Table of contents - tweaking

[jackie_w]

Originally Posted by webfolk

The EPUB file required a bit more tweaking, some aesthetic formatting issues: ... reduced the indentation of each paragraph ... The style I had assigned to these in Word didn't carry over. I'm speculating this is because the style name I had used started with the numeral "1".

I agree the smaller text-indent looks better. I'm not totally sure but I think you may be right about problems with named styles beginning with a digit rather than alpha. Best to just avoid using them in future.

Originally Posted by webfolk

I think I've resolved the table of contents by going into the T of C editor in Sigil ...

Hmm...The TOC is no longer nested into 3 levels. Chap 1 is still nested but Chaps 2-23 aren't. Is this really what you wanted?

Also, I suspect when you removed duplicates in Sigil you may have removed the wrong ones. For instance, if you select Chap 5 you now go to the line 'Chapter 5: The Cave of Trees' instead of the image at the top of the page. i.e. the TOC is no longer pointing at the "invisible" <h2> chapter heading at the top of the page but at the "visible" styling <h2> lower down instead. Of course, if you like it better the way it is feel free to ignore my comments

Originally Posted by webfolk

I did this in Sigil and "Saved as" with an alternate file name (preserving the original file unchanged as backup). The new file, however, has a blank page inserted before the book cover. I've noticed this before when I saved from Sigil. How do I get rid of this and, for that matter, how do I prevent it from happening?

I'm afraid I can't help with Sigil as I've never used it. Word/Calibre/Notepad++ text editor is my toolkit. For no reason other than my own gut-feel I'm not comfortable with moving back and forth between Calibre and Sigil because I'd heard they do things differently especially in the TOC area.

It should be easy enough to remove it manually. You need to edit the file called content.opf inside the epub (Tweak-epub in Calibre will let you do it). The bare minimum you need is to go to the spine section and remove the line marked in red below:

<spine toc="ncx">
        <itemref idref="filelist.xml"/>
        <itemref idref="titlepage.xhtml"/>

The file called filelist.xml was contained in your original HTML zip file and I suspect is one of those MS excess baggage files you get when you don't save Word as webpage-filtered. It probably would have been safe to remove it before importing to Calibre for the initial epub conversion. Calibre knows about the file but seems to sensibly exclude it from the epub spine. It would appear that Sigil doesn't ignore it. You could ask for help in the Sigil forum.

[webfolk]

Originally Posted by jackie_w

I agree the smaller text-indent looks better. I'm not totally sure but I think you may be right about problems with named styles beginning with a digit rather than alpha. Best to just avoid using them in future.

I agree.

Originally Posted by jackie_w

Hmm...The TOC is no longer nested into 3 levels. Chap 1 is still nested but Chaps 2-23 aren't. Is this really what you wanted?

I think Chapters 3, 4, 8, & 13 are also 3 levels, each at the beginning of its respective part. What reduced the other chapters to 2 levels was removing the part headings that preceded them in the T of C ... which was done because they seemed redundant and cumbersome. I'm not sure what to do with this, need to play with it more, especially in light of the following:

Originally Posted by jackie_w

Also, I suspect when you removed duplicates in Sigil you may have removed the wrong ones. For instance, if you select Chap 5 you now go to the line 'Chapter 5: The Cave of Trees' instead of the image at the top of the page. i.e. the TOC is no longer pointing at the "invisible" <h2> chapter heading at the top of the page but at the "visible" styling <h2> lower down instead. Of course, if you like it better the way it is feel free to ignore my comments

I tried several different ways of removing headings and this seemed to work best among them. Interestingly, when I open the EPUB in Adobe Digital Editions and the SONY Reader Library (EPUB readers I downloaded onto my Mac), clicking on Chapter 5 and the other chapters which contain illustrations (10, 11, 14, 16, 17, 20, 21) in the table of contents results in the illustration appearing at the top of the page ... part of the reason I thought this change was OK. However, when I do this in the Calibre reader, the results are as you describe. As the Yul Brynner character in the original King and I film would say, "Is a puzzlement!" I'm not sure what to do with this except sleep on it for now.

