I'll take that into account if I manage to get to this. (Anyone else should feel free to overtake this task.)

I used the following on Test Wiki to test it:

jayvdb, I think there is.

For now, sleep time can be specified through a CLI option. If this option is specified, is it better to ignore it or to output a deprecation notice? Or is it better to leave the CLI option as is but change its default value from 10 to 0?

Given that the script uses a library, it can say for sure whether an isbn is correct and add Template:Please_check_ISBN or Template:Listed Invalid ISBN if it's not, so it's going to be more efficient.

I can't reproduce it too. Sorry, I got it wrong.

You would probably need bs4 4.2.1 (depends on python >= 2.7.5-5~) so mf2py works correctly.

I think to implement it in the following way: catch all page which link to a given template, get HTML for each page, look for table with id="template_name" inside of HTML, parse key-values in the table and add them to Wikibase.

Copying the GCI task description:

Right now replicate_wiki.py uses argparse to parse local arguments. Once handle_args is added to parse global arguments, should it parse local arguments too? If use combination of both handle_args for global and argparse for local arguments then -h would look like