duanyao
commented
8 years ago Converting output of pdf2htmlex to epub3 is definitely doable (we did that actually). Of cause you have to do some programing to generate the navigation document and the OPF.
What epub readers are you targeting? I believe most epub3 readers are based on morden browser engines, and SVG produced by pdf2htmlex (via cario) should be viewable in browsers. Morden browsers support a mixture of SVG 1.1 and SVG 1.2 tiny, by the way.
I think unicode-bidi:bidi-override is related to right-to-left scripts (see #595). Why it must be removed?
YenForYang
anteprandium
RNCTX
I realize that pdf2htmlex can't make reflowable text from PDFs atm, but I've been wanting to generate ebooks with help of pdf2htmlex and/or some other tools from PDFs to some ereader format. For now I see that a fixed-layout EPUB is the only semi-viable opton that hopefully can temporarily satisfy my needs.
Basically, I've been trying to figure out a way to configure pdf2htmlex to generate xhtml documents that can be imported easily into an EPUB3 (at the moment I've been trying to get it to work with Sigil). I've been looking at Eric Dodémont's book called "Fixed Layout ePub: A Practical Guide to Publish eBooks from PDF Files" (I actually have a PDF of it, so if @coolwanglu wants to a look or anyone else I can send them a copy perhaps--some basic info/introduction about the topic is here) the book (and the webpage I guess) has some useful but rather basic information on getting pdf2htmlex to generate the files, but the method is essentially outlined as:
(note that this is considered one method)
It is mentioned, however, that the pages generated from pdf2htmlex require a deal of further editting, (the editting is not automated at all really, but I have found workarounds for these............but perhaps pdf2htmlex can be configured to not require it...?). Basically, the split *.pages generated from pdf2htmlex have to be prepended and appended with some html code so that each .page can act like a seperate xhtml page (pretty much like the main html page, but each page has its own page-container and etc.). Also, The HTML viewer that pdf2htmlEX produces and integrates in the result has to be removed (not quite sure how to do this, but think I do in a way---do I edit the manifest file in the share directory in some way (I'm not too sure what I can and cannot touch in that file, or what to edit out). Other things that won't be used include:
One thing that I really really want to figure out is how to integrate is the .outline file in the epub. I cannot figure out how to get the links to work from the outline (I renamed .outline to .html and clicked the links, but, sadly, they did not work.) If I could use the .outline to create a linked table of contents (in other words, to create the nav.xhtml file) it would save SO much time. There has to be some way to modify it to get it working with epubs right--perhaps the page id's/no's need to be changed to work with seperate xhtml pages? The nav.xhtml file basically has to end up looking like this (which is different from the .outline file):
Probably important to note here that the href's (I noticed) are all to a specific .page/.xhtml file rather than in the outline file which has #pf... hrefs. Also, it doesn't have to be unpacked/decompressed/beautified code like I have above (this is true for all of the code pretty much).
Also, a question about using svg images, because I think the pdf2htmlex generator creates v1.2 svg but epub3 requires v1.1...I don't know, but according to the book I can safely change the 1.2 value to 1.1 without any problems using a sed command. Are there are any problems with using 1.1 for 1.2? If there are, let me know if there are any workarounds.
Also having trouble with getting the content.opf file to work (it's the file containing all the references to all the files with their type (manifest) and page order (spine). I believe it is this file that is causing the outline/nav.xhtml/tableofcontents to not work. The overall opf file has to look something like this:
Also, I believe I have to edit the base.min.css to be not include this at all:
and I believe it is actually included twice in the base.min.css file. Is there any problem to doing this? I think not, but just want to make sure...
Luckily, that is all haha Those are all the issues I believe...I'm hoping that there files I can modify so that the conversion/generation process will automatically resolve the above things to work.