD to run Python. If I type and enter python
in cmd, the Python interpreter provided by anaconda comes up. But I also get a warning:
Warning: This Python interpreter is in a conda environment, but the environment has not been activated. Libraries may fail to load. To activate this environment please see https://conda.io/activation
I didn't expect to get this output. Anyway, there's another problem in VS code. But first I would like to mention that I have checked "Add to PATH" when installing Anaconda so no issues there. Now, when I open a new Terminal in VS Code, automatically C:/Users/User/Anaconda3/Scripts/activate
is run and then conda activate base
is run. But when conda activate base
is run, automatically, as mentioned, I get a CommandNotFoundError. It states Your shell has not been properly configured to use 'conda activate'.
If using 'conda activate' from a batch script, change your
invocation to 'CALL conda.bat activate'
And then I am told to initialize my shell, so I did conda init bash
but still no luck. And this brings me to talk about .bash_profile
. I think it has to do something with this bash profile. Anyway, this is what is in my bash profile
# >>> conda initialize >>>
# !! Contents within this block are managed by 'conda init' !!
eval "$('/C/Users/User/Anaconda3/Scripts/conda.exe' 'shell.bash' 'hook')"
# <<< conda initialize <<<
Just a summary of the problem:
Unexpected warning in CMD when running Anaconda Python interpreter
Automatically run Anaconda Scripts and conda activate base when opening new Terminal in VS Code
Conda init bash not helping
P.S I have tried using conda activate [env_name]
in CMD and also in Git Bash and they work without any issues. In other words, Anaconda and Conda work perfectly outside of VS Code terminal.
To do this, simply open the Anaconda Prompt. You can find the Anaconda Prompt by searching in the Windows menu. This will open a command line terminal like the Windows command prompt, except this terminal is configured to use all the Anaconda tools. Once the Anaconda Prompt is open, type “conda” and press enter.
From within VS Code, you can create non-global environments, using virtual environments or Anaconda, by opening the Command Palette (Ctrl+Shift+P), start typing the Python: Create Environment command to search, and then select the command. The command presents a list of environment types: Venv or Conda.
Visual Studio Code includes a full featured integrated terminal that conveniently starts at the root of your workspace. It provides integration with the editor to support features like links and error detection. To open the terminal: Use the Ctrl+` keyboard shortcut to toggle the terminal panel.
Ensure you have installed VS Code externally. Open VS Code, click "Extensions" on left-side of the panel. Ensure "Anaconda Extension Pack" is installed, if not, install it first. After finished installing "Anaconda Extension Pack", close VS Code.
I have figured out the answer myself and would like to share it here. First of all at the time of writing the question I was using Git Bash as my Terminal in VS Code (am still using it). So the issue was that when I ran the command conda init bash
in Git Bash or the VS Code Terminal, Conda just basically put the command used for activating Conda environments in the .bash_profile since it is sourced during logging into Bash. But the integrated Terminal in VS Code is a subshell of a Git Bash session. That is why .bash_profile
is NOT sourced in VS Code since .bash_profile
is only sourced during the main Bash session. The .bashrc
file is the file that is sourced when creating a Terminal session in VS Code. So what you actually need to do is take the code that is put into .bash_profile
by conda init bash
and paste it into your .bashrc
file and make the .bash_profile
source that .bashrc
file automatically.
So, to sum up, using conda init bash
will put the conda command in the .bash_profile
and it is usually sourced by Git Bash, but VS Code Git Bash terminal will use .bashrc
.
So you can just cut and paste the code from .bash_profile
to .bashrc
(as already mentioned) or if you want, just simply follow this: put this code in your .bash_profile
:
if [ -f ~/.bashrc ]; then
source ~/.bashrc
fi
And in your .bashrc
, put this code:
# >>> conda initialize >>>
# !! Contents within this block are managed by 'conda init' !!
eval "$('{path_to_your_conda.exe}' 'shell.bash' 'hook')"
# <<< conda initialize <<<
# You can get conda.exe path by using `which conda` in Git Bash
Now, when you open up a Git Bash session in VS Code Terminal, you can use conda activate env_name
to activate any environments you have.
Everything is now supposed to work as expected in VS Code terminal but I would like to further elaborate about something. If you want, you can skip the conda init bash
process (NOT recommended, just read on for additional knowledge but please follow the above steps only). This is a feature that was introduced in conda 4.4.0
. Till then the way of activating conda environments was by using source activate
but that command was NOT cross-platform, meaning that the command could not be used in OSes like Windows.
So they made this change by introducing commands like: conda activate env_name
so that conda environments become much easier to use despite the OS platform.
conda activate
also has other advantages. This is directly from their release docs:
conda activate: The logic and mechanisms underlying environment activation have been reworked. With conda 4.4, conda activate and conda deactivate are now the preferred commands for activating and deactivating environments. You'll find they are much more snappy than the source activate and source deactivate commands from previous conda versions. The conda activate command also has advantages of (1) being universal across all OSes, shells, and platforms, and (2) not having path collisions with scripts from other packages like Python virtualenv's activate script.
I used this question as a reference. Check it out to learn more.
Having said that, using source activate env_name
will still work if you are using Git Bash, even in VS Code Git Bash terminal. source activate env_name
requires no prior set up or config. But it is highly recommended that you only use conda init
to set everything up and then use conda activate env_name
.
[NOTE]: Locating and modifying the said .bashrc
and .bash_profile
on Windows is usually not as simple as it is on Linux. But can be done fairly easily like this:
It goes without saying but, you should have Git Bash installed. Having Git Bash installed should, as far as I know, automatically create .bashrc
or .bash_profile
or maybe both. These files are called "dotfiles" (since they start with a dot) and these are by default hidden on most OSes and definitely on Windows. If they were auto-created by Git Bash on your system, it is most likely that they are placed in your home directory. Home directory on Windows is C:\Users\<you>\
. With that said, follow this:
cd
. Just type this and you will be in your home directoryls -a
and you will see all your files, even hidden ones. Look for .bash_profile
and .bashrc
. Both should be present. If they are, you are ready to follow the above instructions. But if one is not there or if both are missing create them using: touch .bashrc && touch .bash_profile
. You should now see these files when you again type: ls -a
.bashrc
and .bash_profile
, you can follow the above instructions. Also, to access these two files quicker, open them like this with VS Code: code ~/.bashrc
or code ~/.bash_profile
. Now, modify these two files as per the instructions.In the question, I have also talked about VS Code activating Conda environments automatically. There's no issue with VS Code doing that since this is the default behavior. I misinterpreted that as something that's an issue. But if anyone was looking to stop VS Code from automatically doing that, I would recommend trying to set this in the user settings:
"python.terminal.activateEnvironment": false
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With