Building Data Science Applications with FastAPI

5 (1 reviews total)
By François Voron
    What do you get with a Packt Subscription?

  • Instant access to this title and 7,500+ eBooks & Videos
  • Constantly updated with 100+ new titles each month
  • Breadth and depth in over 1,000+ technologies
  1. Free Chapter
    Chapter 1: Python Development Environment Setup
About this book

FastAPI is a web framework for building APIs with Python 3.6 and its later versions based on standard Python-type hints. With this book, you’ll be able to create fast and reliable data science API backends using practical examples.

This book starts with the basics of the FastAPI framework and associated modern Python programming language concepts. You'll be taken through all the aspects of the framework, including its powerful dependency injection system and how you can use it to communicate with databases, implement authentication and integrate machine learning models. Later, you’ll cover best practices relating to testing and deployment to run a high-quality and robust application. You’ll also be introduced to the extensive ecosystem of Python data science packages. As you progress, you’ll learn how to build data science applications in Python using FastAPI. The book also demonstrates how to develop fast and efficient machine learning prediction backends and test them to achieve the best performance. Finally, you’ll see how to implement a real-time face detection system using WebSockets and a web browser as a client.

By the end of this FastAPI book, you’ll have not only learned how to implement Python in data science projects but also how to maintain and design them to meet high programming standards with the help of FastAPI.

Publication date:
October 2021
Publisher
Packt
Pages
426
ISBN
9781801079211

 

Chapter 1: Python Development Environment Setup

Before we can get started on our FastAPI journey, we need to configure a clean and efficient Python environment. This chapter will show you the best practices and conventions that Python developers use daily to run their projects.

By the end of this chapter, you'll be able to run Python projects and install third-party dependencies in a contained environment that won't raise conflicts if you happen to work on another project that uses different versions of the Python language or different dependencies.

In this chapter, we're going to cover the following main topics:

  • Installing a Python distribution using pyenv
  • Creating a Python virtual environment
  • Installing Python packages with pip
  • Installing the HTTPie command-line utility
 

Technical requirements

Throughout this book, we'll assume you have access to a Unix-based environment, such as a Linux distribution or macOS.

If they haven't done so already, macOS users should install the Homebrew package (https://brew.sh), which helps a lot in installing command-line tools.

If you are a Windows user, you should enable Windows Subsystem for Linux (https://docs.microsoft.com/windows/wsl/install-win10), WSL, and install a Linux distribution (such as Ubuntu) that will run alongside the Windows environment, which should give you access to all the required tools. There are currently two versions of WSL, WSL and WSL2. Depending on your Windows version, you might not be able to install the newest version. However, we do recommend using WSL2 if your Windows installation supports it.

 

Installing a Python distribution using pyenv

Python is already bundled with most Unix environments. To ensure this is the case, you can run this command in a command line to show the version of the currently installed Python:

$ python3 --version

The output version displayed will vary depending on your system. You may think that this is enough to get started, but it poses an important issue: you can't choose the Python version for your project. Each Python version introduces new features and breaking changes. Thus, it's important to be able to switch to a recent version for new projects to take advantage of the new features but still be able to run older projects that may not be compatible. This is why we need pyenv.

pyenv (https://github.com/pyenv/pyenv) is a tool that helps you manage and switch between multiple Python versions on your system. It allows you to set a default Python version for your whole system but also per project.

Beforehand, you need to install several build dependencies on your system to allow pyenv to compile Python on your system. The official documentation provides clear guidance on this (https://github.com/pyenv/pyenv/wiki#suggested-build-environment), but here are the commands you should run:

  1. Install the build dependencies:
    • macOS users, use this:
      $ brew install openssl readline sqlite3 xz zlib
    • Ubuntu users, use this:
      $ sudo apt-get update; sudo apt-get install --no-install-recommends make build-essential libssl-dev zlib1g-dev libbz2-dev libreadline-dev libsqlite3-dev wget curl llvm libncurses5-dev xz-utils tk-dev libxml2-dev libxmlsec1-dev libffi-dev liblzma-dev

      Package managers

      Brew and APT are what are commonly known as package managers. Their role is to automate the installation and management of software on your system. Thus, you don't have to worry about where to download them and how to install and uninstall them. The commands just tell the package manager to update its internal package index and then install the list of required packages.

  2. Install pyenv:
    $ curl https://pyenv.run | bash

    Tip

    If you are a macOS user, you can also install it with Homebrew: brew install pyenv.

