[GUI Plugin] Modify ePub

[samy2]

In many ebooks i found font face entries for fonts not available in the ebook. It there a way to remove this entries with Modify EPUB? If not, is it possible to add such a feature to this tool?

Thank you very much for reading and considering this. And thank for your great tool, which saves me much time! :-)

[jackie_w]

Originally Posted by samy2

In many ebooks i found font face entries for fonts not available in the ebook. It there a way to remove this entries with Modify EPUB? If not, is it possible to add such a feature to this tool?

Have you tried the option 'Remove embedded fonts'? I think it's supposed to remove the font files (if any) and the @font-faces.

[theducks]

Originally Posted by samy2

In many ebooks i found font face entries for fonts not available in the ebook. It there a way to remove this entries with Modify EPUB? If not, is it possible to add such a feature to this tool?

Thank you very much for reading and considering this. And thank for your great tool, which saves me much time! :-)

Use the Quality check PI to determine which books have:
embedded fonts
and which have @fonts
(you will probably want to tag each case so you can the do a search for those with @fonts, but no embedded fonts.

Select all results

Now the Modify can remove the @fonts

[samy2]

I don't wanna remove any font. I just wanna remove font face entries which referring to fonts not available in the ebook. The calibre editor shows this entries but i have to remove it manually. Some have embedded fonts, which i don't wanna remove. But still there some wrong entries in font face. I want "quality check" just remove very wrong wrong entry.

If I have too look beforehand if there are embedded fonts, I just can remove those entries by myself with the editor. I just thought there is a more convenient way.

[theducks]

Originally Posted by samy2

I don't wanna remove any font. I just wanna remove font face entries which referring to fonts not available in the ebook. The calibre editor shows this entries but i have to remove it manually. Some have embedded fonts, which i don't wanna remove. But still there some wrong entries in font face. I want "quality check" just remove very wrong wrong entry.

If I have too look beforehand if there are embedded fonts, I just can remove those entries by myself with the editor. I just thought there is a more convenient way.

First you have to determine which have @fonts and no fonts
That is what the steps I explained do.
Without doing those,you run the risk of removing @fonts from books WITH fonts

Once you select the results of the combined search, Modify EPUB can remove the @fonts from the CSS of selected files

[JSWolf]

I found a bug...

