While typos in alt text go unnoticed to sighted individuals, they create a jarring, unpleasant experience for screen reader users. Given one or more URLs, the spellchecker will find potential typos in images' alternative text.

Resources on Alternative Text for Images (W3C)

This project requires Python3, beautifulsoup4, pyspellchecker, and lxml.

pip install beautifulsoup4
pip install pyspellchecker
pip install lxml

Given a file name and one or more links, alt_text_spellchecker.py creates the file and lists the potential typos their respective alt texts. The file name must have the extension included. The recommended file type is a text file (*.txt).

The two supported output file types for output are HTML files (*.html and *.htm) and text files (*.txt). Other file types will be treated as text files.

When outputting to an HTML file, heading structure is used to organize each page title and URL. Note that the content of the current HTML file output is what would go in the body section of a full HTML file. That is, the current files do not include the head and body structure, nor do they include a DOCTYPE declaration.

The heading levels correspond to the following:

1. Page title and URL. The URL is an active link to the page.
2. Alt text issue (typos or empty/missing). Plain text under this heading contains the full alt text.
3. Image source attribute value. Plain text under this heading contains the full text of the img tag.

When outputting to a text file, tabs are used to indent each typo listed after its page title and URL. This output can be made into an ordered or unordered list in a Word document by:

1. Copy the output from the text file into a Word document.
2. Select all content (ctrl+a).
3. Manually format as a list. It can be unordered (alt h u then select type) or ordered (alt h n then select type).

Word preserves the indentation levels, resulting in a useful formatted list with the following characteristics:

• Top-level list item: Link title and URL.
• Nested list item: Alt text issue.
• Sub-nested list item: src attribute value and img tag of element with empty or missing alt text.

If there is a potential typo you want alt_text_spellchecker.py to ignore, you can create a custom dictionary file with 1 word per line and reference it with the dict flag (see Usage below).

usage: alt_text_spellchecker.py [-h] [--ignore_empty] [--output OUTPUT] [--dict DICT] links [links ...]

Find potential typos in alt text.

positional arguments:
  links            links to check alt text on

optional arguments:
  -h, --help       show this help message and exit
  --ignore_empty   ignore empty alt text; do not record it in output file
  --output OUTPUT  file to store output in; default output.html
  --dict DICT      dictionary file of typos to ignore

Checking the GitHub home and explore pages:

python ./alt_text_spellchecker.py https://github.com/ https://github.com/explore

Checking the Github home and explore pages, saving to a custom output file github.txt, and adding a dictionary file dict.txt (see Content of dict.txt below) to add "github" and "changelog" as correct words:

python ./alt_text_spellchecker.py --output github.txt --dict dict.txt --ignore_empty https://github.com/ https://github.com/explore

github
changelog