A description of a PDF's structure - Page 2

chikamichi · 12-13-2015, 01:37 PM

Thank you, I was finally able to tweak ebook-convert flags and generate the debug output!

Actually, all my CSS rules are properly applied and merged with Calibre's, as inspecting processed/stylesheet.css revealed. The structure of the HTML file at this stage in the process is:

Code:

<?xml version='1.0' encoding='utf-8'?>
<html xmlns="http://www.w3.org/1999/xhtml" lang="en">
  <head>
    <title>Table of Contents | Willkommen, bienvenue, benvenuto, bienvendia, bem-vindo, ahlan wa sahlan, добро пожаловать, 환영, 欢迎,welcome!</title>
    <meta content="" name="description"/>
    <meta content="GitBook 2.6.5" name="generator"/>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
  <link href="stylesheet.css" rel="stylesheet" type="text/css"/>
<link href="page_styles.css" rel="stylesheet" type="text/css"/>
</head>
  <body class="dir">        
<div class="page">
<div class="section">
    <h1 class="calibre">Table of Contents</h1>
    <ol class="calibre1">
        <li class="calibre2">
            <span class="inner">
                        <a href="index.html" class="calibre3 pcalibre1 pcalibre2 pcalibre">Introduction</a>
                <span class="page1">0</span>
            </span>
        </li>
        <li class="calibre2">
            <span class="inner">
                        <a href="quick-start.html" class="calibre3 pcalibre1 pcalibre2 pcalibre">Quick-start</a>
                <span class="page1">1</span>
            </span>
        </li>
    </ol>
</div>
</div>
</body></html>

So, actually, what seems to happen is that when applying a CSS styling to "body" or ".dir" (both cases), the Output Plugin only considers what's inside the body tag, so it doesn't apply the rules.

Back to square one, then: is it possible to instruct Calibre/ebook-converter to apply a background that would extend the whole page? The only option I can think of is to apply that background to ".page", reduce its margins to 0 somehow while adding some padding. Not the best idea ever I'm afraid, though.

kovidgoyal · 12-13-2015, 01:46 PM

PDF output is generated by a browser (webkit) engine running with CSS 3 columns defined on body. In fact the same engine that powers the calibre viewer. If you can figure out how to convince a browser to show a background behind css3 columns defined on body, you can use it with ebook-convert.

chikamichi · 12-13-2015, 01:54 PM

Interesting. What exactly do you mean by "css3 columns defined on body" though?

chikamichi · 12-13-2015, 01:57 PM

Were you referring to https://github.com/kovidgoyal/calibr.../pdf_output.ui used for Calibre's UI ?

Edit: Oh, I guess it's actually https://github.com/kovidgoyal/calibr...r/from_html.py

kovidgoyal · 12-13-2015, 02:28 PM

google css3 columns. ANd if you want to find it in the code, look for paged.coffee

chikamichi · 12-13-2015, 02:32 PM

Thank you.

chikamichi · 12-13-2015, 05:05 PM

Ok, I've been trying to set CSS rules that would affect the entire PDF document (as produced by Calibre's Output Filter leveraging the logic of paged.coffee), but no luck.

I'm not giving up, but meanwhile, if you have any idea how one could instruct the webkit-based renderer to apply CSS rules to a PDF page as a whole/unit, that would be great. Especially, something that would target the whole page. It's kind of disappointing setting a simple background for the PDF document has to be so damn hard a task

kovidgoyal · 12-13-2015, 11:02 PM

I know of know way to do that in HTML. But a bit of googling will show you plenty of ways you can postpocess a pdf to add background images, for example pdftk multistamp.

chikamichi · 12-14-2015, 06:08 AM

Thank you, that's good to know.

I'd rather have ebook-convert handle and automate that operation though, for having the end-user use the CLI in an extra step doesn't fit my use-case.

chikamichi · 12-14-2015, 06:16 AM

Extra question: how about extending the current behavior found in https://github.com/kovidgoyal/calibr...r/from_html.py so instead of enforcing a white background (as per line 353), it would first check whether a value has been supplied in the CSS? Something like paged_display.set_background() to be implemented within paged.coffee (at least being able to change the color would be nice; of course it would be even nicer being able to set backgroundImage and backgroundCover properties).

kovidgoyal · 12-14-2015, 07:12 AM

That will not work, the body background color does not apply to margins.

kovidgoyal · 12-14-2015, 07:53 AM

Quote:

Originally Posted by chikamichi

I'd rather have ebook-convert handle and automate that operation though, for having the end-user use the CLI in an extra step doesn't fit my use-case.

What exactly are you doing? I'm guessing you are trying to embed ebook-convert inside some other software.

chikamichi · 12-14-2015, 09:34 AM

Quote:

Originally Posted by kovidgoyal

What exactly are you doing? I'm guessing you are trying to embed ebook-convert inside some other software.

The exact context I'm working on is GitBook. It's an open-source project, one can use it their CLI locally, or one can leverage the website itself to benefit from a fully automated process, from writing to hosting books. Basically, one writes markdown+css and gets a fully-fledged book in different format.

In a team I'm working with, I'd like to have non-tech people use GitBook's website to manage some public-facing documentation, so that 1/ rendered documents are all in sync, no matter the format (html, pdf, mobi, etc.) 2/ it becomes easier for translators to follow edits using git/GitHub (for GitBook's website can push to GitHub).

