Although not strictly required, it is highly recommended to start your project in a "virtual environment." A virtual environment is a container (a directory) that holds a specific version of Python and a set of modules (dependencies), and which does not interfere with the operating system's native Python or other projects on the same computer.
By setting up a different virtual environment for each project you work on, various Django projects can run on different versions of Python, and can maintain their own sets of dependencies, without risk of conflict.
Python 3.3+ already includes a standard
venv module, which you can usually call as
pyvenv. In environments where the
pyvenv command is not available, you can access the same functionality by directly invoking the module as
python3 -m venv.
To create the Virtual environment:
$ pyvenv <env-folder> # Or, if pyvenv is not available $ python3 -m venv <env-folder>
If using Python 2, you can first install it as a separate module from pip:
$ pip install virtualenv
And then create the environment using the
virtualenv command instead:
$ virtualenv <env-folder>
The virtual environment is now set up. In order to use it, it must be activated in the terminal you want to use it.
To 'activate' the virtual environment (any Python version)
$ source <env-folder>/bin/activate
This changes your prompt to indicate the virtual environment is active.
From now on, everything installed using
pip will be installed to your virtual env folder, not system-wide.
To leave the virtual environment use
(<env-folder>) $ deactivate
You may also consider using virtualenvwrapper which makes virtualenv creation and activation very handy as well as separating it from your code:
# Create a virtualenv mkvirtualenv my_virtualenv # Activate a virtualenv workon my_virtualenv # Deactivate the current virtualenv deactivate
In environments where you need to handle multiple Python versions you can benefit from virtualenv together with pyenv-virtualenv:
# Create a virtualenv for specific Python version pyenv virtualenv 2.7.10 my-virtual-env-2.7.10 # Create a vritualenv for active python verion pyenv virtualenv venv34 # Activate, deactivate virtualenv pyenv activate <name> pyenv deactivate
When using virtualenvs, it is often useful to set your
DJANGO_SETTINGS_MODULE in the
#!/bin/sh # This hook is sourced after this virtualenv is activated # Set PYTHONPATH to isolate the virtualenv so that only modules installed # in the virtualenv are available export PYTHONPATH="/home/me/path/to/your/project_root:$VIRTUAL_ENV/lib/python3.4" # Set DJANGO_SETTINGS_MODULE if you don't use the default `myproject.settings` # or if you use `django-admin` rather than `manage.py` export DJANGO_SETTINGS_MODULE="myproject.settings.dev"
It is often also helpful to set your project path inside a special
.project file located in your base
<env-folder>. When doing this, everytime you activate your virtual environment, it will change the active directory to the specified path.
Create a new file called
<env-folder>/.project. The contents of the file should ONLY be the path of the project directory.
Now, initiate your virtual environment (either using
source <env-folder>/bin/activate or
workon my_virtualenv) and your terminal will change directories to