Originally Posted by jackie_w

I'm afraid I can't help with Sigil as I've never used it. Word/Calibre/Notepad++ text editor is my toolkit. For no reason other than my own gut-feel I'm not comfortable with moving back and forth between Calibre and Sigil because I'd heard they do things differently especially in the TOC area.

I've found Sigil useful, but prefer to keep my use of it to a minimum.

Originally Posted by jackie_w

It should be easy enough to remove it manually. You need to edit the file called content.opf inside the epub (Tweak-epub in Calibre will let you do it).

Done. Thanks!

Originally Posted by jackie_w

The bare minimum you need is to go to the spine section and remove the line marked in red below:

<spine toc="ncx">
        <itemref idref="filelist.xml"/>
        <itemref idref="titlepage.xhtml"/>

I am dazed and amazed at the detailed level of knowledge and understanding this exemplifies. (Perhaps this reveals my own lack of same. ) Nonetheless, as Darth Vader would say, "Impressive."

I will come back to this in the morning, perhaps with a more fresh and insightful perspective. Again, I can't say enough, thank you, Jackie!

[jackie_w]

Hi webfolk,

If you prefer the 2-level TOC (see attached image below), i.e. Parts and Chapters at same level in TOC with Sections still nested under Chapters, then there is a simple solution not involving Sigil.

Import your cleaned up HTML (blank lines removed etc) into Calibre and convert to EPUB using the settings detailed in post #9 step 2, with the following minor changes:

Table of Contents - Level 1 TOC - //*[@class="invisible"]
Table of Contents - Level 2 TOC - //h:h3
Table of Contents - Level 3 TOC - leave empty

Afterwards, you would still need to manually tweak the EPUB's toc.ncx for the 5 'Part' entries, as detailed in post #9 step 3.

[webfolk]

This looks very good, Jackie! I like the way the table of contents looks ... it really does it! I'll get to work on this. If you don't hear from me for awhile, know that I'm busy "having fun".

Originally Posted by jackie_w

Anyway I'm sure you'll have fun playing around ...

[webfolk]

Hi Jackie. Here's what's been happening. I've done all of the following, starting with the ZIP file you attached in Post 9:

Originally Posted by jackie_w

Import your cleaned up HTML (blank lines removed etc) into Calibre and convert to EPUB using the settings detailed in post #9 step 2, with the following minor changes:

Table of Contents - Level 1 TOC - //*[@class="invisible"]
Table of Contents - Level 2 TOC - //h:h3
Table of Contents - Level 3 TOC - leave empty

but I haven't been able to do this:

Originally Posted by jackie_w

Afterwards, you would still need to manually tweak the EPUB's toc.ncx for the 5 'Part' entries, as detailed in post #9 step 3.

because I can't get a workable edited HTML file to convert to EPUB. I did the editing in Kompozer 0.8b3, an open source WYSIWYG HTML editor that I used to create our website. I started with the Post 9 ZIP file & unzipped it with "The Archiver", then did the edits in source code. When I checked the text view in Kompozer, all looked as it should. But when I saved it, some of the punctuation in the resulting html went awry: an apostrophe becomes ’, left quote becomes “, right quote becomes � and so forth.

I posted a query on a Kompozer forum and got several suggestions that didn't work and one that shows promise. Here it is:

The examples you cite are characteristic of character encoding issues. You can change the character encoding on the format>page title and properties window (in Kompozer). Select the character encoding to match your original document. If you look at the html file in a text editor there should be a line in the head section similar to:

Code: Select all
<meta http-equiv="content-type" content="text/html; charset=UTF-8" />

You would be interested in the charset= part and see if it is in the KompoZer list.

However, I'm having a problem viewing the HTML file in the resident Mac text editor, TextEdit (all that can be seen are the illustrations on black background with no visible text) so I'm at an impasse.

Do you know what the character encoding is? Is it derived from the original source file in Word or does Calibre alter it? Do you have any other insight or approach?

One possibility is to avoid working in HTML and convert the Post 9 Zip to EPUB using the parameters you stated in Posts 9 & 18, then do the cleanup in Sigil ... which is the way I created the EPUB attached in Post 15. I share your preference to avoid Sigil, but the only known glitch was the blank page inserted before the book cover, which you resolved.

I do hope these complications aren't vexing to you.

[jackie_w]

Oh dear...

