https://5percado.hu/munkaidokeret-a-gyakorlatban-1-resz/ (Citoid query) fails to detect the author, even though the author name appears as a link with rel="author".
While [[http://microformats.org/wiki/rel-author|rel="author"]] doesn't necessarily guarantee that the link text will be the author name, it seems like a reasonably safe bet.
I agree it seems a safe bet, although zotero (who we're using for all the scraping now) is a little more cautious about using metadata that appears in the body (as opposed to head), unless there's a custom built translator for it, because you can't guarantee the metadata points to the article itself and not something else on the page, like a comment for example.
If this is a popular site in czech it might make sense to build a translator for it specifically.
Relevant comment from zotero: https://github.com/zotero/translators/issues/1092#issuecomment-468308989
If this is a popular site in czech it might make sense to build a translator for it specifically.
In the meantime, you can use Web2Cit. Just defined some configuration for this domain: a test and a template based on the URL you provided (see corresponding translation summary here). It should work for similar webpages from the same domain. Feel free to change the test or template as needed (e.g., change the item type, etc).
You may use this from Wikipedia installing the Web2Cit user script: https://en.wikipedia.org/wiki/User:Diegodlh/Web2Cit/script