Skip to main content

Python

Python Package Index - pypi.org

Cheatsheet: https://www.pythonsheets.com

Steps to create a Python project​

  1. Set Python version: pyenv local 3.10.0. Check the latest Python version with pyenv versions.
  2. Init Pypenv: pipenv --python 3.10.2. This creates the Pipfile.
  3. Enable the virtual environment: pipenv shell.
  4. Set the Python interpreter in PyCharm.
    • At the Preferences, go to 'Python Interperter' (you can click 'Interpreter Settings...' at the bottom-right).
    • Open the 'Python Interperter' drop-down and click 'Show All...'.
    • At the dialog that opens (Python Interpreters), click '+'.
    • At the dialog that opens (Add Python Interpreter), select 'Existing environment' and set the path to the location of new Python interpreter given by pipenv --venv + /bin/python.
    • Close the dialogs and at the drop-down select the new interpreter just added and then click 'OK'.
  5. Install Black and configure it in PyCharm: follow instructions.
    • Doing which black should be something like /Users/albert/.local/share/virtualenvs/project-name--zmW4vHg/bin/black. → This is not true now, it prints /Users/albertvilacalvo/.pyenv/shims/black.
  6. Install libraries, eg: pipenv install flask.

-m flag​

From 'man python': -m <module-name> Searches sys.path for the named module and runs the corresponding .py file as a script.

https://stackoverflow.com/questions/50821312/meaning-of-python-m-flag

pyenv​

https://github.com/pyenv/pyenv

Python versions are installed in pyenv root.

Installed with Homebrew (see 'Notes Python.odt').

Upgrade it with brew upgrade pyenv.

The global Python version is set in ~/.pyenv/version. We can use system as a value.

pyenv commands​

https://github.com/pyenv/pyenv/blob/master/COMMANDS.md

List commands: pyenv commands

Set local version (overrides the global version): pyenv local 3.10.0

Display current version: pyenv version

List versions: pyenv versions

List available versions to install: pyenv install --list

Install version: pyenv install 3.9.7

After installing a new Python version we can upgrade pip and pipenv, so that we use their latest versions:

  • pip install --upgrade pip or /Users/albert/.pyenv/versions/3.10.2/bin/python3.10 -m pip install --upgrade pip
  • pip install --user pipenv
  • Important: after upgrading pipenv we need to close the current shell and open a new one (see upgrade pipenv)

Packaging​

info

Best practice: pin package versions!

To avoid surprises and bugs, always use exacte versions (==) for all dependencies, including transitive/secondary dependencies.

Python Packaging User Guide: https://packaging.python.org

Managing Application Dependencies (Python Packaging User Guide): https://packaging.python.org/tutorials/managing-dependencies

Tools​

github.com/pypa/pipenv - Python Development Workflow for Humans (Pipfile.lock)

github.com/python-poetry/poetry - Python dependency management and packaging made easy

github.com/jazzband/pip-tools - A set of tools to keep your pinned Python dependencies fresh (requirements.in)

github.com/bndr/pipreqs - Generate pip requirements.txt file based on imports of any project

github.com/naiquevin/pipdeptree - A command line utility to display dependency tree of the installed Python packages

github.com/pypa/pipx - Install and Run Python Applications in Isolated Environments

github.com/astral-sh/ruff - An extremely fast Python linter, written in Rust

Search pacakges​

pypi.org

awesome-python.com

python.libhunt.com

pip​

Python package installer - https://github.com/pypa/pip

Docs: https://pip.pypa.io/en/stable

https://github.com/ohmyzsh/ohmyzsh/tree/master/plugins/pip

pip installs packages systemwide (globally) by default if we are not on a venv

Commands​

Docs: https://pip.pypa.io/en/stable/cli/

pip3 --version

List commands: pip3 --help

Help of a command: pip install --help

Upgrade pip itself:

Install package: pip install requests

Install specific version of a package: pip install requests==2.1.3 or pip install requests>=2,<3

If we want a version compatible with eg 2.1.3 (2.1.4 or 2.1.5) we can do: pip install requests~=2.1.3. You can also do: pip install package==2.1.*

If we want a lower version than X we can do: pip install "flask<2".

Install package from Git repo: pip install git+https://github.com/user/repo.git@branch or pip install git+https://github.com/user/repo.git@commit-hash or pip install git+https://github.com/user/repo.git@tag

List installed packages: pip list

Info about installed package (eg show package version): pip3 show requests

List outdated packages: pip list --outdated

Upgrade package: pip install --upgrade requests or pip install -U requests

Uninstall package: pip uninstall requests. Important: this does not remove the uninstalled package dependencies! Those will remain :/

venv​

Docs: https://docs.python.org/3/library/venv.html

Integrated into the Python standard library, available since Python 3.3.