Firstly, let me say I am definitely not vexed

Unfortunately I'm struggling with my zero-knowledge of Macs and anything to do with Kompozer. However, I'm pretty sure the suggestion about 'encoding problem' is correct. Sadly, my understanding of encoding is also fairly sketchy.

I had a look at the original HTML you supplied. The first line is

<?xml version='1.0' encoding='utf-8'?>

which suggests UTF-8. My text editor describes it as "ANSI as UTF-8" I'm not sure if/how this differs from ordinary UTF-8. I didn't change the encoding in any of the previous posts.

I believe that when you drag a raw HTML file into Calibre it zips it up and converts the encoding to UTF-8 from whichever encoding it currently is. Calibre tries to figure out the source encoding. I'm not sure what happens if you drag in pre-zipped HTML.

When I looked more closely at your original HTML in a text editor, I also noticed that the line-endings are Unix-style, i.e. LF only. My editor tells me that Mac-style should be CR only. So, clutching at straws, I've edited the raw HTML to have Mac-style. I've also changed the encoding so the editor just says UTF-8 and, belt & braces, I've added an extra line 4

<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />

so this HTML should, no ifs no buts, be UTF-8.

I've dragged it into Calibre and it still converts to EPUB OK on my Windows PC.

The attached ZIP contains the raw HTML plus the images. Everything else (including the rogue filelist.xml) has been removed. It is pre-Calibre. Try opening this HTML in your text editor. It obviously still needs your edits.

I have to say that this statement:

...However, I'm having a problem viewing the HTML file in the resident Mac text editor, TextEdit (all that can be seen are the illustrations on black background with no visible text) so I'm at an impasse...

has totally confused me. When I open an HTML file in a text editor I cannot see any images only the html code for the images

[webfolk]

Originally Posted by jackie_w

Firstly, let me say I am definitely not vexed

I'm relieved ... and thanks for saying this right up front!

The following is utterly irrelevant:

Originally Posted by jackie_w

Unfortunately I'm struggling with my zero-knowledge of Macs and anything to do with Kompozer. ... Sadly, my understanding of encoding is also fairly sketchy.

because what you did was spot on!

I immediately expanded and tried to open the ZIP you attached in TextEdit and the same thing happened: images only, black background, no visible text. Playing with it in different ways went nowhere. I strongly feel this is a Mac problem and was about to call Apple when I was interrupted by some incoming phone calls. Later, after having let it go for awhile, the thought came to just go ahead and open the HTML in Kompozer and--voila!--it worked. The first thing I did in Kompozer was make a minor change, save the file, then open it in my browser and all the punctuation anomalies were gone. So I went on to make all the changes, import it into Calibre and convert it to an EPUB with all the settings you recommended. And it looks just fine.

The last thing is to implement step 3 of post 9

Originally Posted by jackie_w

... you would still need to manually tweak the EPUB's toc.ncx for the 5 'Part' entries, as detailed in post #9 step 3.

I found and opened the toc.ncx in TextEdit but I don't know how to do this:

Originally Posted by jackie_w

I rectified this by using Calibre's EPUB-tweak feature to manually edit the toc.ncx file in a text editor. Only 5 small changes required to correct the destination of each of the Level 1 items (or Parts in your terms) so it points to the first Chapter in the Part, i.e. the same piece of HTML as the next item down in the toc.ncx file.

because I don't understand the toc.ncx language well enough to know what's going on in there. What are the 5 changes?


The rest of your reply, about encoding, is a bit over my head at its present elevation and, quite frankly, I'm willing to put understanding on the back burner for a time, being a bit head-weary. That it works, however, is invigorating! What I do understand (after a quick Google) is "belts and braces." Never heard it before ... really like it!

[jackie_w]

The current problem is that, in the Viewer, each TOC "Part" is pointing to the wrong place. Each "Part" should be pointing to the same place that the first "Chapter" following the "Part" points i.e.
Part 1 to Chapter 1
Part 2 to Chapter 3
Part 3 to Chapter 4
Part 4 to Chapter 8
Part 5 to Chapter 13
So we need to use Calibre Tweak-epub to tweak toc.ncx, in your text editor, so that it does.

