$ python -m venv .venv
$ source .venv/bin/activate
$ (.venv) pip install -r requirements.txt
...
(.venv) $ pytest
============================= test session starts ==============================
platform darwin -- Python 3.11.6, pytest-8.2.2, pluggy-1.5.0
rootdir: /Users/andrea/python-starter/python-kata-starter
collected 2 items
tests/test_message.py .. [100%]
============================== 2 passed in 0.01s ===============================
(.venv) $We don't want that libraries used for this project will pollute the global Python installation.
Usually this is solved creating a so-called "virtual environment" which act as a local installation of Python.
The simplest way to create a new virtual environment (in Python 3.x) is launching the following command:
$ python -v venv .venv After that command a .venv directory will be created.
Other alternatives are virtualenv which works with Python 2.x, and poetry that does also some other things.
The .venv directory should be not tracked by the version control system because their content may work only on the current machine.
$ cat .gitignore
/.venv/The environment can be used only if it is activated.
$ source .venv/bin/activate
(.venv) $After sourcing this script the shell will load some environment variables that permit to use the Python executable in this environment. For the PATH variable is updated in order to find the python executable in the .venv/bin directory.
With the virtual environment activated we can install some libraries in it,
just using the pip command.
For example if we want need a yaml library we can install
(.venv) $ pip install PyYAMLUnfortunately this is operation will be not saved automatically in any configuration file tracked by the version control.
But pip provides a method for listing all the libraries currently installed comprising the exact version used. The output of this command if saved can be used to reproduce a similar environment in another machine.
To see the output of the command we need just to call it:
(.venv) $ pip freeze
PyYAML==6.0.1It's common to save the output of the command in a file called requirements.txt in order to be used later for reinstalling all the dependencies.
(.venv) $ pip freeze > requirements.txtUsually this file is tracked using version control.
(.venv) $ git add requirements.txt
(.venv) $ git commit -m "Adding requirements.txt"The requirements.txt file can be later used to restore all the libraries using
the -r option
(.venv) $ pip install -r requirements.txt
Collecting PyYAML==6.0.1 (from -r requirements.txt (line 1))
Using cached PyYAML-6.0.1-cp311-cp311-macosx_11_0_arm64.whl.metadata (2.1 kB)
Using cached PyYAML-6.0.1-cp311-cp311-macosx_11_0_arm64.whl (167 kB)
Installing collected packages: PyYAML
Successfully installed PyYAML-6.0.1Is not mandatory to insert the library version, it can be omitted, for kata I prefer use the last images of the libraries.
(.venv) $ pip uninstall PyYAML
(.venv) $ vim requirements.txt
(.venv) $ cat requirements.txt
PyYAML
(.venv) $ pip install -r requirements.txt
Collecting PyYAML (from -r requirements.txt (line 1))
Using cached PyYAML-6.0.1-cp311-cp311-macosx_11_0_arm64.whl.metadata (2.1 kB)
Using cached PyYAML-6.0.1-cp311-cp311-macosx_11_0_arm64.whl (167 kB)
Installing collected packages: PyYAML
Successfully installed PyYAML-6.0.1In order to exit the virtual environment we need to call the deactivate
command.
(.venv) $ deactivate
$ As you can see the prompt will lose the "(.venv) " prefix.
The environment can be completely deleted if you want.
$ rm -Rfv .venvPython itself already include a testing library called unittest. It's a solid
and portable solution, and you don't necessarily need other libraries. It
already includes a mock library called unittest.mock.
The official documentation is:
- https://docs.python.org/3/library/unittest.html
- https://docs.python.org/3/library/unittest.mock.html
PyTest is a third party library with a lighter-weight syntax for writing tests (as cited in the unittest documentation) which provides some useful test tagging mechanism that can be used to set apart integration tests. I use it in the trash-cli project.
I think is a good choice for professional project if you don't have limitation on which library you can add to the project.
When I start a kata I just add the pytest library in the requirements.txt file:
(.venv) $ echo pytest >> requirements.txt
(.venv) $ cat requirements.txt
PyYAML
pytest
(.venv) $ pip install -r requirements.txtSome IDE like PyCharm recognized the requirements.txt file and will install automatically the libraries listed in there.
There is a little problem with JetBrains IDE (like PyCharm), they don't always understand that always want to use the virtual environment under the .venv subdirectory of the project.
This can be solved from the Python Interpreter in Preferences.
- You need to click Add Interpreter > Add Local Interpreter...
- Then: Virtualenv environment > Existing
- At this point you should choose the path that points to:
<YOUR PROJECT DIR>/.venv/bin/python