Skip to content

README.md

Latest commit

 

History

History
92 lines (81 loc) · 2.51 KB

README.md

File metadata and controls

92 lines (81 loc) · 2.51 KB

PythonDataScience with venv and pip

(The fork for using pipenv is here: https://github.com/DrOzturk/PythonDataScience. Bevare: If you are using PyCharm, pipenv may cause some difficulties.)

Starter data science project using Python with common dependencies included...
Pull requests welcome...

# clone this repo to local
git clone git@github.com:DrOzturk/PythonDataScience.git

Create a Virtual Environment and activate it:

python -m venv myenv

On Windows, run:

myenv\Scripts\activate

On Unix or MacOS, run:

source myenv/bin/activate

More: https://docs.python.org/3/tutorial/venv.html

If you need to exit the virtual environment, command:

deactivate

To connect from Jupyter to this kernel in the virtual env,

After above virtual environment activation

# install ipykernel, which consists of IPython as well
pip install ipykernel
# create a kernel that can be used to run notebook commands inside the virtual environment
python -m ipykernel install --user --name=myenv
# you can now find this myenv as an available kernel in your Jupyter if it is already running
# else you can launch JupyterLab from virtual environment:
jupyter lab

More: https://medium.com/@royce963/setting-up-jupyterlab-and-a-virtual-environment-c79002e0e5f7

Installing the dependencies

pip install -r requirements.txt

Adding new dependencies

If you add a new dependency to the project using pip like:

pip install pandas

You will also need to update the requirements.txt

pip freeze > requirements.txt

before checking into the repository.

Developer Tools

Linting

  • pycodestyle : use to check if code complies with code style guide ex:
pycodestyle example_package/example.py

Setuptools

  • Command to Create the package to distribute in dist folder -.tar.gz (like dist/example_package-0.0.1.tar.gz)
python setup.py sdist
  • For More info type:
python setup.py --help-commands
  • helps in module discovery using find_packages(), so we can refer to all modules without relative import

Unit Test Running

  • Running all unit tests in the command line:
python -m unittest -v example_package/tests/test_example.py
  • Running a specific test in a TestExample class in test_example test module:
python -m unittest example_package.tests.test_example.TestExample.test_greater_than

Clean Compiled

find . -type f -name "*.py[co]" -delete -or -type d -name "pycache" -delete

.gitignore