csmcallister / fed-a11y-scan Goto Github PK

Describe the bug
The build job fails when trying to send the cdk diff to the Pull Request as a comment via the gh api:

{
  "message": "Bad credentials",
  "documentation_url": "https://docs.github.com/rest"
}

As the project owner, I want to a11y scan more than just a site's homepage so that I can have a more accurate measure of a website's accessibility.

Acceptance Criteria

• Given the list of federal dot gov domains online, when we ingest them with our domain gatherer, we should crawl the homepage and figure out all (or most of) the child pages.
• Given the fact that it's possible to sitemap a domain, when we get a list of the domain's child pages, then I expect that we can use pa11y to scan all of them
• Given the ability to a11y scan all of a site's children pages, when we finish scanning them, then I expect the results to be written to a more suitable/performant datastore, such as dynamodb.
• Given the fact that we're writing our data to a database, when we want to get data in a natural pattern, such as all of the results for an agency, organization, domain, or subdomain, then I expect the database's key/index to support these queries.
• Given that all of the above can be implemented, when we do implement it, the old process should continue unabated in order to support the site until changes are made there.

Additional Context
At present, we only a11y scan the homepage of ~9,500 domains/subdomains. Scanning the homepage does not give an adequate picture of a website's accessibility. A homepage could be accessible while every child page has issues. Or vice-versa.