Adventures in Omnibus Creating

[Fitz Frobozz]

Ahoy, omnibus creators. For a few weeks now I've been working on a "personal consumption only" ePub project involving combining a bunch of books by a single author into a 1-3 volume omnibus (still working out space req's), which when completed will be how I dive into reading the books for the first time. In addition to arranging the books and stories in the order that I've decided on, I'll be reformatting them as well.

I'm hoping to fill in some knowledge gaps, if possible, related to anything that comes up for other folks here. E.g. ideas about specific things like structure, reading flow, to stuff like best practices, pitfalls, ethics, etc., or even more general thoughts/experiences,observations. Fwiw I feel like I probably have a decent plan in place and feel good about the technical and logistical side of things, but I'm open to input on that front, too.

Please share your thoughts!

[DNSB]

Personally, I use the EpubMerge plugin for calibre to merge epubs into a single volume. It does some funky stuff such as putting the book files and stylesheets for each book in a subdirectory so each book has access to it's CSS but that is more or less necessary.

[Fitz Frobozz]

Originally Posted by DNSB

Personally, I use the EpubMerge plugin for calibre to merge epubs into a single volume. It does some funky stuff such as putting the book files and stylesheets for each book in a subdirectory so each book has access to it's CSS but that is more or less necessary.

Oh nice. I think I installed that plugin. Haven't gotten around to looking at it closely yet. I didn't realize that it compartmentalizes individual books into subfolders like that, which strikes me as a good idea if you're dealing with a lot of content.

Are you retaining separate stylesheets for each book just because it's easier that way, or because you prefer to style the books differently? My plan has been to use pretty much the same styling for all the books using my own style guide, which is based on elements from the source book, previous editions of said book, general best practices, and my own preferences.

My current approach:

1. Export original book to DOCX, convert DOCX to GDOC
2. Reformat in GDOC, export GDOC to Markdown
3. Remove stuff like XHTML comments from Markdown and clean up/fix structure & formatting
4. Import Markdown into empty Calibre book
5. Check structure & formatting again, add image files, clean up CSS
6. Merge into master Omnibus document

[JSWolf]

This thread seems to be broken. I read it and it still shows as unread.

Edit: Adding this post fixed the problem.

[j.p.s]

Originally Posted by JSWolf

This thread seems to be broken. I read it and it still shows as unread.

Edit: Adding this post fixed the problem.

It seems to be happening on a lot of threads lately.

I think it is because of the presence of posts waiting for moderator approval.

[Fitz Frobozz]

Hmm, I just posted a reply in this thread and a short time later my post vanished. Was I just moderated? Didn't notice any notifications about it, if so. (And to be clear, my post was definitely live for at least a few minutes, as I was able to edit it and still see it after refreshing the page.)

[AlanHK]

Originally Posted by Fitz Frobozz

1. Export original book to DOCX, convert DOCX to GDOC

I'd never do that if I'm starting from an ePub, or any ebook format really.
Structural things like h1, h2, etc can get lost. Images can get converted and degraded.

I use mostly Sigil, though there are a few Calibre utilities that are useful.
Like epub to epub conversion can clean up a lot of junk.

After importing all the separate books, each with its own stylesheet, I make a new empty stylesheet and add that to all files.

Then put basic styles in it, taking care not to duplicate the names of styles in other files.

Look out for numbered styles, like "calibre123" that can have a different definition in each book.

Then select all the files for a book, convert the original style to the corresponding one in my master style sheet. Start with indented body text, chapter heads, etc.

After doing a few, run "Delete unused stylesheet classes" and see what is left. Continue until all the original stylesheets are empty, then delete them and I now have consistent styles throughout, and can reformat to my taste.

I also like to get the covers for the original books and use them instead of a bare heading.

Consolidate all the copyright pages to one, and remove any duplications (e.g. the same map or glossary or author bio in every book). Compare and use the best one only.
Delete any previews or promo, redundant if the next book is included.

[Fitz Frobozz]

Originally Posted by AlanHK

I'd never do that if I'm starting from an ePub, or any ebook format really.
Structural things like h1, h2, etc can get lost. Images can get converted and degraded.

Most of those considerations aren't an issue for me in this project, but your POV is totally valid, and I'm scaling down the process I described anyway.

Initially I came at this out of a desire for a clean slate. I wanted to start out (in the calibre editor) with a clean ePub that would be free of unwanted source related things: formatting errors, quality issues, and other stuff I just didn't want for whatever reason. I also had hopes that my approach would give me a convenient way to maintain a design that could be imported easily.

It didn't really work out the way I had hoped, though, because now I see that the import process is going to take a lot more work to set up in the way I had envisioned. I'm now bypassing all of that and just doing what you described. (On the plus side, this experience taught me a lot and I now have a pretty strong set of project and design documents.)

