View Single Post
Old 12-19-2017, 06:29 PM   #24
geek1011
Wizard
geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.geek1011 ought to be getting tired of karma fortunes by now.
 
Posts: 2,805
Karma: 7423683
Join Date: May 2016
Location: Ontario, Canada
Device: Kobo Mini, Aura Edition 2 v1, Clara HD
Quote:
Originally Posted by davidfor View Post
I'm not sure if I see the point of importing the book, maintaining the metadata in calibre but not writing that metadata back to the book at sometime. Because if I don't then, when I look at the book on my devices, then I won't have the right title, author, description or cover.

And don't say to use the Modify ePub plugin, the Embed metadata function or a conversion (yes, I know you don't) to do this. Then you are back to the situation that geek1011 is complaining about. All of these methods update the metadata in the book in exactly the same way. They will all put the things in the OPF that geek1011 doesn't like.

@geek1011: I'm not a fan of how the extra metadata is written to the OPF, and every now and then I consider looking at how to block it. Adding an option to only write the DC metadata, or calibre core metadata is probably not hard. But, it just hasn't been important. I have almost zero need to look at the OPF and I know that while there is a lot of seeming cruft, it is all valid and usable if the application wants it.

As to the library file structure, it's a database. I treat it that way and it doesn't bug me. I do play in it, but I do understand what is happening and knowing when the files will be updated is easy: when I change something. If I change some metadata, the OPF file will change. And the book will change if I push the change to the book. And if I change some common metadata element (rename a tag, fix an authors name) then all the OPF files for books with that element will change. Calibre updates the database and then updates the files in the background, so this can take a bit of time for a change that affects a lot of books. But if you want to put the library into a version control system, then you really have to just accept all changes. Push them once a day or each time you make bulk changes or something like that.
Yeah. I'm kind of overly pure about ebook formats. I usually remove everything but cover (I scale it, and convert it to JPG), title, author (I fix the sorting to match the epub spec), series, publisher, description (with custom fonts and styles stripped), date, and a few more elements. I also usually rename CSS classes, and strip extraneous styles. I really cannot stand seeing a cluttered epub file. I'd use fb2 if it had some more formatting features, better compatibility, and worked better with nonfiction. It usually takes me about 4-6 minutes to go through each new book.

As for VCS, I like to track changes properly. I usually store my epubs unzipped, with CI/CD setup to zip them and push them to my devices. Also, using a sqlite database in VCS gets large quite quickly.

To update metadata, I prefer to do it manually in the command line. I find I have more control that way, and it is more flexible.

Also, don't get me wrong: Calibre is great for most people, it's just that I kind of have the opposite ideology. I'm still grateful for the devs, as I learn quite a bit from reading the calibre source code (yes, I converted it into a book ). IMO it's the best documentation for some obsure ebook formats.

Last edited by geek1011; 12-19-2017 at 06:36 PM.
geek1011 is offline   Reply With Quote