Thank you Tpt, your tool is *awesome*.
I suggest everyone to use it:
using Internet Archive to get the djvu and then Tpt's tool it's probably the†
best "from-scan-file-to-Commons" cycle I know, the most efficient and easy one.
We are rewriting the italian wikisource Help pages accordingly, it's probably the only option I would like to mention :-)
I always suggest this option to libraries and GLAMs, as they can harness Internet Archive facilities to get different file formats (as the djvu)
and also a good OCR.†
Right now, we only miss to steps to close the cycle:
* have an easy way to Upload a book from Google Books to Internet Archive
* have an easy way to upload metadata from Commons to Wikisource. Maybe Wikidata will help use in this.