[GUI Plugin] Modify ePub

[paulfiera]

Many thanks, kiwidude.

Got it.

Originally Posted by kiwidude

...leaving such orphaned declarations in the css file does absolutely no harm, and the reader will just fall back to using its default font which is what people wanted.

The only downside is that calibre removes these font declarations from the css and puts them in every html/xhtml file when doing a epub -> epub conversion - I don't know why calibre does this. So I usually open the css file after removing the fonts and delete the font declarations.

[DoctorOhh]

Originally Posted by paulfiera

The only downside is that calibre removes these font declarations from the css and puts them in every html/xhtml file when doing a epub -> epub conversion - I don't know why calibre does this. So I usually open the css file after removing the fonts and delete the font declarations.

During the ePub to ePub conversion have you tried selecting the option to completely remove all fonts? See attached.

[paulfiera]

Originally Posted by dwanthny

During the ePub to ePub conversion have you tried selecting the option to completely remove all fonts?

Thanks, Walt.

Just tried it. Same results. Font declarations are moved from the css file to every html/xhtml file.

I really wonder why calibre does that.

[kiwidude]

I've moved a bunch of recent posts from this thread about other additions I may make to this plugin over on the related thread in the Development forums, but will try to keep the hopefully slightly less technical discussions on this thread.

The new features intended for this next release can be seen in the screenshot below:

• Remove Adobe .xpgt files and links - for completely clean xpgt file removal (deletes the file and the <link> references from the html files)
• Remove Adobe resource DRM meta tags - to strip some Adobe cruft of a <meta> tag from the html files.
• Remove embedded fonts - extended to also remove @font-face declarations from the CSS and html files
• Remove unused image files - remove orphaned images not referenced from the html content to save space.
• Flatten TOC hierarchy in NCX file - move all the navPoints in the NCX file to a single level if they are nested, to support devices that don't like a nested TOC

This is a beta version!. Please make sure you have a backup of your ePub(s) you are modifying when you are experimenting with these new options in case I have badly screwed up. I've given them a quick whirl as a sanity check but there could be edge cases I've missed.

Also note there is a new beta version of the Quality Check plugin here which has matching checks for all of the above. So please feel adventurous and install both together to more easily test them.

Feedback very much welcomed before I officially release it. The other features I am considering working on can wait for a future version...

[JSWolf]

When this plugin is used, does Calibre do any modification to the ePub other then the function(s) of the plugin being used? For example, removing the page-template file and links, is the ePub the same other then these missing things?

[paulfiera]

Originally Posted by kiwidude

• Remove Adobe .xpgt files and links - for completely clean xpgt file removal (deletes the file and the <link> references from the html files)

Thanks, kiwidude.

I never understood the meaning of this xpgt files. They are not really needed?

[paulfiera]

@kiwidude: currently testing this beta and it's working great but it looks like you have to do a couple of passes to remove every checked option.

With "Remove inline @font-face declarations" checked and the three Adobe options checked, this is what I've seen:

In the first pass, it removes the declarations and the xpgt file.
In the second pass, it removes the Adobe DRM meta tags if found.
In the third pass, it removes the link to the xpgt file.

No ill effects observed.

Great work !

[kiwidude]

Originally Posted by JSWolf

When this plugin is used, does Calibre do any modification to the ePub other then the function(s) of the plugin being used? For example, removing the page-template file and links, is the ePub the same other then these missing things?

That is exactly the ethos of this plugin. It does not use any of calibre's conversion pipeline code, nor does it do any css or html parsing/rewriting.

So if you choose the xpgt files & links option for example, the only change that should be made is to the opf file to remove the reference to the xpgt, the xpgt file itself will be deleted, and then on each html page that had a <link .../> it will do a simple raw text regex based replacement of just that.

[kiwidude]

@paulfiera - three passes should not be necessary (that is what it does internally in the code, but you should only have to click the ok button once). I must admit I only tested the options individually as I wrote them, I'll take a look at this shortly.

[paulfiera]

Thanks, kiwidude.

This seems to happen when the three conditions are met:

• - in line @font-face declarations.
• - Meta DRM tag.
• - Link to Adobe xpgt file.

I've tested it by running the plugin and opening the epub in Sigil. And it looks like the removal of these elements always happens in the same sequence.

[kiwidude]

Ok, cool. It does bring up another question I meant to ask - do the removal of @font-face declarations have to be so granular? At the moment removing embedded fonts, css @font-face and html @font-face delcarations are three separate options.

I would anticipate most users being happy to just tick "Remove embedded fonts" and have it also remove *any* font-face declarations. Are there any scenarios at all where this would not be the case?

[paulfiera]

Originally Posted by kiwidude

I would anticipate most users being happy to just tick "Remove embedded fonts" and have it also remove *any* font-face declarations. Are there any scenarios at all where this would not be the case?

As I see it, once you have decided to remove the embedded fonts, both the inline and css @font-face declarations make no sense at all.

My vote for the "Remove embedded fonts" checkbox taking care of everything. Fonts and declarations.

[jackie_w]

My 2p's worth... epubs with embedded fonts should also have @font-faces (in either css or html). Other epubs will have @font-faces (in either css or html), but no embedded font files, i.e. referencing external font files. I would be happy to have a single 'Remove embedded fonts' option which handles both scenarios. I don't need to know which scenario was found, I just want it all gone.

[kiwidude]

Thanks folks, I have rolled them all together for the next version I will post shortly. Paul I found the bug in the "individual step" stuff you were having to do so that will be sorted too.

It does beg the question as to whether Quality Check needs two checks then - should it just have a single "Check @font-face" which looks at both css and html files? The log can display where it found them...

[paulfiera]

Originally Posted by kiwidude

It does beg the question as to whether Quality Check needs two checks then - should it just have a single "Check @font-face" which looks at both css and html files? The log can display where it found them...

A single "Check @font-face" would be perfect.