Look out for numbered styles, like "calibre123" that can have a different definition in each book.

Yes, dealing with those styles is a pain. I used to do what you suggested (try to salvage some of it for my own stylesheets) but I grew tired of getting lost in all the crazy organization, redundancies, etc. and I no longer even bother to look at any of it these days. Instead, I just spend time up front (right after starting a new ePub document) running global copy and paste operations and deleting it all.

I also like to get the covers for the original books and use them instead of a bare heading.

I haven't yet decided which covers I'll end up going with, but yes, I'll be devoting time to covers soon. (Also considering a possible "covers gallery" side project featuring all covers from all editions.)

Oh, and footnotes. I ended up going with an endnote approach, where each book has its own endnotes section.

[DNSB]

Originally Posted by Fitz Frobozz

Are you retaining separate stylesheets for each book just because it's easier that way, or because you prefer to style the books differently? My plan has been to use pretty much the same styling for all the books using my own style guide, which is based on elements from the source book, previous editions of said book, general best practices, and my own preferences.

The separate stylesheets are due to stylesheets often being different even with books from the same publisher and/or author. It was just easier to stick with EpubMerge's approach than to try to match the stylesheets in each book that are going into the omnibus. This approach also allows you to have identical filenames in different directories rather than renaming them. The quick and dirty approach.

I have gone the uniform stylesheets approach for a couple of omnibus editions I put together for one author. Since there were class names that were reused between the books with different information, I located those and renamed them as needed and then merged the stylesheets into a single stylesheet. Redo this effort for the images. After this an import and rename of each book's text files and finally removing any duplicate images, classes, etc. Many thanks to Kovid Goyal, KevinH and diapdealer for the tools supplied with Sigil and calibre's e-book editor that allowed me to automate much of the process.

[Quoth]

Yes, for an automatic merge the stylesheets need renamed (if the same name) as they often have same name CSS entries (styles) that are totally different.

A merged CSS approach needs renaming of styles that don't match. I've done that for a small omnibus that needs the CSS hugely edited anyway.

Of course if you have the source, or export docx/RTF and properly re-edit in MS Word (or LO Writer as odt) and re-import to Calibre as docx, then you'll have one sensible CSS.

[Fitz Frobozz]

Update!

I just hit a major milestone and am feeling pretty good. All three volumes are structurally complete and seem to have passed all major tests required as a minimum for submission. I'm sure I've more testing to do but I'm starting to relax.

I've learned a TON during this project, and if/when I do this again I won't be using many/any of the strategy ideas I mentioned in the OP. I'll post again at some point when I have more time to outline the kind of approach and strategy that I think would make more sense now that I know what I know. Edit: Oh, and Sigil wound up becoming my editor of choice. I should have explored it more before embarking.

Next on my list:

1. Finalize main volume cover images (created by me — book title pages use original book cover images)
2. Finish the volume frontmatter pages which are only about halfway done
3. Media queries, device-specific styling changes, other optimizations
4. Semantics and other backend improvements/optimizations
5. Styling tweaks & fixes (most if not all significant issues have already been fixed)

Regarding semantics, I'm a little unfamiliar with how they should be implemented (if at all — I should probably mention that (1) I'm still an ePub semantics noob and (2) the project has evolved a bit since the OP and it may actually be put out into the wild, down the road.)

This is the tree I'm currently going with as of this writing, which reflects the structure I'm using and the semantic terms that I think would make sense.

Thoughts?

volume
├──cover (Volume cover page)
├──titlepage (Volume title page)
├──copyright-page ("Copyright Info")
├──preamble ("Editors Note")
├──colophon ("Edition Notes")
├──acknowledgements ("Acknowledgements")
├──preamble ("Tentative Timeline")
├──division (Book)
│....├──titlepage? (Book title pages with book cover images)
│....├──toc (Book TOCs - not sure how to distinguish, semantically, from Volume TOCs)
│....└──bodymatter
│..........├──part (Book sections, level 1)
│..........│....├──preamble (Maps - not sure at all about this one, just picked it as it seemed to be the closest fit)
│..........│....├──introduction (Book introductions)
│..........│....├──prologue (Book prologues)
│..........│....└──chapter (Book sections, level 2)
│..........│..........├──endnotes ("Notes" — chapter endnotes)
│..........│..........└──chapter part Book section, level 3 — not sure what to call this, semantically)
│..........│................└──endnotes ("Notes" — chapter part endnotes)
│..........├──glossary
│..........└──afterword
├──toc (Volume TOCs - not sure how to distinguish, semantically, from Book TOCs)
└──appendix (haven't actually developed this idea much yet, still considering ideas)