As an employee of MediaLibrary Online, I worked in these months to get all the metadata and index the texts. A bit of perspective
* All the texts have been either proofread or validated.
They are not all whole books: we indexed the texts directly in namespace 0 (and not in the Index: namespace), in order to provide the user a better search experience and findability.
This, in our opinion, is very important: we used the MediaWiki API to retrieve all the data, and some HTML scraping for the rest :-)
* We link directly the EPUB generated by the awesome tool from Tpt, but also to the page in Wikisource.
* We automatically generated the EPUB covers for every text which didn't have one.
MediaLibraryOnline is a digital platform that provide Italian libraries with the possibility of lend digital resource, as ebook or audiobooks.
It's not just "a portal on the Internet", but a service used and managed by single libraries for their uses.
It also has an "Open" collection, freely accessible and downloadable for everyone (and not only the patrons of the libraries which have access to MediaLibrary).
I've been hired few months ago to develop such collection, and this is a major milestone for us (and for me :-).
I'm expecially excited by the fact that now Wikisource ebooks "enter" in the collection of libraries, and often in their very catalog.