Before uploading the CSV data to create new items, several things need to be cleaned up. But this cleaning should be done at the end because otherwise the edits would get overwritten if the table is regenerated by running the script again.
[x] Check for errors on the catalog numbers. Some were split between two cells and others are just stupid.
[x] The descriptions need to be checked and corrected. In cases where the only artist name came from Commons, the artist name parts sometimes are silly, or have dangling stuff like open parentheses. In other cases, parts are missing, like the descriptions (but that should probably be fixed in an earlier step).
Before uploading the CSV data to create new items, several things need to be cleaned up. But this cleaning should be done at the end because otherwise the edits would get overwritten if the table is regenerated by running the script again.