Download and Process Public Domain Works from Project Gutenberg

Download and process public domain works in the Project Gutenberg collection < http://www.gutenberg.org/>. Includes metadata for all Project Gutenberg works, so that they can be searched and retrieved.

gutenbergr: R package to search and download public domain texts from Project Gutenberg

Authors: David Robinson
License: MIT

Download and process public domain works from the Project Gutenberg collection. Includes

A function gutenberg_download() that downloads one or more works from Project Gutenberg by ID: e.g., gutenberg_download(84) downloads the text of Frankenstein.
Metadata for all Project Gutenberg works as R datasets, so that they can be searched and filtered:
- gutenberg_metadata contains information about each work, pairing Gutenberg ID with title, author, language, etc
- gutenberg_authors contains information about each author, such as aliases and birth/death year
- gutenberg_subjects contains pairings of works with Library of Congress subjects and topics

Install the package with:

install.packages("gutenbergr")

Or install the development version using devtools with:

devtools::install_github("ropenscilabs/gutenbergr")

Examples

The gutenberg_works() function retrieves, by default, a table of metadata for all unique English-language Project Gutenberg works that have text associated with them. (The gutenberg_metadata dataset has all Gutenberg works, unfiltered).

Suppose we wanted to download Emily Bronte's "Wuthering Heights." We could find the book's ID by filtering:

library(dplyr)
library(gutenbergr)
 
gutenberg_works() %>%
  filter(title == "Wuthering Heights")
#> # A tibble: 1 × 8
#>   gutenberg_id             title        author gutenberg_author_id language
#>          <int>             <chr>         <chr>               <int>    <chr>
#> 1          768 Wuthering Heights Brontë, Emily                 405       en
#>                                   gutenberg_bookshelf
#>                                                 <chr>
#> 1 Gothic Fiction/Movie Books/Best Books Ever Listings
#> # ... with 2 more variables: rights <chr>, has_text <lgl>
 
# or just:
gutenberg_works(title == "Wuthering Heights")
#> # A tibble: 1 × 8
#>   gutenberg_id             title        author gutenberg_author_id language
#>          <int>             <chr>         <chr>               <int>    <chr>
#> 1          768 Wuthering Heights Brontë, Emily                 405       en
#>                                   gutenberg_bookshelf
#>                                                 <chr>
#> 1 Gothic Fiction/Movie Books/Best Books Ever Listings
#> # ... with 2 more variables: rights <chr>, has_text <lgl>

Since we see that it has gutenberg_id 768, we can download it with the gutenberg_download() function:

wuthering_heights <- gutenberg_download(768)
wuthering_heights
#> # A tibble: 12,085 × 2
#>    gutenberg_id                                                                    text
#>           <int>                                                                   <chr>
#> 1           768                                                       WUTHERING HEIGHTS
#> 2           768                                                                        
#> 3           768                                                                        
#> 4           768                                                               CHAPTER I
#> 5           768                                                                        
#> 6           768                                                                        
#> 7           768   1801.--I have just returned from a visit to my landlord--the solitary
#> 8           768 neighbour that I shall be troubled with.  This is certainly a beautiful
#> 9           768 country!  In all England, I do not believe that I could have fixed on a
#> 10          768    situation so completely removed from the stir of society.  A perfect
#> # ... with 12,075 more rows

gutenberg_download can download multiple books when given multiple IDs. It also takes a meta_fields argument that will add variables from the metadata.

# 1260 is the ID of Jane Eyre
books <- gutenberg_download(c(768, 1260), meta_fields = "title")
books
#> # A tibble: 32,744 × 3
#>    gutenberg_id                                                                    text
#>           <int>                                                                   <chr>
#> 1           768                                                       WUTHERING HEIGHTS
#> 2           768                                                                        
#> 3           768                                                                        
#> 4           768                                                               CHAPTER I
#> 5           768                                                                        
#> 6           768                                                                        
#> 7           768   1801.--I have just returned from a visit to my landlord--the solitary
#> 8           768 neighbour that I shall be troubled with.  This is certainly a beautiful
#> 9           768 country!  In all England, I do not believe that I could have fixed on a
#> 10          768    situation so completely removed from the stir of society.  A perfect
#>                title
#>                <chr>
#> 1  Wuthering Heights
#> 2  Wuthering Heights
#> 3  Wuthering Heights
#> 4  Wuthering Heights
#> 5  Wuthering Heights
#> 6  Wuthering Heights
#> 7  Wuthering Heights
#> 8  Wuthering Heights
#> 9  Wuthering Heights
#> 10 Wuthering Heights
#> # ... with 32,734 more rows
 
books %>%
  count(title)
#> # A tibble: 2 × 2
#>                         title     n
#>                         <chr> <int>
#> 1 Jane Eyre: An Autobiography 20659
#> 2           Wuthering Heights 12085