  3. This will download and execute an installation script that will handle everything for you. At the end, it'll prompt you with some instructions to add some lines to your shell scripts so that pyenv is discovered properly by your shell:

a. Open your ~/.profile script in nano, a simple command-line text editor:

$ nano ~/.profile

b. Add the following lines before the block containing ~/.bashrc:

export PYENV_ROOT="$HOME/.pyenv"
export PATH="$PYENV_ROOT/bin:$PATH"
eval "$(pyenv init --path)"

c. Save by using the keyboard shortcut Ctrl + O and confirm by pressing Enter. Then, quit by using the keyboard shortcut Ctrl + X.

d. Open your ~/.bashrc script in nano. If you are using zsh instead of Bash (the default on the latest macOS), the file is named ~/.zshrc:

$ nano ~/.bashrc

e. Add the following line at the end:

eval "$(pyenv init -)"

f. Save by using the keyboard shortcut Ctrl + O and confirm by pressing Enter. Then, quit by using the keyboard shortcut Ctrl + X.

  1. Reload your shell configuration to apply those changes:
    $ source ~/.profile && exec $SHELL
  2. If everything went well, you should now be able to invoke the pyenv tool:
    $ pyenv
    pyenv 1.2.21
    Usage: pyenv <command> [<args>]
  3. We can now install the Python distribution of our choice. Even though FastAPI is compatible with Python 3.6 and later, we'll use Python 3.7 throughout this book, which has more mature handling of the asynchronous paradigm. All the examples in the book were tested with this version but should work flawlessly with newer versions. Let's install Python 3.7:
    $ pyenv install 3.7.10

This may take a few minutes since your system will have to compile Python from the source.

  1. Finally, you can set the default Python version with the following command:
    $ pyenv global 3.7.10

This will tell your system to always use Python 3.7.10 by default, unless specified otherwise in a specific project.

  1. To make sure everything is in order, run the following command to check the Python version that is invoked by default:
    $ python --version
    Python 3.7.10

Congratulations! You can now handle any version of Python on your system and switch it whenever you like!

 

Creating a Python virtual environment

As for many programming languages of today, the power of Python comes from the vast ecosystem of third-party libraries, including FastAPI, of course, that help you build complex and high-quality software very quickly. The Python Package Index (https://pypi.org), PyPi, is the public repository that hosts all those packages. This is the default repository that will be used by the built-in Python package manager, pip.

By default, when you install a third-party package with pip, it will install it for the whole system. This is different from some other languages, such as Node.js' npm, which by default creates a local directory for the current project to install those dependencies. Obviously, this may cause issues when you work on several Python projects with dependencies that have conflicting versions. It also makes it difficult to retrieve only the dependencies necessary to deploy a project properly on a server.

This is why Python developers generally use virtual environments. Basically, a virtual environment is just a directory in your project containing a copy of your Python installation and the dependencies of your project. It's quite similar to the node_modules directory in Node.js. This pattern is so common that the tool to create them is bundled with Python:

  1. Create a directory that will contain your project:
    $ mkdir fastapi-data-science
    $ cd fastapi-data-science

    Tip

    If you are on Windows with WSL, we recommend that you create your working folder on the Windows drive rather than the virtual filesystem of the Linux distribution. It'll allow you to edit your source code files in Windows with your favorite text editor or IDE while running them in Linux.

    To do this, you can actually access your C: drive in the Linux command line through /mnt/c. You can thus access your personal documents using the usual Windows path, for example, cd /mnt/c/Users/YourUsername/Documents.

  2. You can now create a virtual environment:
    $ python -m venv

Basically, this command tells Python to run the venv package of the standard library to create a virtual environment in the venv directory. The name of this directory is a convention, but you can choose another name if you wish.

  1. Once this is done, you have to activate this virtual environment. It'll tell your shell session to use the Python interpreter and the dependencies in the local directory instead of the global ones. Simply run the following command:
    $ source venv/bin/activate

After doing this, you may notice that the prompt adds the name of the virtual environment:

(venv) $

Remember that the activation of this virtual environment is only available for the current session. If you close it or open other command prompts, you'll have to activate it again. This is quite easy to forget, but it will become natural after some practice with Python.

You are now ready to install Python packages safely in your project!

 

Installing Python packages with pip

As we said earlier, pip is the built-in Python package manager that will help us install third-party libraries. To get started, let's install FastAPI and Uvicorn:

$ pip install fastapi uvicorn[standard]

We'll talk about it in later chapters, but Uvicorn is required to run a FastAPI project.

