[GUI Plugin] EpubMerge

[Mandabar]

Originally Posted by JimmXinu

Calibre's epub->epub conversion, in the 'EPUB output' section, has an option for 'Flatten EPUB file structure'.

Ah. Thanks I'll give that a try!

Edit: Nope, didn't seem to work. I'll have to check to compare to be sure, and see if it made other changes to make it still work, but I went from 5 stylesheets to one. And the named styles don't match any of the original style of the sheets that were removed. Different settings for the same name I mean.
Hmm.

Edit2:Maybe it worked, and it combined the stylesheets? It's hard to tell but I can't see anything obviously wrong popping out at me... Close enough probably...

[JimmXinu]

Yeah, Calibre's convert is very aggressive about rewriting everything. I assume it has to use a very generic intermediate form internally because it supports so many formats.

[Jennifer Stevens]

Hi, JimmXinu, I've been using Epubmerge for a while. Lately my omnibus editions (a whole series merged via your plugin) have been getting rejected for upload at Draft2Digital with an error code like this:

ERROR(NCX-001): /1/toc.ncx(-1,-1): NCX identifier ("13915734-ff33-4049-849a-4f06657573ae") does not match OPF identifier ("epubmerge-uid-1747862829").

I get one line of that for every epub merged into the final file. No clue what I'm doing wrong! Maybe it's because I am using an older version of Calibre, because the newer versions blow away all my stylenames. That was unacceptable because I use those names for a reason, thanks, and want them available for final HTML massage before publication.

I prepare all my omnibus editions by adding the title page and cover and other frontmatter to the first epub in the merge, then adding the backmatter to the last epub in the merge. I've learned to give up on TOCs in omnibus merges; they seem to trigger lots of validation hate.

Most important is fixing this NCX identifier:OPF identifier match. Help?

[JimmXinu]

Not having seen the epubs in question, I am speculating.

But what I would guess, is that Draft2Digital is looking at the extra toc.ncx files retained by EpubMerge for UnMerging. Evidence: "/1/toc.ncx"

IMO, this is a mistake in Draft2Digital's processing. Those files should be ignored. OTOH, they are unnecessary and in a clean published book should technically be removed.

It isn't practical for EpubMerge to change the identifier in those files--they need to match the merged sub-book for UnMerge to work.

I suggest that, before doing your final merge, you go into EpubMerge's config and uncheck 'Keep UnMerge Metadata?'. Then EpubMerge won't include those extra toc.ncx (and content.opf files).

Or, if retaining UnMerge is important to you, use Edit book just before uploading to manually remove the extra toc.ncx and content.opf files--ONLY the ones with a number in front: 1/toc.ncx, 1/content.opf, 2/toc.ncx, 2/content.opf, etc.

[Jennifer Stevens]

JimmXinu, THANK YOU! I will totally uncheck UnMerge Metadata. I don't need it.

I also have a bunch of merged omnibus editions that I'd love not to have to remerge (because authors have galley proofs even after merging what they SWEAR are final files) so the clue for removing extra toc.ncx and content.opf files is huge, too.

THANK YOU SO MUCH.

I kind of thought D2D was being a little bit lazy about this, but hey, knowing how to use my own software is important. Thanks for this. I'll pass it along to my other formatters.

[serveth]

Hello there,
Is there any plan to update the plugin for Epub 3 support?
The plugin merges all ebooks into Epub 2 even if they are Epub 3. And ebook platforms started to use Epub check and other tools for new accessibility rules (EU) which reject files with errors.
It would be great to just let version 3 files into a version 3 file without modification.

[JimmXinu]

Originally Posted by serveth

Hello there,
Is there any plan to update the plugin for Epub 3 support?
The plugin merges all ebooks into Epub 2 even if they are Epub 3. And ebook platforms started to use Epub check and other tools for new accessibility rules (EU) which reject files with errors.
It would be great to just let version 3 files into a version 3 file without modification.

I asked a year ago and basically all responses where in favor of epub2 as the default. Click through here see the responses:

Originally Posted by JimmXinu

I've been revisiting the idea of including epub version 3 support natively in EpubMerge.

Currently, EM uses Calibre's Convert to downgrade epub3 to epub2. This works, but Convert is rather intensive with what it changes.

I think I have a handle on how I'd implement epub3 natively in EM. The problem is that I don't really want to deal with epub3 AND epub2 input at the same time, or with epub3 AND epub2 output.

Calibre's Polish Upgrade book internals appears to be much less heavy handed about what it changes when going from epub2 to epub3. The epubs it produces are NOT epub2 back compatible, however. EM output would NOT be epub2 compliant (no NCX TOC). But Convert is still there.

Any opinions on whether reversing EM from epub2 only (albeit with auto-convert) to epub3 only (with auto-polish) is worthwhile?

I will also be looking at epub3 in EpubSplit and FanFicFare (anthology feature, at least).

Calibre's Polish book feature's Upgrade book internals option changes an epub2 to an epub3 mostly painless. But going from an epub3 to an epub2 via Convert rewrites the entire thing.

I am curious tho, what 'ebook platforms' are rejecting epub2s?

[JSWolf]

When EpubMerge merges eBooks with at least one ePub3, does it remove the ePub3 specific code?

[serveth]

Originally Posted by JimmXinu

I am curious tho, what 'ebook platforms' are rejecting epub2s?

Some bookstores in Poland started using w3c/epubcheck and Ace by DAISY last year. They require books to be in version 3 and compatible with the latest accessibility standards.

[JimmXinu]

My tools are really intended to be user grade, not publisher grade. If it works for you, great, but it's not my priority.

IMO, if you're publishing books, you should be using some other tool to do final checks and verification.