calibre, version 1.39.0
ERROR: Modify ePub failed: No ePub files were updated

Modify ePubs
Logfile for book ID 12248 (The Silver Spike / Glen Cook)
12248
  Modifying:  C:\Users\Jon\AppData\Local\Temp\calibre_2aeddf\onhui4_modify_epub\12248.epub
Parsing xml file: OEBPS/package.opf
Parsing xml file: OEBPS/toc.ncx
	Looking for broken links in the NCX
Changing playOrder from: 0 to: 1
Changing playOrder from: 0 to: 2
Changing playOrder from: 0 to: 3
Changing playOrder from: 0 to: 4
Changing playOrder from: 0 to: 5
Changing playOrder from: 0 to: 6
Changing playOrder from: 0 to: 7
Changing playOrder from: 0 to: 8
Changing playOrder from: 0 to: 9
Changing playOrder from: 0 to: 10
Changing playOrder from: 0 to: 11
Changing playOrder from: 0 to: 12
Changing playOrder from: 0 to: 13
Changing playOrder from: 0 to: 14
Changing playOrder from: 0 to: 15
Changing playOrder from: 0 to: 16
Changing playOrder from: 0 to: 17
Changing playOrder from: 0 to: 18
Changing playOrder from: 0 to: 19
Changing playOrder from: 0 to: 20
Changing playOrder from: 0 to: 21
Changing playOrder from: 0 to: 22
Changing playOrder from: 0 to: 23
Changing playOrder from: 0 to: 24
Changing playOrder from: 0 to: 25
Changing playOrder from: 0 to: 26
Changing playOrder from: 0 to: 27
Changing playOrder from: 0 to: 28
Changing playOrder from: 0 to: 29
Changing playOrder from: 0 to: 30
Changing playOrder from: 0 to: 31
Changing playOrder from: 0 to: 32
Changing playOrder from: 0 to: 33
Changing playOrder from: 0 to: 34
Changing playOrder from: 0 to: 35
Changing playOrder from: 0 to: 36
Changing playOrder from: 0 to: 37
Changing playOrder from: 0 to: 38
Changing playOrder from: 0 to: 39
Changing playOrder from: 0 to: 40
Changing playOrder from: 0 to: 41
Changing playOrder from: 0 to: 42
Changing playOrder from: 0 to: 43
Changing playOrder from: 0 to: 44
Changing playOrder from: 0 to: 45
Changing playOrder from: 0 to: 46
Changing playOrder from: 0 to: 47
Changing playOrder from: 0 to: 48
Changing playOrder from: 0 to: 49
Changing playOrder from: 0 to: 50
Changing playOrder from: 0 to: 51
Changing playOrder from: 0 to: 52
Changing playOrder from: 0 to: 53
Changing playOrder from: 0 to: 54
Changing playOrder from: 0 to: 55
Changing playOrder from: 0 to: 56
Changing playOrder from: 0 to: 57
Changing playOrder from: 0 to: 58
Changing playOrder from: 0 to: 59
Changing playOrder from: 0 to: 60
Changing playOrder from: 0 to: 61
Changing playOrder from: 0 to: 62
Changing playOrder from: 0 to: 63
Changing playOrder from: 0 to: 64
Changing playOrder from: 0 to: 65
Changing playOrder from: 0 to: 66
Changing playOrder from: 0 to: 67
Changing playOrder from: 0 to: 68
Changing playOrder from: 0 to: 69
Changing playOrder from: 0 to: 70
Changing playOrder from: 0 to: 71
Changing playOrder from: 0 to: 72
Changing playOrder from: 0 to: 73
Changing playOrder from: 0 to: 74
Changing playOrder from: 0 to: 75
Changing playOrder from: 0 to: 76
Changing playOrder from: 0 to: 77
Changing playOrder from: 0 to: 78
Changing playOrder from: 0 to: 79
Changing playOrder from: 0 to: 80
Changing playOrder from: 0 to: 81
Changing playOrder from: 0 to: 82
Changing playOrder from: 0 to: 83
Changing playOrder from: 0 to: 84
Changing playOrder from: 0 to: 85
Changing playOrder from: 0 to: 86
Changing playOrder from: 0 to: 87
Changing playOrder from: 0 to: 88
ePub not changed after 0.22 seconds

That is a change to the ePub and it should allow the ePub to be updated. That was just using the option "Remove broken TOC entries in NCX file".

[JSWolf]

Originally Posted by Rev. Bob

Those make sense; added.

Attached are two versions of the latest code. One is the latest version of the one you've seen, with the de-Kobo and strip-span functions combined. The other separates them into two functions, both of which perform the self-closing combination and removal. The GUI, batch, and help files contain the proper information, except that I have not updated the help file to document the self-closing feature, the empty-removal feature, or to remove the "hey, this is experimental" note. In short, consider this a late-stage beta.

Feedback and testing is, naturally, more than welcome.

Very nice. I really dislike spans that do nothing at all. All that does is code bloat.

Now, can you come up with a mod to change those <span class="italics">, <span class="italic">, or <span class="bold"> into proper <em> and <b> of course with the proper close? That would solve one other pet peeve I have.

[JSWolf]

Originally Posted by samy2

In many ebooks i found font face entries for fonts not available in the ebook. It there a way to remove this entries with Modify EPUB? If not, is it possible to add such a feature to this tool?

Thank you very much for reading and considering this. And thank for your great tool, which saves me much time! :-)

That is a good idea.

[JSWolf]

Originally Posted by jackie_w

Have you tried the option 'Remove embedded fonts'? I think it's supposed to remove the font files (if any) and the @font-faces.

The idea here is to remove something like "font-family: Times-Roman, serif;" where there is no embedded font for Times-Roman thus leving all the properly embedded font code in place.

