I can activate my python virtual environment from it's folder by entering . bin/activate. I'd like to instead type a single word alias, such as shazam, from the home folder (or anywhere else) that activates the environment, changes to my master project folder, and lists my projects.
I tried creating an alias in .bashrc that pointed to an .sh file containing:
cd ~/path-to-virtual-environment
. bin/activate
cd ~/path-to-master-project-folder
ls -a
I was getting a permission denied error, so I ran chmod u+x <script file>. The script now runs, but the VE does not activate and while the project folders are listed, the shell is not in the master project folder. I would appreciate some guidance. Thanks.
Recreate all your environments in ~/.virtualenvs and install virtualenvwrapper. The command to activate an env is workon shazam. Command line completion is supported.
Now about your problem: you've tried to activate an environment in a shell script. That doesn't work because shell scripts run with another shell and env activation change their environments, not the current. I.e., the environment briefly activated but then at the end of the script the new shell exits and the environment deactivated.
There are two ways to overcome this.
Use aliases or shell functions instead of scripts — they are the only way to change the current shell's environment.
Run interactive shell at the end of your script (exec $SHELL). It inherits activated environment and gives you a command prompt. To deactivate simply exit the shell (exit or [Ctrl]+[D].).
Related
If I run the command:
pipenv shell
in my Mac shell, everything works fine and a new virtual environment is created and activated:
.../django_celery_rabbit_flower$ pipenv shell
Launching subshell in virtual environment...
bash: parse_git_branch: command not found
.../django_celery_rabbit_flower$ . /.../django_celery_rabbit_flower-rEt8HW1V/bin/activate
bash: parse_git_branch: command not found
(django_celery_rabbit_flower) .../django_celery_rabbit_flower$
but a bash error is displayed:
bash: parse_git_branch: command not found
I do not understand where it come from. Any idea?
UPDATE
Jen answer trigger a little thought. I have checked my ./bash_profile and I can see the line:
export PS1="\[\033[36m\]\u\[\033[m\]#\[\033[32m\]\h:\[\033[33;1m\]\w\[\033[m\]\$(parse_git_branch)\$ "
This shows the git branch on the bash prompt. So I believe vscode uses this settings. The folder I am working on is not a git folder. Can I write an if statement to avoid the error being displayed when running the python virtual environment?
Instead make sure that the command does actually exists before running it.
PS1="...."'$(if hash parse_git_branch >/dev/null 2>&1; then parse_git_branch; fi)'"..."
I am trying to setup my project environment from a shell script on ubuntu so that I can skip basic setup every time.
I use my virtual environmet cv using the command workon cv.
But I am having trouble doing so using a shell script. I tried the the script
#!/bin/bash
workon cv
But I get the error
workon: command not found
I try to list all the venv I have
pran#pran-HP-65-Notebook-PC:~$ lsvirtualenv
cv
==
virtual-py2
===========
Also, I thought of locating it
(cv) pran#pran-HP-65-Notebook-PC:~/.virtualenvs$ l
cv/ postdeactivate preactivate* prermvirtualenv*
get_env_details* postmkproject* predeactivate virtual-py2/
initialize postmkvirtualenv premkproject*
postactivate postrmvirtualenv* premkvirtualenv*
How can I do it?
You cannot execute workon command outside of Python virtual environment.
If you are interested, check out this article that will walk you through all of the steps required to set up your virutal env.
If it is not what you need then please clarify your requirements for a project environment.
I found the solution:
My venv was located in .virtualenvs. So, I put the command in the bash file start.sh (meant to be executed before working on my project).
#start.sh
source ~/.virtualenvs/cv/bin/activate
And run it using
$ source start.sh
It works perfectly 👍
I try to open terminal in XFCE and activate python3 virtualenv using the following line in a bash script:
xfce4-terminal --working-directory=$HOME/path/to/project --maximize \
-e 'bash -c "source $HOME/path/to/project/venv/bin/activate; bash"'
The strange thing is that the virtualenv gets kind of activated since:
which python
shows the correct path to the virtualenv directory and the project seems to be working fine.
