Python setup and installation
=============================

There are many ways to use and install Python. Here we list some
recommended ways to get started in increasing order of complexity.

Google Colab (Easiest):
-----------------------

Want to use Python without installing anything? Enter `Google
Colab <https://research.google.com/colaboratory/>`__, a free environment
for running Jupyter notebooks, allowing you to create text and code
cells side by side with a Google account being all you need to start.

Colab is a popular choice for quick and dirty data science projects. It
has many popular libraries pre-installed, and you have the ability to
install any others you need as well. You even have access to Google’s
GPUs for free, which is great for many deep learning pipelines. Finally,
it is incredibly easy to share your notebooks, with the same sharing
tools as other Google Drive products.

Cons
~~~~

There are some downsides to Colab, however. It is not incredibly
straightforward to connect to `version control </git/>`__. Uploading
data sets can also be a drag, as you have to re-upload your files every
time you refresh the page, though you can connect your Google Drive to
Colab to in theory get around this.

Installation
~~~~~~~~~~~~

Again, there is no installation process for Colab. Simply go to
`research.google.com/colaboratory <https://research.google.com/colaboratory/>`__
and connect your Google Account to get started.

Anaconda (Most tools):
----------------------

`Anaconda <https://www.anaconda.com/products/distribution>`__ is an
incredibly popular distribution of Python. It comes pre-loaded with a
plethora of Python libraries, as well as environments in which to run
your scripts and Jupyter notebooks, like JupyterLab and Spyder.

Anaconda also comes ready with the ``conda`` environment managing
system. ``conda`` allows you to separately install and download versions
of libraries and Python itself for individual projects. This can be
incredibly useful if you need to use Python 3.7 to use ``tensorflow``
for a project, which depends on 3.7, but would like to use 3.9 or other
versions for other projects. ``conda`` makes sure your libraries always
use the correct versions and dependencies for that environment.

.. _cons-1:

Cons
~~~~

Anaconda’s main downside is its large installation size, due to the
numerous packages and programs that come with it. Make sure you have at
least 3 GB of disk space free before you install.

With everything that is installed, it can also be a bit confusing which
editor, IDE, or environment to use for the job. Spyder or Jupyter
notebook are safe bets, and you can also use other environments like VS
Code or PyCharm, as well.

.. _installation-1:

Installation
~~~~~~~~~~~~

1. Download the installer for the free individual edition of Anaconda -
   Python 3.9 edition. You will find installers for Windows, Mac, and
   Linux here: https://www.anaconda.com/distribution.
2. Proceed through a standard installation, which will require several
   clicks.
3. Once your installation is completed, open **Anaconda Navigator** from
   your applications. Here you will see lots of useful applications!
   |anaconda navigator|
4. In Anaconda Navigator on the Home tab, you will see a tile that says
   Jupyter Notebook. Click **Launch** (or **Install** and then
   **Launch**) to open Jupyter Notebook in a web browser.
5. Jupyter Notebook opens in a web browser but is running locally on
   your computer. On the main tab **Files** you will see a location in
   your file system. This startup folder is called the root directory of
   the app. Make sure you have a folder you can navigate to where you
   can put any files and datasets you’ll be using.

Miniconda (Most streamlined)
----------------------------

Want access to Python conda without having to download 3 GB of programs
and files?
`Miniconda <https://docs.conda.io/en/latest/miniconda.html>`__ is here
for you. You get all of the environment control features of ``conda``
without any of the fluff you don’t specifically want.

.. _cons-2:

Cons
~~~~

Because using ``conda``, and therefore Miniconda, requires you to use
terminal and the command line. That being said, the commands to run
``conda`` are relatively straightforward (e.g. ``conda activate my-env``
or ``conda install numpy``).

.. _installation-2:

Installation
~~~~~~~~~~~~

Follow the instructions for your operating system from `the conda
docs <https://conda.io/projects/conda/en/latest/user-guide/install/index.html>`__.

You will also need a text editor or IDE. We recommend
`Spyder <https://www.spyder-ide.org/>`__, `VS
Code <https://code.visualstudio.com/>`__, or
`PyCharm <https://www.jetbrains.com/pycharm/>`__.


.. |anaconda navigator| image:: /_static/images/python/setup-install/anaconda_nav.jpeg