[Rev. Bob]

Originally Posted by JSWolf

I found a bug...

That is a change to the ePub and it should allow the ePub to be updated. That was just using the option "Remove broken TOC entries in NCX file".

Not something I've touched, so at least it's not something I broke! I'd suggest, right off the bat, making sure the output file isn't write-protected or otherwise a bad destination. If that's good, there may be other problems with the NCX file that changing the PlayOrders is not enough to fix. Perhaps if you could extract and attach that NCX file...?

Originally Posted by JSWolf

Now, can you come up with a mod to change those <span class="italics">, <span class="italic">, or <span class="bold"> into proper <em> and <b> of course with the proper close? That would solve one other pet peeve I have.

While I would LOVE that feature, it's simply beyond what I'm able to achieve as a "guest contributor" to the plugin. First of all, of course, there's the fact that CSS class names are inherently meaningless, which in turn renders them dramatically unique...so searching for class names is useless unless you can actually look at the class contents. That capability would be a real hassle to build, and even if we magically posit its existence, the only proper targets would be classes where the only property specified is the application of bold weight or italic styling. In my experience, published CSS sheets are seldom so cleanly made; there are huge amounts of over-specification to be found almost everywhere, and the rare exceptions are precious, valuable jewels.

In short, we'd be adding a lot of code for a minimal actual benefit, which really demonstrates the law of diminishing returns in dramatic ways. Such a feature would be more appropriate to an actual ebook editing program than a "simple touch-ups" plugin, and even there it's a perk that should come below many other CSS optimization features. I would truly enjoy having such a program, but I do not know of one in existence.

[JSWolf]

Originally Posted by Rev. Bob

Not something I've touched, so at least it's not something I broke! I'd suggest, right off the bat, making sure the output file isn't write-protected or otherwise a bad destination. If that's good, there may be other problems with the NCX file that changing the PlayOrders is not enough to fix. Perhaps if you could extract and attach that NCX file...?

It isn't something you broke. It's something that's been broken before you modified the plugin to add your nice features. The fact that the plugin says it failed because there's noting updated is 100% wrong. Changing the NCX is changing the ePub and that is something that should be saved.

Any chance you could maybe have a look at it and see where it goes wrong? Or for that matter, can anyone have a go at it?

[Rev. Bob]

Originally Posted by JSWolf

Any chance you could maybe have a look at it and see where it goes wrong? Or for that matter, can anyone have a go at it?

That would be why I suggested attaching the NCX, so I could take a look at it and see if that file shows any other problems. Really, though, I'd suggest contacting the author of the main plugin; although I'm about ready to call my two-function update "gold," he's the one who maintains it and keeps the Official Version.

Plus, y'know, it's his code. He understands it better.

[JSWolf]

Originally Posted by Rev. Bob

That would be why I suggested attaching the NCX, so I could take a look at it and see if that file shows any other problems. Really, though, I'd suggest contacting the author of the main plugin; although I'm about ready to call my two-function update "gold," he's the one who maintains it and keeps the Official Version.

Plus, y'know, it's his code. He understands it better.

But, I don't want the official one fixed without yours also being fixed. And yes, I did PM the author.

[Rev. Bob]

Originally Posted by JSWolf

But, I don't want the official one fixed without yours also being fixed. And yes, I did PM the author.

Okay, I took a look and I think I see the problem. Please try the attached two-routine version; the only change is a one-line attempted fix to the NCX routine. If that works, I'll recommend including it as an official patch.

The plugin generally uses a variable called "Dirtied" to indicate whether changes need to be written, but in this case it uses a dual test. My fix is to add a line at the end of that dual test to force "Dirtied" to be set to true, which I think will fix everything elegantly.

EDIT: The original fix didn't work, apparently because the error was coded in two places and I only found it in one. I've removed the first attempt from here and will post a new message with the new attempt.

[Rev. Bob]

Dirty little bug...

Let's see if this two-line fix does the trick. It looks like it works on this end, but I want outside verification...