It can also take the output of gutenberg_works directly. For example, we could get the text of all Aristotle's works, each annotated with both gutenberg_id and title, using:

aristotle_books <- gutenberg_works(author == "Aristotle") %>%
  gutenberg_download(meta_fields = "title")
 
aristotle_books
#> # A tibble: 39,950 × 3
#>    gutenberg_id                                                                   text
#>           <int>                                                                  <chr>
#> 1          1974                                               THE POETICS OF ARISTOTLE
#> 2          1974                                                                       
#> 3          1974                                                           By Aristotle
#> 4          1974                                                                       
#> 5          1974                                         A Translation By S. H. Butcher
#> 6          1974                                                                       
#> 7          1974                                                                       
#> 8          1974        [Transcriber's Annotations and Conventions: the translator left
#> 9          1974 intact some Greek words to illustrate a specific point of the original
#> 10         1974   discourse. In this transcription, in order to retain the accuracy of
#>                       title
#>                       <chr>
#> 1  The Poetics of Aristotle
#> 2  The Poetics of Aristotle
#> 3  The Poetics of Aristotle
#> 4  The Poetics of Aristotle
#> 5  The Poetics of Aristotle
#> 6  The Poetics of Aristotle
#> 7  The Poetics of Aristotle
#> 8  The Poetics of Aristotle
#> 9  The Poetics of Aristotle
#> 10 The Poetics of Aristotle
#> # ... with 39,940 more rows

FAQ

What do I do with the text once I have it?

The Natural Language Processing CRAN View suggests many R packages related to text mining, especially around the tm package.
The tidytext package is useful for tokenization and analysis, especially since gutenbergr downloads books as a data frame already.
You could match the wikipedia column in gutenberg_author to Wikipedia content with the WikipediR package or to pageview statistics with the wikipediatrend package.
If you're considering an analysis based on author name, you may find the humaniformat (for extraction of first names) and gender (prediction of gender from first names) packages useful. (Note that humaniformat has a format_reverse function for reversing "Last, First" names).

How were the metadata R files generated?

See the data-raw directory for the scripts that generate these datasets. As of now, these were generated from the Project Gutenberg catalog on 05 May 2016.

Do you respect the rules regarding robot access to Project Gutenberg?

Yes! The package respects these rules and complies to the best of our ability. Namely:

Project Gutenberg allows wget to harvest Project Gutenberg using this list of links. The gutenbergr package visits that page once to find the recommended mirror for the user's location.
We retrieve the book text directly from that mirror using links in the same format. For example, Frankenstein (book 84) is retrieved from http://www.gutenberg.lib.md.us/8/84/84.zip.
We retrieve the .zip file rather than txt to minimize bandwidth on the mirror.

Still, this package is not the right way to download the entire Project Gutenberg corpus (or all from a particular language). For that, follow their recommendation to use wget or set up a mirror. This package is recommended for downloading a single work, or works for a particular author or topic.

Code of Conduct

This project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.

News

gutenbergr 0.1.4

Added curl to SUGGESTS, since if it's not installed readr::read_lines could fail

gutenbergr 0.1.3

The Project Gutenberg mirror in Maryland Public Libraries (http://www.gutenberg.lib.md.us) has been broken for months. When it is provided from robot/harvest, replaces with http://aleph.gutenberg.org.
Changed test of .zip capability not to run on CRAN
Removed rvest dependency

gutenbergr 0.1.2

Made compatible with change to distinct in dplyr 0.5 (which is about to be submitted to CRAN)
Removed xml2 dependency

gutenbergr 0.1.1

Transferred repo ownership to ropenscilabs
The license was changed from MIT to GPL-2. This is based on the realization that the catalog data is licensed under the GPL, and the package includes a processed version of the catalog data. (See here).
Updated datasets to 5/5/2016 and added a "date_updated" attriute to tell when they were last updated
Added all_languages and only_languages arguments to gutenberg_works, allowing fine-grained control of languages. (For example, "either English or French" or "both English and French")
Changed get_gutenberg_mirror to use xml2 directly, in order to handle AppVeyor
Removed use of data() in gutenberg_works, since it slows down gutenberg_works about 2X
Various documentation, vignette, and README adjustments in response to ROpenSci feedback.
Added AppVeyor for Windows continuous integration
Added code coverage information through codecov.io and covr, along with tests to improve coverage

gutenbergr 0.1

First version of package, including
- gutenberg_download function, for downloading one or more works from Project Gutenberg using Gutenberg IDs
- Datasets of Project Gutenberg metadata: gutenberg_metadata, gutenberg_subjects, gutenberg_authors
- gutenberg_works function to retrieve filtered version of gutenberg_metadata
- Introductory vignette including basic examples of downloading books
- Unit tests for gutenberg_download and gutenberg_works
Added a NEWS.md file to track changes to the package.