This package defines a few useful functions for keyword searching using the pdftools package developed by rOpenSci.

The package can be installed from CRAN directly:


To install the development version you use devtools:


Basic Usage

There are currently two functions in this package of use to users. The first keyword_search takes a single pdf and searches for keywords from the pdf. The second keyword_directory does the same search over a directory of pdfs.

Example with keyword_directory

The keyword_directory function allows users to search for keywords in multiple PDF files in one function call. The same functionality from the keyword_search function can be invoked, specifically remove_hyphen and surround_lines. Below is an example of searching a single directory.

directory <- system.file('pdf', package = 'pdfsearch')

# do search over two files
directory_result <- keyword_directory(directory, 
       keyword = c('repeated measures', 'measurement error'),
       surround_lines = 1)

head(directory_result, n = 2)

A few other useful arguments are possible when searching for keywords within multiple PDF files in a directory. One is the recursive (default is FALSE), where if set to TRUE will search within subdirectories as well, the default function behavior will not venture into subdirectories. Finally, if the directory has many PDF files, testing the function first on a handful of PDF files may be desired. The number of PDF files can be limited with the argument max_search where a positive integer can be specified indicating the number of PDF files to search. For example, is max_search = 2, only the first two PDF files will be searched within the directory.

Shiny App

The package also has a simple Shiny app that can be called using the following command

Usage in Research

The pdfsearch package may be most useful to those conducting research syntheses or meta-analyses. The package can allow users to search for keywords related to a research question; therefore, instead of searching the entire text of a document, specific portions of the text can be identified to be searched. This could increase the reproducibility and reduce the time needed to collect the data for the research synthesis or meta-analysis.

As an example, the package is currently being used to explore the evolution of statistical software and quantitative methods used in published social science research ( This process involves getting PDF files from published research articles and using pdfsearch to search for specific software and quantitative methods keywords within the research articles. The results of the keyword matches will be explored using research synthesis methods. A pre-print of the paper and slides from the presentation will be posted to the GitHub repo as part of the package later this summer.