Provides isolated Python environments, with project-specific dependencies and Python interpreter. pip installs dependencies systemwide (globally), which means that you can only have 1 version of a library. venv allows 2 projects to use different versions of the same library, and also different Python versions.

Create virtual environment: python3 -m venv <DIR>. Eg python3 -m venv venv or python3 -m venv ~/.virtualenvs/djangodev.

(This creates lots of files and folders - see them with tree venv.)

Activate the virtual environment: source venv/bin/activate or source ~/.virtualenvs/djangodev/bin/activate. This needs to be done for each new terminal session. Doing . venv/bin/activate alwo works.

(After activating, doing which pip points to the binary in the venv folder, not the global /usr/local/bin/pip3. Same with the Python interpreter: which python points to the venv interpreter, not /usr/bin/python.)

Leave active venv and go back to the global environment: deactivate

To destroy the venv first run deactivate (if active) and then delete the folder (eg rm -rf ./venv).

virtualenv​

Similar than venv but with more features.

Docs: https://virtualenv.pypa.io/en/latest

https://packaging.python.org/key_projects/#virtualenv

Install it using pipx: pipx install virtualenv.

Create virtual environment: python3 -m virtualenv <DIR>.

Activate the virtual environment: source <DIR>/bin/activate.

requirements.txt​

Docs: https://pip.pypa.io/en/latest/user_guide/#requirements-files

It includes transitive (secondary) dependencies. Does not specify the Python version.

Print packages: pip freeze

Create file: pip freeze > requirements.txt

Install dependencies: pip install -r requirements.txt or pip install -r requirements.txt -t ./venv

Every time we do pip install somepackage we need to do pip freeze > requirements.txt.

Example​

# This is a comment
requests>=2.20
flake8>=3.8,<=4.0
schedule==0.4.2

See the example in the docs: https://pip.pypa.io/en/stable/cli/pip_install/#example-requirements-file

dev dependencies​

requirements.txt

Flask==1.0

requirements-dev.txt

-r requirements.txt
mock==2.0.0

Example: https://github.com/Yelp/love

pip-tools​

https://github.com/jazzband/pip-tools

https://stackoverflow.com/questions/61536466/pips-requirements-txt-best-practice

Pipenv​

github.com/pypa/pipenv

Docs: https://pipenv.pypa.io/en/latest

Install:

Upgrade Pipenv​

pip3 install --user --upgrade pipenv

Important note after upgrading. If we are using a Python version from pyenv (ie we have a .python-version file created with eg pyenv local 3.10.2), after upgrading pipenv it's possible that doing pipenv --version still shows the old/previous pipenv version. We need to open a new shell (and close the current one) so that it picks up the new pipenv (just doing source ~/.zshrc is not enough). Note that we have an 'eval' for pyenv at ~/.zprofile and another at ~/.zshrc, which explains this behavior (see https://github.com/pyenv/pyenv#basic-github-checkout).

Commands​

Create a new project: pipenv --three or pipenv --python 3.7 or pipenv --python 3 or pipenv --python 3.10.2. Note that even though we may do pipenv --python 3.10.2, and it does indeed use 3.10.2 (it says 'Using /Users/albertvilacalvo/.pyenv/versions/3.10.2/bin/python3 (3.10.2) to create virtualenv...'), it writes python_version = "3.10" at Pipfile, not 3.10.2.

☢️ Activate project's virtualenv: pipenv shell. Type 'exit' or 'Ctrl+D' to return. Virtual environments are located in ~/.local/share/virtualenvs.

(If we do which python or which pip3 outside the environment we get /usr/bin/python and /usr/local/bin/pip3. Inside we get /Users/albertvilacalvo/.local/share/virtualenvs/.../bin/....)

