[FM Discuss] Challenges in rendering automatic covers

[Tuukka Hastrup]

2011-10-14 05:34, adam kirjoitti:
> actually, why not just use html and convert to pdf?

Yeah, maybe that's ok for now although it doesn't give as much options 
as Inkscape for the artist.

To get the required PDF with front, spine and back on one page from HTML 
format, we would need to figure out a way to combine the three separate 
PDF pages, *or* make the combined layout as well as the spine text 90 
degree rotation in CSS. Would anyone know how to do a HTML/CSS design 
like this right away?

> On 10/14/2011 01:50 PM, Tuukka Hastrup wrote:
>>
>> I made some progress with help from Elisa on IRC.
>>
>> First, she sent this example of what kind of a result we're ideally
>> targetting, although it could be much simpler at this point:
>> http://tuukka.iki.fi/tmp/example-cover.png
>>
>> Second, a new idea to alternatively use Scribus for this came up in the
>> discussion, as Scribus is meant for page-making given text and image
>> assets. Unfortunately, on the IRC channel #scribus, I heard that the
>> efforts to make Scribus suitable for automatic rendering are still
>> experimental. Thus, someone could explore e.g. running Scribus headless
>> with Xvfb and scripting it with Scribus Scripter2 which was part of the
>> GSOC this year:
>> http://scripter2.jainbasil.net/
>> http://jainbasil.net/scripter2/
>>
>> Third, on the IRC channel #inkscape, I learned about the current
>> limitations of the SVG approach such as flowed text (automated
>> text-layout) being Inkscape-only and Inkscape not supporting
>> high-quality rendering of SVG images nested within an outer SVG image.
>> It might be possible to convert the flowed text to non-flowed with
>> Inkscape then use something like wkhtmltopdf to convert standard SVG to
>> PDF. wkhtmltopdf crashes when I try to convert an SVG though. The link I
>> got about nested SVG in Inkscape was this:
>> http://article.gmane.org/gmane.comp.graphics.inkscape.user/14208
>>
>> The three earlier issues still remain as well:
>>
>> 2011-10-14 02:05, Tuukka Hastrup wrote:
>>> I hit some "challenges" trying to render book covers automatically as
>>> needed for lulu.com integration. At the moment, I can render a blank PDF
>>> of the right dimensions ;-) I'm wondering if anyone else has ideas on
>>> how to solve (or work around) these:
>>>
>>> 1. How to express in SVG a float layout similar to an IMG tag floating
>>> on the right of a text paragraph in HTML?
>>>
>>> 2. Alternatively, how to combine separate HTML-to-PDF renderings of a
>>> front cover, spine, and back cover into a single PDF *page*?
>>>
>>> 3. In either case. how to scale the font size of the book title in SVG
>>> *or* HTML so that the title width matches the horizontal page available
>>> in the cover?
>>>
>>> Actually, I wouldn't mind if someone else wanted to whip up something
>>> for this as it's only tangential even though required for the lulu.com
>>> integration.
>>>
>>>
>>> Regards
>>> Tuukka
>>>
>>> P.S. Here's what I use for the blank test PDFs: wkhtmltopdf-i386
>>> --page-width 210 --page-height 297 --load-error-handling ignore
>>> /dev/zero a4.pdf
>>>
>>> _______________________________________________
>>> Booki-dev mailing list
>>> Booki-dev at lists.flossmanuals.net
>>> http://lists.flossmanuals.net/listinfo.cgi/booki-dev-flossmanuals.net
>> _______________________________________________
>> Discuss mailing list
>> Discuss at lists.flossmanuals.net
>> http://lists.flossmanuals.net/listinfo.cgi/discuss-flossmanuals.net
>