Tip

You have probably noticed the word standard inside square brackets just after uvicorn. Sometimes, some libraries have sub-dependencies that are not required to make the library work. Usually, they are needed for optional features or specific project requirements. The square brackets are here to indicate that we want to install the standard sub-dependencies of uvicorn.

To make sure the installation worked, we can open a Python interactive shell and try to import the FastAPI package:

$ python
>>> from fastapi import FastAPI

If it passes without any errors, congratulations, FastAPI is installed and ready to use!

 

Installing the HTTPie command-line utility

Before getting into the heart of the topic, there is one last tool that we'll install. FastAPI is, as you probably know, mainly about building REST APIs. To do so, you have several options:

  • FastAPI automatic documentation (we'll talk about this later in the book)
  • Postman, a GUI tool to perform HTTP requests
  • cURL, the well-known and widely used command-line tool to perform network requests

Even if visual tools are nice and easy to use, they sometimes lack some flexibility and may not be as productive as command-line tools. On the other hand, cURL is a very powerful tool with thousands of options but can be complex and verbose for testing simple REST APIs.

This is why we'll introduce HTTPie, a command-line tool aimed at making HTTP requests with an intuitive syntax, JSON support, and syntax highlighting. It's available to install from most package managers:

  • macOS users, use this:
    $ brew install httpie
  • Ubuntu users, use this:
    $ sudo apt-get update; sudo apt-get install httpie

Let's see how to perform simple requests on a dummy API:

  1. First, let's retrieve data:
    $ http GET https://603cca51f4333a0017b68509.mockapi.io/todos
    HTTP/1.1 200 OK
    Content-Length: 195
    Content-Type: application/json
    [
        {
            "id": "1",
            "text": "Island"
        }
    ]

As you can see, you can invoke HTTPie with the http command and simply type the HTTP method and the URL. It outputs both the HTTP headers and the JSON body in a clean and formatted way.

  1. HTTPie also supports sending JSON data in a request body very quickly without having to format the JSON yourself:
    $ http -v POST https://603cca51f4333a0017b68509.mockapi.io/todos text="My new task"
    POST /todos HTTP/1.1
    Accept: application/json, */*;q=0.5
    User-Agent: HTTPie/2.3.0
    {
        "text": "My new task"
    }
    HTTP/1.1 201 Created
    Content-Length: 31
    Content-Type: application/json
    {
        "id": "6",
        "text": "My new task"
    }

By simply typing the property name and its value separated by =, HTTPie will understand that it's part of the request body in JSON. Notice here that we specified the -v option, which tells HTTPie to output the request before the response, which is very useful to check that we properly specified the request.

  1. Finally, let's see how we can specify request headers:
    $ http -v GET https://603cca51f4333a0017b68509.mockapi.io/todos "My-Header: My-Header-Value"
    GET /todos HTTP/1.1
    Accept: */*
    My-Header: My-Header-Value
    User-Agent: HTTPie/2.3.0
    HTTP/1.1 200 OK
    Content-Length: 227
    Content-Type: application/json
    [
        {
            "id": "1",
            "text": "Island"
        }
    ]

That's it! Just type your header name and value separated by a colon to tell HTTPie it's a header.

 

Summary

You now have all the tools and setup required to confidently run the examples of this book and all your future Python projects. Understanding how to work with pyenv and virtual environments is a key skill to ensure everything goes smoothly when you switch to another project or when you have to work on somebody else's code. You also learned how to install third-party Python libraries using pip. Finally, you saw how to use HTTPie, a simple and efficient way to run HTTP queries that will make you more productive while testing your REST APIs.

In the next chapter, we'll highlight some of Python's peculiarities as a programming language and get a grasp of what it means to be Pythonic.

About the Author
  • François Voron

    François Voron is graduated from the University of Saint-Étienne (France) and the University of Alicante (Spain) with a master's degree in Machine Learning and Data Mining. A full-stack web developer and a data scientist, François has a proven track record working in the SaaS industry, with a special focus on Python backends and REST API.

    He is also the creator and maintainer of FastAPI Users, the #1 authentication library for FastAPI, and is one of the top experts in the FastAPI community.

    Browse publications by this author
Latest Reviews (1 reviews total)
These are some good books I was waiting to get my hands on and learn the new stuff, it is a great resource for learning.
Building Data Science Applications with FastAPI
Unlock this book and the full library FREE for 7 days
Start now