Book Image

Python Automation Cookbook - Second Edition

By : Jaime Buelta
Book Image

Python Automation Cookbook - Second Edition

By: Jaime Buelta

Overview of this book

In this updated and extended version of Python Automation Cookbook, each chapter now comprises the newest recipes and is revised to align with Python 3.8 and higher. The book includes three new chapters that focus on using Python for test automation, machine learning projects, and for working with messy data. This edition will enable you to develop a sharp understanding of the fundamentals required to automate business processes through real-world tasks, such as developing your first web scraping application, analyzing information to generate spreadsheet reports with graphs, and communicating with automatically generated emails. Once you grasp the basics, you will acquire the practical knowledge to create stunning graphs and charts using Matplotlib, generate rich graphics with relevant information, automate marketing campaigns, build machine learning projects, and execute debugging techniques. By the end of this book, you will be proficient in identifying monotonous tasks and resolving process inefficiencies to produce superior and reliable systems.
Table of Contents (16 chapters)
14
Other Books You May Enjoy
15
Index

Creating a Telegram bot

Telegram Messenger is an instant messaging app that has good support for creating bots. Bots are small applications that aim to produce automatic conversations. The big promise of bots is that, as machines, they can create any kind of conversation, totally indistinguishable from a conversation with a human being, and pass the Turing Test, but that objective is quite ambitious and not realistic for the most part.

The Turing Test was proposed by Alan Turing in 1951. Two participants, a human and an Artificial Intelligence (AI) machine/software program, communicate via text (like in an instant messaging app) with a human judge that decides which one is human and which one is not. If the judge can only guess correctly 50% of the time, it can't be easily differentiated and therefore the AI passes the test. This was one of the first attempts to measure AI.

But bots can be very useful for a more limited approach, similar to phone systems where...