jericho : Break Down the Walls of ‘HTML’ Tags into Usable Text

Structured ‘HTML’ content can be useful when you need to parse data tables or other tagged data from within a document. However, it is also useful to obtain “just the text” from a document free from the walls of tags that surround it. Tools are provied that wrap methods in the ‘Jericho HTML Parser’ Java library by Martin Jericho Martin’s library is used in many at-scale projects, icluding the ‘The Internet Archive’.

As a result of using a Java library, this package requires rJava.

The following functions are implemented:

  • html_to_text: Convert HTML to Text
  • render_html_to_text: Render HTML to Text


If you do use devtools, then it should pickup the Remotes: section in DESCRIPTION. Until the package is on CRAN, you might want to also invoke the installation of jerichojars as shown below:

install.packages(c("jerichojars", "jericho"), repos = "")


Let’s use this NASA blog post as an example.


# current verison
## [1] '0.2.0'
URL <- ""
doc <- paste0(readr::read_lines(URL), collapse = "\n")

This is pure text extraction:

This provides a human readable version of the segment content that is modelled on the way Mozilla Thunderbird and other email clients provide an automatic conversion of HTML content to text in their alternative MIME encoding of emails.

You should run each to see and compare the output (GitHub markdown documents aren’t the best viewing medium).

jericho Metrics

Lang # Files (%) LoC (%) Blank lines (%) # Lines (%)
Java 2 0.18 49 0.38 9 0.19 14 0.13
R 6 0.55 40 0.31 10 0.21 62 0.56
Maven 1 0.09 23 0.18 1 0.02 1 0.01
Rmd 1 0.09 9 0.07 24 0.50 33 0.30
make 1 0.09 8 0.06 4 0.08 0 0.00