The only "special" use-case I'd have is for the PDF rendering of a document, to have an image as its background (same bg on each page, full-extend, covering margins). Although I can provide CSS styling in GitBook for the body tag of the HTML that's going to be provided to ebook-convert, it eventually gets wiped out for the reasons we demonstrated above (thx again for your insights).

kovidgoyal · 12-14-2015, 09:36 AM

So why not just use pdftk, make it part of a script that first runs ebook-convert and then calls pdftk?

chikamichi · 12-14-2015, 10:08 AM

As I said, I'd rather have the non-tech people use a down-to-earth solution like GitBook: "click a button, and it's ready"

I'll see whether adding post-processing (using pdftk for instance) to GitBook's workflow is an option, then. Thank you!

12-13-2015, 01:57 PM	#19
chikamichi Member Posts: 17 Karma: 10 Join Date: Dec 2015 Device: none	Were you referring to https://github.com/kovidgoyal/calibr.../pdf_output.ui used for Calibre's UI ? Edit: Oh, I guess it's actually https://github.com/kovidgoyal/calibr...r/from_html.py Last edited by chikamichi; 12-13-2015 at 02:03 PM.

12-13-2015, 05:05 PM	#22
chikamichi Member Posts: 17 Karma: 10 Join Date: Dec 2015 Device: none	Ok, I've been trying to set CSS rules that would affect the entire PDF document (as produced by Calibre's Output Filter leveraging the logic of paged.coffee), but no luck. I'm not giving up, but meanwhile, if you have any idea how one could instruct the webkit-based renderer to apply CSS rules to a PDF page as a whole/unit, that would be great. Especially, something that would target the whole page. It's kind of disappointing setting a simple background for the PDF document has to be so damn hard a task Last edited by chikamichi; 12-13-2015 at 06:30 PM.

12-14-2015, 06:16 AM	#25
chikamichi Member Posts: 17 Karma: 10 Join Date: Dec 2015 Device: none	Extra question: how about extending the current behavior found in https://github.com/kovidgoyal/calibr...r/from_html.py so instead of enforcing a white background (as per line 353), it would first check whether a value has been supplied in the CSS? Something like paged_display.set_background() to be implemented within paged.coffee (at least being able to change the color would be nice; of course it would be even nicer being able to set backgroundImage and backgroundCover properties). Last edited by chikamichi; 12-14-2015 at 06:43 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[old thread] filename and library structure /author and titel structure	tscamera	Library Management	4	05-31-2011 05:44 PM
Interesting behavior of Structure Detection PDF to MOBI	tleon	Conversion	8	05-04-2011 05:29 PM
description for the lrf structure	joblack	LRF	2	08-01-2009 03:35 PM
Other Non-Fiction Abelson, H; Sussman G: Structure and Interpretation of Computer Programs. PDF. v1.0	scottdw	Other Books	0	07-05-2008 08:43 AM

12-13-2015, 01:46 PM	#17
kovidgoyal creator of calibre Posts: 43,858 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	PDF output is generated by a browser (webkit) engine running with CSS 3 columns defined on body. In fact the same engine that powers the calibre viewer. If you can figure out how to convince a browser to show a background behind css3 columns defined on body, you can use it with ebook-convert.

12-13-2015, 01:54 PM	#18
chikamichi Member Posts: 17 Karma: 10 Join Date: Dec 2015 Device: none	Interesting. What exactly do you mean by "css3 columns defined on body" though?

12-13-2015, 02:28 PM	#20
kovidgoyal creator of calibre Posts: 43,858 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	google css3 columns. ANd if you want to find it in the code, look for paged.coffee

12-13-2015, 02:32 PM	#21
chikamichi Member Posts: 17 Karma: 10 Join Date: Dec 2015 Device: none	Thank you.

12-13-2015, 11:02 PM	#23
kovidgoyal creator of calibre Posts: 43,858 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	I know of know way to do that in HTML. But a bit of googling will show you plenty of ways you can postpocess a pdf to add background images, for example pdftk multistamp.

12-14-2015, 06:08 AM	#24
chikamichi Member Posts: 17 Karma: 10 Join Date: Dec 2015 Device: none	Thank you, that's good to know. I'd rather have ebook-convert handle and automate that operation though, for having the end-user use the CLI in an extra step doesn't fit my use-case.

12-14-2015, 07:12 AM	#26
kovidgoyal creator of calibre Posts: 43,858 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	That will not work, the body background color does not apply to margins.

12-14-2015, 09:36 AM	#29
kovidgoyal creator of calibre Posts: 43,858 Karma: 22666666 Join Date: Oct 2006 Location: Mumbai, India Device: Various	So why not just use pdftk, make it part of a script that first runs ebook-convert and then calls pdftk?

12-14-2015, 10:08 AM	#30
chikamichi Member Posts: 17 Karma: 10 Join Date: Dec 2015 Device: none	As I said, I'd rather have the non-tech people use a down-to-earth solution like GitBook: "click a button, and it's ready" I'll see whether adding post-processing (using pdftk for instance) to GitBook's workflow is an option, then. Thank you!

Advert

Advert