Context: Within the response for Wikimedia Enterprise APIs, we include all categories (meaning non-hidden and hidden categories) that are applied to a particular article. There are lots of lovely benefits that come from all categories but for the sake of this feature to replace the need for parsing categories off the wikitext, it has more than is on the DOM - thus too much.
What we need: For the response from APIs, just include categories that are not hidden. But store the hidden ones (if possible, we can use these for credibility signals!)
Changes: We need to gather if a category is hidden or not hidden via our ingestion of categories per article, omit the non-hidden. Also we should increase the limit of categories we have to the max (either Max or 500) so that we don't miss any categories in the cutoff.