Skip to content

package_management.md

Latest commit

 

History

History
165 lines (110 loc) · 3.88 KB

package_management.md

File metadata and controls

165 lines (110 loc) · 3.88 KB
Raw

This is a guide to manage Python packages.

Preliminaries

You can either use pip or anaconda to manage your python packages. If you want to use pip, then follow the pip guide, otherwise the anaconda steps.

Pip

  1. Install pip

Depending on you OS, install pip. The following snippet is for Ubuntu. pip is a package manager which uses Pypi as repository.

sudo apt-get update && sudo apt-get -y upgrade
sudo apt-get install python-pip python-dev build-essential 
sudo pip install --upgrade pip
  1. Install virtualenv.

Virtualenv is a tool to create isolated development environment. That is, it enables the user to customise the Python packages for each particular project.

sudo pip install virtualenv
  1. Initiate virtual environment.

To initiate your virtual environment, execute the following.

virtualenv venv

After you have initiated the environment, you can activate it with

source venv/bin/activate

You will see the prefix (venv) appended to your command line. This should be done everytime you start working on the project.

  1. Install and update requirements.

After you have initiated the virtual environment, now we can start installing the required package for the project.

Pull the requirements.txt from the master branch, then install the packages.

NOTE (Michael): Make sure you are in the virtual environment, otherwise, it would be a system wide installation.

git checkout master -- requirements.txt
pip install -r requirements.txt

All future dependent packages should be installed via pip.

If you have updated the package requirements, make sure to update the requirements.txt with

pip freeze > requirements.txt

After updating the package and current development, you can make a pull request to update your developments and the package requirements.

  1. Leaving the virtual environment.

After the developments, you can exit the virtual environment simply by typing

deactivate

Anaconda

NOTE (Michael): Given that most of the team members are using pip, the following guide is slightly adjusted to pip flavoured.

  1. Install Anaconda

Follow the Guide to install Anaconda.

  1. Initiate conda.

Create the environment with.

conda create -n the_reading_machine

Then start the environment

source activate the_reading_machine

After the activation you should see the (the_reading_machine) appended to your command line.

  1. Install and update requirements.

After you have initiated the conda environment, now we can start installing the required package for the project.

Pull the requirements.txt from the master branch, then install the packages.

NOTE (Michael): Make sure you are in the virtual environment, otherwise, it would be a system wide installation.

git checkout master -- requirements.txt
conda install --yes --file requirements.txt

NOTE (Michael): Unfortunately, certain packages have inconsistent naming and thus need to be manually installed. For example, the prompt toolkit is named prompt-toolkit in pip but prompt_toolkit in anaconda.

For example, the prompt-toolkit is not available on the anaconda repository and thus need to be installed with.

conda install -c https://conda.anaconda.org/conda-forge prompt_toolkit

All future dependent packages should be installed via conda.

If you have updated the package requirements, make sure to update the requirements.txt with

conda ls -e > requirements.txt

After updating the package and current development, you can make a pull request to update your developments and the package requirements.

  1. Leaving the virtual environment.

After the developments, you can exit the virtual environment simply by typing

source deactivate