How to influence the ToC when creating an ebook

[Skeeve]

Hi!

I'm compiling an ebook from 26 short stories which appeared in a german magazin.

My HTML files contain headlines similar to this:

<div class="headline">
<h1 id="u094aaa39-0c47-43d1-9343-d5190bf2d74d">Unter allen Wipfeln (1)</h1>
<address class="author">Arno Endler</address>
<address class="published">c't 2014, Heft 1, Seite 192</address>
</div>

When I let calibre (1.48) generate a ToC I get lines like this:

<li><a href="ct-2014-01.html#u094aaa39-0c47-43d1-9343-d5190bf2d74d">Unter allen Wipfeln (1)</a></li>

But I'd like to have the author added like so:

<li><a href="ct-2014-01.html#u094aaa39-0c47-43d1-9343-d5190bf2d74d">Arno Endler: Unter allen Wipfeln (1)</a></li>

Is there any way to achieve that?

I already found the ToC generation using XPath, but this did not seem to help as it only allows to select elements and each selected element's text content seems to become an ToC entry. It seems to me there's no way to pick more than one element and somehow style them to an entry.

Am I right or do you have any proposal what I can do?

[thewitt]

Edit the TOC after you create the book. Though I don't use Calibre, I use Sigil for this and it works beautifully.

[theducks]

Use some fancy Regex (not shown) on the source lines in the book to add a:
title="\3: \2 " inside the H1 tag

fwiw
\3 is the author captured
\2 is the title captered

title= "" overrides the normal H# text choice for the TOC generator

[Skeeve]

Thanks for your replies.

So it seems what I wanted to do isn't possible.

Sigil I can't use (at the moment) as my OS X is too old and it won't work here.

Regex to set the title attribute could be an idea - maybe I'll take that approach. Thanks for that hint!

Update: Tried it by manually setting the title attribute in one of the stories: No success. The title attribute wasn't taken by Calibre 1.48's automatic ToC creation. Maybe my Calibre is too old (same issue as with sigil: My OS X is too old)

[Ghitulescu]

You can always edit the TOC by hand to point to the exact targets.

[theducks]

@Skeeve
Seems to work in Sigil but not in Calibre TOC generator (2.19 W7x64)

the title =solution does work if you get it correct

title = must be part of the H tag

[Skeeve]

Originally Posted by Ghitulescu

You can always edit the TOC by hand to point to the exact targets.

Sure I can. But I'm asking for an automated solution.

[BetterRed]

Originally Posted by theducks

Use some fancy Regex (not shown) on the source lines in the book to add a:
title="\3: \2 " inside the H1 tag

fwiw
\3 is the author captured
\2 is the title captered

title= "" overrides the normal H# text choice for the TOC generator

Originally Posted by Skeeve

Thanks for your replies.

So it seems what I wanted to do isn't possible.

Sigil I can't use (at the moment) as my OS X is too old and it won't work here.

Regex to set the title attribute could be an idea - maybe I'll take that approach. Thanks for that hint!

Update: Tried it by manually setting the title attribute in one of the stories: No success. The title attribute wasn't taken by Calibre 1.48's automatic ToC creation. Maybe my Calibre is too old (same issue as with sigil: My OS X is too old)

Originally Posted by theducks

@Skeeve
Seems to work in Sigil but not in Calibre TOC generator (2.19 W7x64)

the title =solution does work if you get it correct

title = must be part of the H tag

I'm wondering if the Function Mode for Search & Replace in the Editor feature could be used to inject metadata element values into the Headers. If not then maybe the editor could be enhanced to make them available.

BR

[Skeeve]

Originally Posted by BetterRed

I'm wondering if the Function Mode for Search & Replace in the Editor feature could be used

I think, that is really a good idea! Thanks a lot!

[Skeeve]

Thanks @BetterRed!

I found a solution that works for me!

First I create the ToC based on XPath.

Level 1 is //h:h1
Level 2 is //h:address[@class="author"]

This will give me a toc.ncx which I open in the editor.

On that I apply this Search:

(<text>)(.*?</text>\s*</navLabel>\s*<content src=".*?)#.*?("/>)\s*<navPoint.*?>\s*<navLabel>\s*<text>(.*?)</text>\s*</navLabel>\s*<conten.*?/>\s*</navPoint>

And apply this replace:

\1\4: \2\3

After that I create the Inline ToC.

[eschwartz]

Originally Posted by BetterRed

I'm wondering if the Function Mode for Search & Replace in the Editor feature could be used to inject metadata element values into the Headers. If not then maybe the editor could be enhanced to make them available.

BR

Function mode was introduced in calibre 2.11 -- not ideal for those stuck on Qt4 versions.

[BetterRed]

Originally Posted by Skeeve

Thanks @BetterRed!

I found a solution that works for me!

Great, its always satisfying to arrive at a destination solution under one's own steam.

Originally Posted by eschwartz

Function mode was introduced in calibre 2.11 -- not ideal for those stuck on Qt4 versions.

I missed the reference to 1.48

BR

[Ghitulescu]

Originally Posted by Skeeve

Sure I can. But I'm asking for an automated solution.

I meant after. Don't tell me you're trusting automations?

[Skeeve]

Originally Posted by Ghitulescu

I meant after. Don't tell me you're trusting automations?

I do trust it. Why not?

I already convert the PDFs, which are the source of this book, using automated scripts I wrote myself. They are not perfect, but I trust that they don't mess everything up. They especially are able to create the meta-info properly. So I also trust my found half-automated way of creating the ToC.

But still I would like to get proof-readers. Unfortunately it's german text.