Setting up environments

Global and Virtual environments

Now that we have installed Python, we mentioned that it comes with certain pre-installed basic libraries. But what happens when you want to use Python for a specific task and need to install additional packages? For this, Python has enabled pip, a recursive acronym for "Pip Installs Packages" or "Pip Installs Python", as its package manager to automate installation, update, and package removal.

However, installing new packages directly into the download of Python can have difficult consequences. This Python we just installed is the Global environment. We mentioned that Python is an open-source interpreted programming language that goes through constant updates. For this reason, newer libraries developed on different Python versions often conflict with libraries without the same updates, and error messages start to pop up.

Dmitriy Zub, Dec 22, 2021. Python Virtual Environments tutorial using Virtualenv and Poetry. Place of publication: SerpApi. Available here.

This image is a beautiful, chaotic example of what happens to your global environment when we do not have order in the different paths and versions of new packages.

To avoid this, a good practice when using Python is to use virtual environments. This environment allows for isolating package dependencies so they do not clash. For example, you may have two projects, one for computer vision and another for Natural Language Processing (NLP). Both of them use similar libraries but in different versions of Python. We can not install both versions system-wide, but we can create isolated environments for each project.

There are different ways to build Python environments:

Python environments (venv):

For this type of environment, the only requirement for your computer is to have a version of Python installed on it. venv is a Python module that supports lightweight virtual environments.

For this type of environment, you need to be familiar with how the terminal works, how you can move from one folder to another, and the Python versions you have installed. If you are unfamiliar with these requirements, please refer to the How to use Anaconda section.

From Python 3.3 onwards, venv should be included in the commands available. To create a virtual environment with this, please open your terminal:

macOS

To enter the terminal, you can search it directly from the Launchpad or application folder. Type Command + Space bar and type terminal for a shortcut. First, we must ensure you are in the folder where you want to save the environment. When you open the terminal, you should see only your user name:

For this example, I am going to access my Documents folder. You can access whatever folder you wish to save your environment on.

python -m venv [name of the environment]

Inside of the brackets, you can change it to whatever name you want. Just make sure that the name of the environment is something easy to remember, or write it down somewhere. The name should also follow the terminal rules: if you are going to name something with more than one word, you need to hyphenate the words with an underscore (_).

• Example: python -m venv example_environment.

The way you activate it is while inside the folder where you created the environment, call source [name of the environment]/bin/activate. The name in front of the dollar sign should change to the name of the environment you are currently in.

• Example:

WINDOWS OS

For the Windows OS you also need to have previously installed Python.

If you have Python already in your Path variables, you can just use the same arguments as the macOS instructions.

• Example: python -m venv example_environment

Please note that the same rules apply to Windows, so make sure to name the environment something unique and easy to remember, and also select the correct folder for your environment.

Conda environments

Another way to create environments in Python is to use the Anaconda distribution. For this, we have two options: we can download Anaconda from the official distribution or a more light version of Anaconda called Miniconda.

If you go for the Anaconda version:

macOS

If you go to the official Anaconda page (https://www.anaconda.com/download#downloads), you will see the Download button for your OS. If you are using macOS, it is also important to know if your hardware settings are an Intel chip or an M1/M2/M3 chip.

You can check this by going to the Apple at the top left corner of your screen and clicking on About this Mac.

This should prompt a Window that will say the chip your computer has:

Go back to the Downloads page on Anaconda and choose the right setting for your computer.

Once the package is done loading, you can choose the predetermined installation and click install. If your terminal was opened, please relaunch it, you will now be able to see the pre fix in your user name as (base). Which means that you are in the base or root environment. We explain this later on.

It goes from:

To look something like this:

For Windows users:

For Windows, there is only one universal installer, so it does not matter the version of Windows that you are using. It is only important to know that it is only available for the 64-bit version. If your computer is a 32-bit, please look into the Anaconda archive (https://docs.anaconda.com/free/anaconda/install/old-os/).

Follow the installation process. Same as with the Python installation, this installer is going to ask you if you want to add this software to your path variables. If you are not sure or do not understand fully what path does, do not activate this function. We can change it later on if you need it.

Installing Miniconda

Graphical installer:

macOS

You must download the Miniconda installer from the official page (https://docs.anaconda.com/free/miniconda/). From here, you can choose the Intel or the M1/2/3 chip version and the bash or pkg versions. We strongly recommend using the bash version since the pkg may skip the "Destination select" process, failing the installation.

bash

• Once you download the document, locate where it is stored (it is most likely in the Downloads folder).
• Go to a new terminal in your computer (Command + Space bar type terminal + Enter)
• Go into the folder where you have the bash file stored (Ej. cd Downloads); you can verify it is there by running ls.
• This is an optional step, but we recommend verifying the download to ensure everything will run correctly. In your terminal run shasum -a 256 filename replacing filename with the downloaded file name. If it does not raise any error, follow the next step.
• Run bash <conda-installer-name>-latest-MacOSX-x86_64.sh. Replace the <conda-installer-name> with the downloaded file's name.
• Once the process is finished, please close and reopen your terminal (Command + q to close/reboot the program completely. You can see if this was successful by clicking Command + tab. It should show you all the applications you have opened. Make sure that the terminal does not show there).
• To verify the installation, open a terminal window and run conda list. If you installed everything correctly, a list of all the installed packages should appear.

Windows

For Windows, you can just download the .exe file and run it. In case you want to install it directly from the command line, you can run the following code:

curl https://repo.anaconda.com/miniconda/Miniconda3-latest-Windows-x86_64.exe -o miniconda.exe

The curl command allows the terminal to exchange data directly from another server, in this case, it is making an HTTP request, just like a web browser. If the request fails, make sure that your internet connection is stable. Once the download is complete, type:

start /wait "" miniconda.exe /S

This will start the installation process. Afterward, you can run del miniconda.exe to delete the executable file.

After installing, close the terminal and open the "Anaconda Prompt (miniconda3)" to use miniconda.