word-extractor

word-extractor

Read data from a Word document (.doc or .docx) using Node.js

Why use this module?

There are a fair number of npm components which can extract text from Word .doc files, but they all appear to require some external helper program, and involve either spawning a process or communicating with a persistent one. That raises the installation and deployment burden as well as the runtime one.

This module is intended to provide a much faster way of reading the text from a Word file, without leaving the Node.js environment.

This means you do not need to install Word, Office, or anything else, and the module will work on all platforms, without any native binary code requirements.

As of version 1.0, this module supports both traditional, OLE-based, Word files (usually .doc), and modern, Open Office-style, ECMA-376 Word files (usually .docx). It can be used both with files and with file contents in a Node.js Buffer.

How do I install this module?

yarn add word-extractor

# Or using npm... 
npm install word-extractor

How do I use this module?

const WordExtractor = require("word-extractor"); 
const extractor = new WordExtractor();
const extracted = extractor.extract("file.doc");

extracted.then(function(doc) { console.log(doc.getBody()); });

The object returned from the extract() method is a promise that resolves to a document object, which then provides several views onto different parts of the document contents.

Methods

WordExtractor#extract(<filename> | <Buffer>)

Main method to open a Word file and retrieve the data. Returns a promise which resolves to a Document. If a Buffer is passed instead of a filename, then the buffer is used directly, instad of reading a disk from the file system.

Document#getBody()

Retrieves the content text from a Word document. This will handle UNICODE characters correctly, so if there are accented or non-Latin-1 characters present in the document, they'll show as is in the returned string.

Document#getFootnotes()

Retrieves the footnote text from a Word document. This will handle UNICODE characters correctly, so if there are accented or non-Latin-1 characters present in the document, they'll show as is in the returned string.

Document#getEndnotes()

Retrieves the endnote text from a Word document. This will handle UNICODE characters correctly, so if there are accented or non-Latin-1 characters present in the document, they'll show as is in the returned string.

Document#getHeaders()

Retrieves the header and footer text from a Word document. This will handle UNICODE characters correctly, so if there are accented or non-Latin-1 characters present in the document, they'll show as is in the returned string.

Document#getAnnotations()

Retrieves the comment bubble text from a Word document. This will handle UNICODE characters correctly, so if there are accented or non-Latin-1 characters present in the document, they'll show as is in the returned string.

License

Copyright (c) 2016-2021. Stuart Watt.

Licensed under the MIT License.

Changelog

1.0.0 / 16th May 2021

• Major refactoring of the OLE code to use promises internally
• Added support for Open Office XML-based (.docx) Word files. See #1
• Added support for reading direct from a Buffer. See #11
• Removed event-stream dependency. See #19
• Fixed an issue with not closing files properly. See #23
• Corrected handling of extracting files with files. See #31
• Corrected handling of extracting files with deleted text. See #32
• Fixed issues with extracting multiple rows of table data. See #33

This is a major release, and while there are no incompatible API changes, it seemed best to bump the version so as not to pick up updates automatically. However, all old applications should not require any code changes to use this version.