@davidfor, will we be able to update the metadata for all the side loaded eBooks we already have on our Readers to fix them to work properly with the new series features?

The metadata update options in the driver and plugin will work in exactly the same way as they currently do. They just update two extra columns in the database.

This will be implemented in seriesmeta tonight at the same time the patches are released. @davidfor, would you mind making the UUID deterministic (preferably a UUIDv5 of the series name with a namespace of 9ea9be91-2966-4fbf-b822-de692259bc9b [I generated a random one just now]), to make compatibility with Calibre easier? If you don't like that, anything deterministic is better than a random one.

No wonder why I couldn't find that. I was seeing if I could patch a few things about it I was planning to try with the previous version.

That's interesting.

^ that :D.

That might have something to do with https://github.com/geek1011/kobopatch-patches/issues/33.

I'm not generating a UUID. I'm reading the database to get the UUID for books from Kobo along with their series name. If a sideloaded book has the same series name and author, I use the UUID. Otherwise, I just use the series name. I have had this working in the driver and plugin with no problems since about 24 hours after the Series tab hit the betas.

The issues I have found have been related to bad series data. The series name from Kobo isn't always consistent. I have a couple of examples where one book has a slightly different series name than the others. (Some have "The xxxx" and others "xxxx"). And I have a few books that download with the series id, but no series name. And this was device dependent. I had to sign-out of the account on the device to fix it. It was something cached on Kobo's side. it seemed to only affected books I had for a while. In at least one case, I think when I bought it it wasn't part of a series and was added to a "Sci Fi Masterworks" series later.

Overall, if the series info in the books is good, the series tab works well. My desire now is for a filter along the lines of "don't show series with only 1 book in it".
I thought the old version was there. There might be some A/B testing going on with it.
And for the record, I'm on Easy level 116.
That would probably fit.

Oh, OK, I see. I'll probably do the same, then. What happens with your implementation if a book already has been updated with Calibre, but someone later bought a book from Kobo in the same series, and they try to sideload another book with the same series? Will the plugin update the first sideloaded book to match?

The driver compares the metadata in the database to the library and updates the database as needed. This should mean all the sideloaded books get updated to the id from Kobo. It isn't something I explicitly tested, but, from memory of the code, it will handle this situation.

And for those worried about performance, there is only one database read. The driver collects all necessary metadata from the database in the first pass. The later comparisons are in memory, and, unless there have been a lot of changes, fast.

OK, that should be simple enough to implement.

Also, do you have any idea of changes made regarding handling of "Extra: " dictionaries? We (jackie_w and I) have been encountering some strange behaviour: https://github.com/geek1011/kobopatc...ment-593735567.

A trivial question perhaps, but how does one play the Word Scramble? I select letters, but see no way to have the selections counted. that count stays at "0" out of however many are listed for that particular grid.

Sorry, no idea. I haven't tried sideloading a dictionary for a while. And, I haven't been paying much attention to how it works recently.

OK, quick question about the database changes for compatibility with the new Series tab.

I sideload everything (even the stuff I buy from Kobo; drm removed, of course) so no need for me to generate a UUID or anything like that, but for various reasons, it'd be easier and faster for me just to copy the database file off my Kobo and run the SQL directly on my computer and then copy it back rather than do it through usb (2.35 GB file, 13,000 books (blame gutenberg), etc, plus I want to compact the database after and that definitely is faster if I do it locally).

I haven't installed the new firmware yet so I can't inspect the database to see what changed, but am I correct to assume that it'd be just as simple as copying over all of the data from the 'Series' column to whatever the new column is (looking at geek1011's seriesmeta commit, is it called SeriesID?) and that's it, or is there more to it than that (sorry if the question sounds dumb; I'm weak in sql)?

Yes, that should work for now. Just copy the contents of 'Series' to 'SeriesID' for sideloaded books (and remember to exclude the rows used for TOC entries). You can also use seriesmeta v3.1.0 to do it for you (it's quite fast, even on large libraries).

Curious: Can seriesmeta be run directly on a database file locally, or is it dependent on having the reader connected too because it also needs to scan the books?

It needs to scan the books (but it's quite efficient at doing so - it scans my ClaraHD with 834 books [5GB] in ~19 seconds).

I wonder why Kobo decided they needed new columns for the series stuff?