[GUI Plugin] Modify ePub

[@Escape]

Originally Posted by theducks

Calibre is now at 1.5
is there a reason not to update?

Many PI needed fine tuning when the backend code changed, the needed calibre level requirements may have been missed and you got a version you should not have.
Kiwidude or Kovid will need to answer that.

I know but in first post => Special Notes: Requires Calibre 0.8.53 or later

[theducks]

Originally Posted by @Escape

I know but in first post => Special Notes: Requires Calibre 0.8.53 or later

That is why I said the requirement update may have been missed. A lot of this is process all done by hand. (Don't try and count my typos: Boundary overflow)

You did no wrong move.

[JSWolf]

Originally Posted by Rev. Bob

Merely importing a book into Calibre adds the metadata. As I understand it, that's a prerequisite to being able to use Calibre's TOC Editor on it...

EDIT: As for an example, I just added an EPUB into my Calibre library, did nothing to it, and used Save to Disk to make a copy. It made the following OPF changes that I consider "dirtying it up":

• Removed the BOM from the file, making it an improper UTF-8 file.
• Moved the title below the author, removing the indentation in the process.
• Modified the date/time.
• Removed the (empty) subject field.
• Inserted Calibre-specific metadata: title_sort, user_categories, and author_link_map (also with botched indentation)
• Inserted a contributor credit for itself.
• Inserted a second, Calibre-specific, identifier field.

You may be fine with those changes. I am not.

Save to disk is what is causing the conversion to happen and that is a USER error in this case. You need to go into Preferences and then Saving to disk. Uncheck all of the options except Show files in file browser after saving to disk. That will prevent the metadata update conversion.

I do so wish Calibre's default options for Saving to disk would be not to do any metadata updating and that in a new version, that the Saving to disk options get changed this way. People can change it back if they want.

[JSWolf]

Originally Posted by @Escape

@kiwidude

My Calibre version is 0.9.0

It's not a good idea to report errors based on an obsolete version of Calibre. Update to the latest version and then try Modify ePub again and if it then fails, report it.

[theducks]

I suspect this may be a minimum version required error
BUT
I only us a few setting/options

We need a screenshot (of the choices) of the configuration were being used when the error pops.

[kiwidude]

As the guys have said it is most likely my fault for not updating the minimum calibre version at some point - either Kovid moved some code around or I added a feature relying on a later version. There is absolutely no reason not to be using the latest calibre code in my opinion (1.x rocks), and your issue should disappear. When I next work on this plugin I will bump the minimum version...

[@Escape]

Thanks for your response. I have found a previous version 1.3.7 and all is OK now :-)

[JSWolf]

Originally Posted by @Escape

Thanks for your response. I have found a previous version 1.3.7 and all is OK now :-)

Why would you update to an old obsolete version of the plugin when it's just so easy to update Calibre and then be able to use the latest and greatest version of the plugin?

[docxtester]

Hello

I have noticed the following bug of this plugin:
When a file name inside an EPUB contains spaces, f.e. "file name.png"
and it is listed in a manifest as "file%20name.png"
It gets qualified as unused and gets deleted from an EPUB if remove unmafisted or unused files option is used.
The same happens if a file with such name is used in any HTML file inside an EPUB.
The same problem exists when any file name contains non ANSI characters and those characters are translated inside manifest according to URL translation rules.

Be careful with those options now and I would appreciate a fix

[ShellShock]