However I don't see the (venv) to the left from the shell prompt. Moreover, when I enter deactivate it complains that no such command can be found.
Is there a proper way to solve this problem?
I created a bash shortcut for this in my /Users/username/.bash_profile (I use mac; on linux use Users/username/.bashrc instead).
function pcd() {
cd /Users/username/Code/"$1"_env/"$1"
source ../env/bin/activate
atom -a .
}
Where project_env is the root folder, which contains the venv, and the project folders (project folder is where code goes)
to execute this simply call pcd project
I don't know much about bash, so I can't really tell you why your code doesn't work.
Also, make sure you open a new terminal window after saving this.
I am not a Linux user so this might be a easy fix but I have tried the following:
first I install it using the command curl -sSL https://dot.net/v1/dotnet-install.sh | bash /dev/stdin for which I get the following result:
dotnet-install: .NET Core SDK version 2.1.403 is already installed.
dotnet-install: Adding to current process PATH:
/home/<!username!>/.dotnet. Note: This change will be visible only when sourcing script.
dotnet-install: Installation finished successfully.
I do . ~/.profile to reload the profile,
but even after this when I run dotnet I get the following error:
Command 'dotnet' not found, but can be installed with:
sudo snap install dotnet-sdk`
I was expecting the script to do everything and make dotnet available.
TLDR: curl | bash can not modify PATH so it will not add dotnet to your PATH. You need to add dotnet to your path manually. Add export PATH="$PATH:/home/<!username!>/.dotnet" to your ~/.profile (or ~/.bashrc or equivalent) and log out and log back in.
Long version:
When you run a command in the shell (for example, bash), the shell tries to find an executible with the name in the all the paths listed in the environment variable PATH. PATH is generally set to something like /bin:/usr/bin. So when you type a command like curl, your shell looks in both /bin and /usr/bin for an executible file named curl.
You can see what your PATH is by doing env | grep PATH or echo $PATH.
The other important piece of information is how environment variables propagate. It's quite simple, actually:
A program (or process) can only modify its own set of environment variables.
Any child processes that a process creates inherit its environment variables.
What this means is that a program that you execute can not modify the environment variables of another random program. The shell actually provides a special command, export to set its own environment variables (and any child processes it later creates will inherit those).
Note the output at the end of step 1.
Note: This change will be visible only when sourcing script.
If you run curl | bash, it runs bash as a child process. That child process can not modify the environment variables of the program that started it (the shell that invoked curl | bash). So it can not modify PATH to add the location of dotnet to it. It even (helpfully) tells you that it can't.
In step 2, you are reloading ~/.profile. But does it contain any commands to add dotnet to PATH? I dont think so. I know the dotnet-install.sh script has not historically added it. You need to add a line like
export PATH="$PATH:/home/<!username!>/.dotnet"
To your ~/.profile (or ~/.bashrc, or equivalent) manually.
Actually, I would write it as follows to make the change more portable to other users:
export PATH="$PATH:$HOME/.dotnet"
Try running this again:
sudo add-apt-repository universe
sudo apt-get install apt-transport-https
sudo apt-get update
sudo apt-get install dotnet-sdk-2.2
I am working with Docker on my windows machine via git bash. Since git bash does not record the current status on closing, I need to set some environment variables related to Docker every time when I start a new terminal. The command I would like to run before start-up is:
eval $(docker-machine env)
Or better yet, have a bash script including other logics. For example if docker machine is not up, start the machine first, etc. Is there a way to automatically run bash command or script before opening a new git bash window?
I would recommend creating a new file under your home folder(~/) namely ~/.bashrc which is read by your terminal when it first starts-up. Add a function say myStartUpFunction() that runs your command as you need.
myStartUpFunction() {
docker-machine env
}
myStartUpFunction
This would enable you to run the docker-machine env every time a new terminal session is opened.