[Jackbishop]

I'm new here (and new to eBook creation) -- let me know if I'm generally screwing up on local etiquette. Or if I'm in the wrong forum -- I thought this might be a MOBI or a Calibre issue, but I'd like to know what my other options are, and, as can be seen below, I'm juggling formats besides MOBI in intermediate steps.

I've been playing with the Kindle as a platform for comic books: chiefly manga scanlations and a handful of the full-page comics which have been made available online. For several months I used the Kindle's (fairly dire and buggy) navigation capabilities for individual images. And then I learned that Calibre can convert CBZs to MOBIs, resulting in better refresh time and better navigability than trying to navigate among individual JPEGs.

So far, a success story. But my ambition got the better of me and I wanted more. Ordinary eBooks, I knew, had chapters to which I could navigate with the 4-way; seeing as how each individual chapter (of a manga) or issue (of an American comic) was in a different folder, it shouldn't be hard to get a chapter-navigable mobi instead of a monolithic one with either chapter-breaks every page (default CBZ->MOBI structure detection) or none at all (forcing Calibre to not detect structure).

I've spent hours on this and am (a) still not really producing satisfactory output, and (b) pretty certain I'm making it more complicated than I have to.

It might be worth mentioning here that I'm running Linux and would prefer a native and/or scriptable (i.e. command line) solution. I'm definitely interested in _any_ solution or constructive input to the problem, but those parameters have, thus far, directed the project.

My first plan was: make each folder into its own CBZ, convert the CBZs to MOBIs, merge the MOBIs so each individual part is a chapter of the monolithic MOBI. That would be perfect and easy except that there's no tool to actually do the last step, as far as I know.

There _is_ an ePub-merging tool (LRFTools has an ePub-merge mode), but direct conversion of CBZ->ePub->MOBI with calibre was, for reasons I couldn't quite get, putting a blank page between each page of comic content. CBZ -> MOBI -> ePub -> MOBI worked, and the intermediate ePub could be worked with.

And that's more or less where I am now. I'm zipping comic-book images folderwise into CBZs, converting CBZs to MOBIs and then those MOBIs into ePubs with calibre, using LRFTools to merge the ePubs, and converting the monolithic ePub back to a MOBI.

Here's the (extremely unpolished at present) script I'm using to do it, for reference:

#!/bin/bash
tempdir="$(mktemp -d)"
for curdir in $*; do
    if test -d $curdir; then
	zip $curdir.cbz $curdir/*.jpg $curdir/*.png $curdir/*.gif
	ebook-convert $curdir.cbz  $tempdir/$curdir.mobi --output-profile kindle_dx --keep-aspect-ratio --output-format png
	ebook-convert $tempdir/$curdir.mobi  $tempdir/$curdir.epub
	rm $curdir.cbz
    fi
done
java -Xms200M -Xmx300M -jar ~/LRFTools-0.9.236.jar mergeEPUB $tempdir -a "Monolithic--EDIT" -t "Monolithic--EDIT" -o monolith.epub

This would actually work OK... but LRFTools's mergeEPUB mode is a bit janky and crude. It takes a folder as a parameter and merges the ePubs in that directory... in no particular order, so the issues are all out of order. Plus it puts a title page on each, which I might be able to hack away by putting the issue cover in metadata instead of in the text, but even with all of that, it gets a bit peculiar about the last navpoint in ways I can't quite suss out and which seem to, under particularly unfavorable circumstances, crash my Kindle. Plus the whole MOBI -> ePub -> MOBI thing rubs me the wrong way.

Am I making this whole thing a lot harder than it has to be? At this point I'd even be amenable to a tool that puts navpoints in manually, even though it'd be better if there were an automatic way to translate directory structure of the source to navigational structure of the output.