We expect a small number of edge cases where the same ref "name" attribute is used for multiple tags on a page, but with a different "group" for each.
- Improve scraper logic to distinguish refs by group and name.
- Add this edge case to the test fixture in docker-dev
- Pull the HTML using the REST interface for the revision, copying it to the scraper repo with a command like: curl http://dev.wiki.local.wmftest.net:8080/w/rest.php/v1/revision/7/html > test/data/fixture4.html
- Adapt the test expectation.
Code to review