I posted about this last year, so apologies for a bit of repetition here. I really like this plug-in, except I find the "Smarten punctuation" a bit of a blunt tool, so I have created my own version of the code in the plug-in:

        def smarten_punctuation_for_page(html):
            preprocessor = HeuristicProcessor(None, self.log)
            start = 'calibre-smartypants-'+str(uuid4())
            stop = 'calibre-smartypants-'+str(uuid4())
            html = html.replace('<!--', start)
            html = html.replace('-->', stop)
            html = preprocessor.fix_nbsp_indents(html)
            # Do not use Calibre smartyPants.
            # html = smartyPants(html)
            html = html.replace(start, '<!--')
            html = html.replace(stop, '-->')
            # convert ellipsis to entities to prevent wrapping
            html = re.sub(r'(?u)(?<=\w)\s?(\.\s?){2}\.', '&hellip;', html)
            # convert double dashes to em-dash
            html = re.sub(r'\s--\s', u'\u2014', html)
            # Convert short em-dash to em-dash.
            html = re.sub(u'\u2013', u'\u2014', html)
            # Convert space, hyphen, space to em-dash.
            html = re.sub(r'\s-\s', u'\u2014', html)
            # Convert punctuation, hyphen to em-dash.
            html = re.sub(r'([\'|"|\u2018|\u201C])-(\w)', r'\1\u2014\2', html)
            # Convert hyphen, punctuation to em-dash.
            html = re.sub(r'(\w)-([\'|"|\u2019|\u201D|?|!|,|\.])', r'\1\u2014\2', html)
            # Convert space, em-dash, space to em-dash.
            html = re.sub(r'\s*\u2014\s*', u'\u2014', html)
            return substitute_entites(html)

You can see I do not use the smartypants within Calibre, and I have added some fixes for em-dash. Perhaps there could be an "Advanced" options in the plug-in UI to give the user finer control over exactly what the smarten punctuation does?

[cybmole]

i cannot puzzle out how this plug in decides what dimensions to use when it does "insert cover image"

I can find an 600 x 800 image via calibre cover search but when that gets put into the epub via this plugin, then the generated code is image height="751" width="563"
in another test, source of 600 x 895 goes in as image height="751" width="489"
I've experimented with various sized cover, and with 2 pass remove cover followed by inset cover, but I cannot suss out how the cover dimensions as displayed in calibre edit metadata screen are translated into image dimensions code placed in the epub.

it is bugging me a little now as I have a sony T2 and on its display of recent books on its home page, my cover images are raggedy. ( they are neatly rescaled on the actual bookshelf pages though). did not happen with my previous prs-650 so I never focussed on this issue before.

I'd really like to use some combo of calibre+ this plug in to get 600 x 800 cover images ( or as close to that as is feasible ) , so they all look the same size on the sony T2 recent books display, rather than display with different heights .any suggestions ?

[kiwidude]

Cybmole, it is all to do with your general defaults, and what profile you have chosen for your device. There are settings to do with proportions if I recall correctly if you go to ePub defaults. And the overall dimensions/rescaling comes from what device you have chosen, pick tablet for no rescaling at all (but be aware that mobi conversions using % for margins etc will be completely screwed up).

[Rev. Bob]

Originally Posted by cybmole

the generated code is image height="751" width="563"
in another test, source of 600 x 895 goes in as image height="751" width="489"

Originally Posted by cybmole

I'd really like to use some combo of calibre+ this plug in to get 600 x 800 cover images ( or as close to that as is feasible ) , so they all look the same size on the sony T2 recent books display, rather than display with different heights .any suggestions ?

Both of your examples show the same height, with different widths to maintain the proper aspect ratios. If you want images that are all the same size in both dimensions, you can't just ignore the aspect ratio...

[cybmole]

thanks guys, that makes sense. the sony T2 reader does something strange when it generates recently added thumbnails as I see different heights there, not just different widths , but without being able to search for T2, that is going to be a pain totrack down. I guess that i can live with it. it may be using cover images sent separately by calibre, not those embedded in the books.

if I am understand the previous answers correctly, there is a dependency between this plug in and the epub output profileselected in calibre, so ( to tlnk to my other thread query) changing my output profile to T3 ( more pixels) will affect the code generated by plugin insert image ?

[jcloggins]

Useful tool but I have had a problem with the preserveAspectRatio property of the svg element when replacing the cover image.

Setting the value to "none" means that the cover scales to fill the entire window. The result is a very distorted cover image when the display is in landscape mode.

My usual fix is to edit the cover page and delete the property, which makes the image scale properly.