Feature request: Add an non-acsii-ized option to calibre.

[Orville Q. Song]

Hello.

I'm a Calibre user from China using Chinese and Japanese, and I'm not sure where to submit Feature request so I end up here.

I saw a lots thread talking about store non-acsii filenames in library with Unicode, but the answer is always no. Still I want to try again cause this is actually a big deal for CJK user's experience (especially for those who use calibre-web like me), or we won't ask this feature for years.

The latest version of calibre-web(a web interface for calibre metadata.db) store books with Unicode by default, because unlike "Saving to disk" in Calibre, it can only serve the file as it in the library. I could open the metadata.db managed by Calibre-web, but if I did something like rename the book, or convert the book, everything will be mess-up. I have about 100 epub books in my server, and if I want to convert them to azw3 without a mess, I have to manually convert each book once a time thorough the Calibre-web.

I do understand this was designed to prevent bugs filenames when sent to devices are ascii-ized, but I also believe if someone use languages other than English, it's basically impossible for him to have an ascii only device. And also, give the user choice is always not a bad idea, right?

So would you please add unicode support to library and e-book convert in Calibre as an option (or a plugin?) to help us life get easier?

Thank you.

[kovidgoyal]

There is absolutely no reason for calibre-web to use ascii filenames if it doesnt want to. It can use whatever filename it likes regardless of the filename in the calibre library.

[Orville Q. Song]

Originally Posted by kovidgoyal

There is absolutely no reason for calibre-web to use ascii filenames if it doesnt want to. It can use whatever filename it likes regardless of the filename in the calibre library.

Yes, they are. But if someone use calibre-web with Language other than English like me and who wants to manage the library with calibre like alter hundreds metadata or convert hundreds books from EPUB to AZW3 at once will end up with a messed up library.

So, I'm wishing there could be an option (or a plugin) to disable the ascii-ize of the author and title. Which I remembered there was a solution before version 5, which just need to comment 4 lines of code in `/usr/lib/calibre/db/backend.py`

So, could you please give the user choice on this feature?

Thank you.

[kovidgoyal]

Sorry I dont follow. Regardless of how many books you convert or metadata your edit, the filenames inside calibre will not get messed up in any way. If that happens then it is a bug in calibre-web and I suggest you use calibre's builtin server instead.

[Orville Q. Song]

The filenames get messed up is because Calibre-web use unicode to store books as it is in the library, but Calibre convert unicode filenames to acsii first.

So if I create a blank library and store a CJK book with Calibre-web, the file structure will looks like this:

.
├ metadata.db
└ 水浒传 (1)
├ cover.jpg
└ 水浒传 - 施耐庵.epub

And if I want to convert this book to azw3, it will end up like this.

.
├ metadata.db
└ 水浒传 (1)
├ cover.jpg
├ 水浒传 - 施耐庵.epub
└ shui hu zhuan - shi nai an.azw3

Then, if I alter some metadata within this book, It will create another folder call "shui hu zhuan (1)" and leave the original folder there.

And since Calibre-web can only serve the book in the library "as it", this will create an absolute mess.

[pdurrant]

The file names used inside the calibre library are up to calibre. If Calibre-web does not follow the same rules as calibre, that is a bug in Calibre-web.

[davidfor]

Originally Posted by Orville Q. Song

The filenames get messed up is because Calibre-web use unicode to store books as it is in the library, but Calibre convert unicode filenames to acsii first.

So if I create a blank library and store a CJK book with Calibre-web, the file structure will looks like this:

.
├ metadata.db
└ 水浒传 (1)
├ cover.jpg
└ 水浒传 - 施耐庵.epub

And if I want to convert this book to azw3, it will end up like this.

.
├ metadata.db
└ 水浒传 (1)
├ cover.jpg
├ 水浒传 - 施耐庵.epub
└ shui hu zhuan - shi nai an.azw3

Then, if I alter some metadata within this book, It will create another folder call "shui hu zhuan (1)" and leave the original folder there.

And since Calibre-web can only serve the book in the library "as it", this will create an absolute mess.

Um no. Calibre-web is a web app that displays data from a calibre library. It will have an option to download books to your local storage to read. Or read them in the browser. What calibre-web send when does not have to have any real relationship to where the data is stored on the server. It is reading the metadata.db, so it know what the metadata is and hence the non-ASCII titles. And that is what the calibre-web should be displaying when you use it. Calibre-web also knows what the file name on the server is. When the file is requested, calibre-web can send it as any file name it wants to. Or it could add the appropriate header to the response to suggest the file name, such as unicode version of the title. You need to ask the calibre-web developer to do this.

Of course, calibre-web might already be doing this, but, it is also up to the browser and OS it is running on to support it.

[PeterT]

Unfortunately calibre-web also allows editting of metadata, deletion and addition of books. I for one would be concerned about using it to manage my library.

Sent from my Pixel 4a using Tapatalk

[Orville Q. Song]

I think we are getting into a dead loop here. Calibre-web is just a reason why I personally wish calibre could store and convert books in library without asciied their names, the real matter here is to give the user option, not whose fault on misuse the software.

I compiled myself an old version of Calibre with only 4 lines code in /src/calibre/db/backend.py commented and it works fine for my situation with zero errors, which I believe lots of people faced the same. I'm not a python programmer after all, I found tutorials online to help me with that.

So, sincerely. I really wish this could be a tweak option so we, the user can easily deal with the situation I faced without pain.

If you guys are not interest with this, I'll try to open a pull request myself which, at least I tried.

And finally, Thanks for your amazing works on this amazing software.

[pdurrant]

You should not be interested in the file names in the calibre database.

Without having tested it, I assume that the name remains in the unicode in the display in calibre, and in exported copies of the book.