In matt-dray/altcheckr: Assess Image Alt Text on a Web Page

knitr::opts_chunk$set(collapse = TRUE, comment = "#>")

library(altcheckr)
library(dplyr)

Install

You can install {altcheckr} from GitHub using the {remotes} package.

install.packages(remotes)
remotes::install_github("matt-dray/altcheckr")
library(altcheckr)

Get image elements

Use the alt_get() function to scrape the attributes of each <img> element on a web page that you name in the url argument,

The function uses {xml2} and {rvest} to scrape a given web page and extract image attributes, with a little bit of {purrr} to get it into a data frame.

get_img <- alt_get("https://www.bbc.co.uk/news")

The function returns a tibble where each row is an image element from that page and columns are the the image source (src), alt text (alt) and link to a file with a longer description (longdesc), if it exists (sometimes used for complex images). The alt column will be created and filled with NA if it isn't present.

Setting the argument all_attributes to TRUE will return all the attributes provided in the <img> element, not just src, alt and longdesc.

Here is a preview of the tibble that is output from alt_get():

print(get_img)

Check alt text

You can then pass the output of alt_get() to alt_check() to perform a series of basic assessments of each image's alt text.

(You can also pass any data frame that contains a src and alt column, where alt contains the text to be assessed by alt_check(). For example, {altcheckr} has a built-in dataset: example_get.)

check_img <- alt_check(get_img)

This will return the same tibble as alt_get(), but new columns have now been appended.

Each new column is the outcome of a check for a possible accessibility issue with the alt text. For example, whether the alt text actually exists and whether it is long.