Here is the relevant section for "Part 1" in the toc.ncx in my version of the epub:

    <navPoint id="e2c7f948-bfe9-41ab-96e3-b664d1569949" playOrder="2">
      <navLabel>
        <text>Part 1 ~ Earth Time</text>
      </navLabel>
      <content src="SMem2utf8_split_002.htm#calibre_toc_2"/>
    </navPoint>
    <navPoint id="7ee7a724-dd85-4146-951f-235ca86c50da" playOrder="3">
      <navLabel>
        <text>Chapter 1: The Awakening</text>
      </navLabel>
      <content src="SMem2utf8_split_003.htm"/>
      <navPoint id="20319fe0-acfe-49bb-9a46-53231822216a" playOrder="4">
        <navLabel>
          <text>A Gift for Someone…</text>
        </navLabel>
        <content src="SMem2utf8_split_003.htm#calibre_toc_30"/>
      </navPoint>

What you need to do to correct "Part 1" is to change the red bit to match the blue bit.

... and similarly for "Part 2":

    <navPoint id="db1618d1-bc2e-412c-ac9e-e5e99bc7dbd8" playOrder="21">
      <navLabel>
        <text>Part 2 ~ Day One</text>
      </navLabel>
      <content src="SMem2utf8_split_004.htm#calibre_toc_5"/>
    </navPoint>
    <navPoint id="1cb9e2c9-392e-46df-aa07-d370c3e7ecf7" playOrder="22">
      <navLabel>
        <text>Chapter 3: The Prophecy</text>
      </navLabel>
      <content src="SMem2utf8_split_005.htm"/>

Then repeat for Parts 3, 4 and 5.

When you've finished, save toc.ncx.
Close the text editor.
Close the file browser window (listing all the files inside the epub).
Click on "Rebuild epub".
View epub in Calibre Viewer and click on each "Part" in the TOC to make sure it now points to the right place.
...and that should be it...

By the way I've opened a new ticket in Calibre Bugs. If the powers-that-be agree that it is a bug then it should get fixed so that no toc.ncx tweaking would be necessary. Sometimes these things get fixed very quickly, occasionally they don't, so don't hold your breath.

Good Luck

[webfolk]

Almost!!! I made all the changes you specified in Post 23, but it's still slightly off ... as follows:

• TOC Part 1 and TOC Chapter 1 point to the Foreword
• TOC Part 2 and TOC Chapter 3 point to Chapter 2
• TOC Part 3 and TOC Chapter 4 point to Chapter 3
• TOC Part 4 and TOC Chapter 8 point to Chapter 7
• TOC Part 5 and TOC Chapter 13 point to Chapter 12

All the other items in the TOC are fine. I suspect it's just a matter of changing some numbers (just off by 1 digit) ... and I tried adjusting toc.ncx a few different ways, but haven't gotten it right. I'll attach the current EPUB which which reflects the changes from Post 23 ... and not my attempts to correct it.

Originally Posted by jackie_w

By the way I've opened a new ticket in Calibre Bugs.

Good idea! I read it. Thanks for doing that.

[jackie_w]

I think you've changed the toc.ncx items the wrong way round, i.e. you've changed blue to match red instead of red to match blue. If it's easier to remember change the upper <content> item to match the lower <content> item and leave the lower item alone.

Taking 'Part 1' as an example, after the edit both the 'Part 1' and 'Chapter 1' should look like

<content src="SMem2utf8_split_003.htm"/>

You will need to re-do the ZIP to EPUB conversion first, so you've got a fresh toc.ncx.

[webfolk]

Originally Posted by jackie_w

I think you've changed the toc.ncx items the wrong way round ...

Right! Well! Ahem ... That would seem to explain things. And having done as you suggested seems to bear this out: the file now looks perfect!

I took the further step of converting the EPUB to MOBI and that also looks perfect when viewed in the Kindle reader I have on my Mac. It looks like Sacred Memories is ready to go up on the website as an ebook.

I cannot thank you enough for all your help, Jackie. Your explanations were precise, clear, remarkably well organized and easy to follow ... including (and especially) Post 23:

Originally Posted by jackie_w

What you need to do ... is to change the red bit to match the blue bit.

And thank you for your diligence and incapacity for vexedness!

[jackie_w]

You're welcome I hope you've learned a few tricks for the rest of your project.