Python Managers Comparison

Note

For Windows user / machine learning / newbie: Recommend to use conda + pip

Manage Python interpreter version Link to heading

conda Link to heading

Installation: https://docs.conda.io/en/latest/miniconda.html Specify the Python version when creating the environment: conda create -n env python=3.8

pyenv Link to heading

Installation:

Mac: https://github.com/pyenv/pyenv#installation
Linux: https://github.com/pyenv/pyenv-installer
Windwos: https://github.com/pyenv-win/pyenv-win

Install a new version: pyenv install -v 3.8.0 Set the global version: pyenv global 3.8.0 Set the local version: pyenv local 3.8.0 List versions: pyenv versions

Manage Python virtual environment Link to heading

Global Link to heading

	virtualenvwrapper	conda
Installation	pip install virtualenv virtualenvwrapper	¹
Env List	lsvirtualenv	conda env list
Env Create	mkvirtualenv `env`	conda create -n `env`
Env Activate	workon `env`	conda activate `env`
Env Deactivate	deactivate	conda deactivate
Env Remove	rmvirtualenv `env`	conda env remove -n `env`
Env Duplicate	cpvirtualenv `orig` `new`	conda create -n `new` --clone `orig`

Virtualenvwrapper setup Link to heading

# In .bashrc/.profile/.zshrc etc.
export WORKON_HOME=<path-to-env-stored>
export VIRTUALENVWRAPPER_PYTHON= $(which python3)
source /usr/local/bin/virtualenvwrapper.sh

Local Link to heading

	venv	virtualenv	conda	pipenv	poetry
Installation	-	pip install virtualenv	¹	²	³
Env Create	python -m venv `env`	virtualenv `env`	conda create -p `env`	pipenv --python `python-version`	poetry init
Env Activate	source `env`/bin/activate	source `env`/bin/activate	conda activate ./`env`	pipenv shell	poerty shell
Env Deactivate	deactivate	deactivate	conda deactivate	++ctrl+d++	++ctrl+d++
Env Remove	Delete `env` folder	Delete `env` folder	conda env remove -p `env`	pipenv --rm	⁴
Env Run	-	-	-	pipenv run python main.py	poetry run python main.py

Manage python packages Link to heading

	pip	conda	pipenv	poetry
Installation	⁵	¹	²	³
Package Install	pip install `pkg`	conda install `pkg`	pipenv install `pkg`	poetry add `pkg`
Package Update	pip install --upgrade `pkg`	conda update `pkg`	pipenv update `pkg`	poetry update `pkg`
Package Remove	pip uninstall `pkg`	conda remove `pkg`	pipenv uninstall `pkg`	poetry remove `pkg`
Env Export	pip freeze > requirements.txt	conda env export > environment.yml	Pipfile	pyproject.toml
Env Deploy	pip install -r requirements.txt	conda env create -f environment.yml	pipenv install	poetry install
Import from pip	-	⁶	pipenv install -r requirements.txt	⁷
Export to pip	-	-	pipenv lock -r > requirements.txt	poetry export -f requirements.txt -o requirements.txt

pip Link to heading

Install packages from PyPI.
Won't remove dependencies when removing explicitly installed packages. (Solution: pip install python3-pip-autoremove, pip3-autoremove <pkg>)
Visualize dependencies in requirements.txt:
1. pip install pipdeptree
2. pipdeptree -f > requirements.txt
Mimic separating development env and production env.
1. Without any tools
2. Using pip-tools

conda Link to heading

Install packages from Anaconda, which has fewer packages than PyPI.
Can use pip/poetry inside conda env. (Set poetry config settings.virtualenvs.create false when using poetry inside conda env)
Support nested env activation: conda activate --stack <env>
For exporting packages which is explicitly installed with conda: conda env export --from-history

pipenv Link to heading

Install packages from PyPI.
Support separating development env and production env.
1. Package Install: pipenv install <pkg> --dev
2. Env Deploy: pipenv install --dev
Support .env file.
Generating Pipfile.lock is time-consuming.

poetry Link to heading

Install packages from PyPI.
Support separating development env and production env.
1. Package Install: poetry add <pkg> --dev
2. Package Remove: poetry remove <pkg> --dev
3. Env Deploy: poetry install --no-dev
It is hard to type poetry fast. 😄

Reference Link to heading

https://docs.conda.io/en/latest/miniconda.html ↩︎ ↩︎ ↩︎
pip install pipenv or curl https://raw.githubusercontent.com/pypa/pipenv/master/get-pipenv.py | python ↩︎ ↩︎
curl -sSL https://raw.githubusercontent.com/sdispater/poetry/master/get-poetry.py | python ↩︎ ↩︎
poetry env list --full-path, poetry env remove <env> ↩︎
Use system package manager (e.g. apt, yum) or curl -sSL https://bootstrap.pypa.io/get-pip.py | python ↩︎
while read requirement; do conda install --yes $requirement || pip install $requirement; done < requirements.txt ↩︎
cat requirements.txt | xargs poetry add ↩︎