You're reading from Building Data Science Applications with FastAPI Develop, manage, and deploy efficient machine learning applications with Python

Product type Paperback

Published in Jul 2023

Publisher Packt

ISBN-13 9781837632749

Length 422 pages

Edition 2nd Edition

Languages

Python

Tools

FastAPI

Concepts

Data Science

Author (1):

François Voron

View More author details

Table of Contents (21) Chapters

Preface

1. Part 1: Introduction to Python and FastAPI

2. Chapter 1: Python Development Environment Setup FREE CHAPTER

3. Chapter 2: Python Programming Specificities

4. Chapter 3: Developing a RESTful API with FastAPI

5. Chapter 4: Managing Pydantic Data Models in FastAPI

6. Chapter 5: Dependency Injection in FastAPI

7. Part 2: Building and Deploying a Complete Web Backend with FastAPI

8. Chapter 6: Databases and Asynchronous ORMs

9. Chapter 7: Managing Authentication and Security in FastAPI

10. Chapter 8: Defining WebSockets for Two-Way Interactive Communication in FastAPI

11. Chapter 9: Testing an API Asynchronously with pytest and HTTPX

12. Chapter 10: Deploying a FastAPI Project

13. Part 3: Building Resilient and Distributed Data Science Systems with FastAPI

14. Chapter 11: Introduction to Data Science in Python

15. Chapter 12: Creating an Efficient Prediction API Endpoint with FastAPI

16. Chapter 13: Implementing a Real-Time Object Detection System Using WebSockets with FastAPI

17. Chapter 14: Creating a Distributed Text-to-Image AI System Using the Stable Diffusion Model

18. Chapter 15: Monitoring the Health and Performance of a Data Science System

19. Index

Why subscribe?

20. Other Books You May Enjoy

Summary

In this chapter, we showed how WebSockets can help us bring a more interactive experience to users. Thanks to the pretrained models provided by the Hugging Face community, we were able to quickly implement an object detection system. Then, we integrated it into a WebSocket endpoint with the help of FastAPI. Finally, by using a modern JavaScript API, we sent video input and displayed algorithm results directly in the browser. All in all, a project like this might sound complex to make at first, but we saw that powerful tools such as FastAPI enable us to get results in a very short time and with very comprehensible source code.

Until now, in our different examples and projects, we assumed the ML model we used was fast enough to be run directly in an API endpoint or a WebSocket task. However, that’s not always the case. In some cases, the algorithm is so complex it takes a couple of minutes to run. If we run this kind of algorithm directly inside an API endpoint, the...