Mastering spaCy: An end-to-end practical guide to implementing NLP applications using the Python ecosystem

I came across Mastering spaCy when I was looking for references for my AI project which involves the use of transformers. I have extensive experience in many aspects of data science, however, NLP has always been an area I want to improve on. As I am reading this book, I can’t help but realize it has gradually become my favorite reference whenever I come across an AI project which more or less involves the implementation of NLP. I like this book for that it is not merely a textbook to the spaCy python library, but also an elaboration to the most commonly seen terms in NLP, such as POS tagging, tokenization, lemmatization, etc, with examples. Thus, reading this book allows me to learn about spaCy and linguistics simultaneously.To gain the most out of this book, I suggest that you prepare a vocabulary list and take notes on the terms appearing in this book that are intuitive to you. You should really take your time to taste the first 6 chapters of Mastering spaCy as it walks you through the general process of completing an NLP project, the concepts (NER and information extraction, etc) you might come across, and the functions in spaCy that apply these concepts.Additionally, I want to express my appreciation to the author, Miss Duygu Altinok, that I particularly enjoy the hands-on practice in chapter 8 and 9 as they are relevant to the project I am currently working on. I am able to borrow part of the codes from these two chapters to improve my own transformer. I strongly recommend this book to everyone who struggles with NLP. It influences me, and I believe it will influence you too!

Disclaimer: The publisher reached out to me to review the book and provided me a copy of it.This book is a comprehensive resource for knowing all the capabilities of spacy package. Book covers a lot of breadth, starting from comparing spacy with nltk, tokenization, lemmatization, POS tagging, Named Entity Recognition, word vectors, semantic similarity computation, text classification, using transformers and chatbot designing. Overall I recommend this book is a very helpful starting resource for someone new to NLP.Author could improve the content in some sections e.g., why Spacy is required to use along with huggingface transformers? and comprehensive code to build a chatbot.Final thoughts:Highly recommend for readers who are new to NLP

The structure and content of the book is really solid, the major downfall is poor editing. Lines of code are frequently incorrect or missing, requiring me to do external research to determine what the correct code is. There’s much knowledge to earn from the author, but it will take effort to get all of it.

Mastering spaCy is a great book for those who want to work with spaCy. It is very thorough in its explanations of spaCy's features, including rule-based and machine learning tasks, and it is suitable for a beginner. Experts will likewise this book very useful as a reference in their daily work. It describes all the important features of spaCy, such as part-of-speech tagging, syntactic and semantic parsing, named entity recognition, word vectors, building and updating machine learning models, using deep learning and transformers, and a complete chatbot put together using spaCy. Overall, an excellent book for the NLP practitioner.

The media could not be loaded.  My Review of Mastering SpaCy by Duygu Altinok - comprehensive, hands-on introduction ofspacy for natural language processing on customizing NLP pipelines, using BERT, huggingface & building your own conversational agents. Recommended reading.

It’s a good book, with good explanations. But it’s getting really old to have to google and troubleshoot every few pages because of typos in the text or lines that are just missing. Even if I copy the text *exactly* I often get errors. The GitHub page doesn’t have all the code, so I can’t always double check there. I feel like I spend a lot of time googling to figure out why I got value errors on code that should have worked. Which might be a good learning experience, but it is intensely frustrating and often completely holds up what I’m trying to do for hours.

I have been practicing NLP for over 10 years now and I'm really happy to see a book on using SpaCy for several NLP applications. This is a good book to have if you are doing a lot of tasks but want to keep it for using a single package to get most of your job done.

I purchased the book because it claims to be written for Spacy 3.0 (See preface ix). But it is not.For one example, and there are many, the code on page 109 uses the nlp = spacy.load('en') convention, which is no longer viable. When you adapt the code for 3.0 using nlp = spacy.load('en_core_web_sm'), you receive an "extra_spans_key" error message. There is nothing wrong with my code. The book is outdated though it claims not to be.

This is probably the most comprehensible data science book I've ever read. The author fully explains each concept and each line of code in readable, approachable English. I wouldn't hesitate to buy any books she publishes in the future.

Ben scritto, ricco di spunti di teoria e di pratica.

Habe das Buch mit sehr viel Hoffnung gekauft und mir sehr viel Zeit genommen, auch wirklich jede Zeile Code abzutippen und auszuprobieren. Als Python Anfänger ist es schon eine recht anspruchsvolle Aufgabe, das Thema NLP. Mit ein wenig Geduld und noch mehr Informatik Hintergrundwissen ist der Einstieg aber machbar. Einige meiner Probleme habe ich auf Gitgub an die nette Autorin Duygu Altinok direkt gerichtet: https://github.com/PacktPublishing/Mastering-spaCy. Die Antworten kamen meinst auch wirklich sehr schnell, was mich sehr beeindruckte!Einige Zeilen des abgedruckten Codes haben die falsche Einrückung, was zuerst zu kleinen aber schnell merkbaren Fehlern führt.Der Beschreibungsstil der einzelnen Codezeilen wechselt ab ungefähr der Hälfte des Buches in eine etwas oberflächlichere Variante. Hier fehlen tatsächlich auch einmal die ein oder andere Zeile.Der code auf Github ist 5 Monate alt und tatsächlich unvollständig.Leider leider jedoch hat sich seit der Drucklegung an spaCy jedoch ein Releasewechsel vollzogen, was einiges von dem abgedruckten code inkompatibel/unbrauchbar macht. Die Suche nach der tatsächlichen Lösung ist nur mit der Originaldokumentation https://spacy.io/usage und ggf. neuen youtube Videos erreichbar. Dadurch wird die Lernkurve erheblich steiler, leider.Habe nun 3/4 des Buches durchgearbeitet. Bisher ca. 35h Arbeit, teilweise auch aufgrund eines Python-Grünschnabels.