Install the packages in [packages] but not in [dev-packages]: pipenv install → Can modify Pipfile.lock

Install all packages, including dev dependencies: pipenv install --dev

Install all packages specified in Pipfile.lock: pipenv sync → Never modifies Pipfile.lock

Enforce that your Pipfile.lock is up to date: pipenv install --deploy → This will fail a build if the Pipfile.lock is out–of–date, instead of generating a new one (source)

pipenv install vs pipenv sync:

pipenv install will install all dependencies from the Pipfile, and update Pipfile.lock with the versions it used. pipenv sync will install the exact versions specified in Pipfile.lock.

pipenv install --ignore-pipfile is nearly equivalent to pipenv sync, but pipenv sync will never attempt to re-lock your dependencies as it is considered an atomic operation. pipenv install by default does attempt to re-lock unless using the --deploy flag.

pipenv sync will never try to relock your dependencies (you can regard it as blind to Pipfile) while pipenv install --deploy will first check the consistency between Pipfile and Pipfile.lock and abort if they don't match

To avoid having '*' as version number, when we add a new dependency we need to set the dependency version number from Pipfile.lock to Pipfile. Then run pipenv lock which updates the Pipfile.lock _meta sha256.

Install package: pipenv install requests

Install package to dev: pipenv install pytest --dev

Uninstall package: pipenv uninstall requests

Generate a lockfile: pipenv lock

Run: pipenv run python main.py

Locate the virtualenv: pipenv --venv. Is something like /Users/albertvilacalvo/.local/share/virtualenvs/My-App-K-OEJVkt

Locate the Python interpreter: pipenv --py

Show graph of dependencies: pipenv graph

Fix errors​

Get "✘ Locking Failed! ResolutionFailure" when doing pipenv install black --dev. Fixed it with pipenv lock --pre --clear.

From https://stackoverflow.com/questions/51540404/how-do-you-resolve-python-package-dependencies-with-pipenv

Black​

https://github.com/psf/black

Docs: https://black.readthedocs.io/en/stable/

Installation​

  • With Pipenv:
    • Make sure that the environment is active, ie run pipenv shell.
    • pipenv install black --dev.
    • Then on the Pipfile replace version value '*' with the actual version in the Pipfile (eg "==21.12b0").
    • Finally run pipenv lock. If the error '✘ Locking Failed! ResolutionFailure' happens, run pipenv lock --pre --clear.
  • With venv or virtualenv:
    • Make sure that the environment is active, ie run source venv/bin/activate, otherwise black is installed globally.
    • pip install black

Commands​

List commands: black --help

black --version

Check: black --check <dir>, eg black --check .

Format: black .

PyCharm integration​

https://black.readthedocs.io/en/stable/integrations/editors.html#pycharm-intellij-idea

Requires the File Watchers plugin, which is installed by default in PyCharm.

which black should be something like /Users/albertvilacalvo/.local/share/virtualenvs/todo-app-quaXzdRU/bin/black if we are on an active virtual environment created with Pipenv and we've installed black on the environment (pipenv install black --dev). → This is not true now, it prints /Users/albertvilacalvo/.pyenv/shims/black.

Add the External Tool​

This needs to be done only once ever.

Preferences → Tools → External Tools. Click +.

  • Name: Black
  • Description: Code formatter
  • Program: $PyInterpreterDirectory$/black
  • Arguments: "$FilePath$"
  • Working directory: $ProjectFileDir$

Create the File Watcher​

This needs to be done for every project.

Preferences → Tools → File Watchers. Click + and select 'custom'.

  • Name: Black
  • File type: Python
  • Scope: Project Files
  • Program: $PyInterpreterDirectory$/black (or the output of which black if we are on a active venv)
  • Arguments: "$FilePath$"
  • Output paths to refresh: $FilePath$
  • Working directory: $ProjectFileDir$
  • Environment variables: leave it emtpy

In Advanced Options:

  • Uncheck "Auto-save edited files to trigger the watcher"
  • Uncheck "Trigger the watcher on external changes"
  • Uncheck "Trigger the watcher regardless of syntax errors"
  • Uncheck "Create output file from stdout"

VSCode integration​

https://code.visualstudio.com/docs/python/editing#_formatting

At the global settings (⌘,):

  • Disable 'Format On Paste', 'Format On Save' and 'Format On Type'
  • Set 'Format On Save Mode' to 'file'
  • Set 'Default Formatter' to 'None'.

At the project (not global) .vscode/settings.json set:

"python.formatting.provider": "black",
"editor.formatOnSave": true,
"editor.formatOnSaveMode": "file",
"editor.formatOnType": true,
"editor.formatOnPaste": true,

pycodestyle​

https://github.com/PyCQA/pycodestyle

Check your Python code against some of the style conventions in PEP 8.

Concurrency​

How to Choose the Right Python Concurrency API - https://news.ycombinator.com/item?id=32408577 - https://superfastpython.com/python-concurrency-choose-api/