All notes of ebook mistakenly link to the book's title page. What's best way to fix?

[Tex2002ans]

Originally Posted by thymesnewroman

I didn't use Calibre at all. [...] Perhaps the original person who created this epub took a "wrong turn" with the notes/linking/coding?

Yes, because index_split_000 is one of the ways Calibre splits files... so it's a pretty sure sign this book was run through a Calibre conversion at some point.

* * *

What's the name of this book?

Perhaps there's a better quality file floating around.

Many times it's better to work from a much better source than to try to clean up a poorly-converted mess.

Originally Posted by exaltedwombat

It's going to be easier to clear this up if you make the code of an entire chapter available to us.

While this might be better... from the samples given, you can already tell the person who initially converted this botched it up so badly.

Look closely at the Footnote code from Example #1:

(I'm going to rip out the class="" crap so we can see

<a href="../Text/index_split_000.xhtml#anchor22">14</a>

[...]

<p><span id="anchor129"></span><a href="../Text/index_split_000.xhtml#anchor60">14</a>

The footnote 14 number says jump to anchor22, but the end of the chapter says anchor129...

I'm assuming the entire book's IDs are completely mangled as well. (They don't match up in any of the examples you gave.)

Originally Posted by thymesnewroman

index_split_001.xhtml is the Table of contents.
index_split_003.xhtml is Preface.
[...]
And so on and so forth.

If you do decide to continue trying to recover this mess of a file...

It may be easier to rename all the files to human-readable names.

In Sigil, on the left-side, you'll see the Book Browser (where it lists all the files). You can Right-Click files, then Rename.

So then you could rename:

index_split_003.xhtml -> Preface.xhtml
index_split_005.xhtml -> Part1.xhtml
[...]

That will at least make the spaghetti of links more readable. Then you would be able to more easily tell:

"Whoops, this Chapter 1 footnote actually points to titlepage.xhtml... no wonder it's broken."

And while this ebook may be recoverable... perhaps this cleanup project may be too advanced for someone who doesn't know the technical innards yet.

[JSWolf]

This mess (to me) is not worth the effort. It would be a heck of a lot easier to get a tablet and read the PDF that way.

[thymesnewroman]

Sorry for my delay in getting back to you. I thought I was getting instant email notifications when there are new responses to this thread.

Originally Posted by Tex2002ans

Yes, because index_split_000 is one of the ways Calibre splits files... so it's a pretty sure sign this book was run through a Calibre conversion at some point.

Yes, you're right. When I view the code in Sigil, I see the word "Calibre" in many places.
* * *

Originally Posted by Tex2002ans

What's the name of this book?

Reformed Dogmatics by Herman Bavinck. (This is the four-volume work, not the "Concise Reformed Dogmatics", which is a one-volume "summary").

[thymesnewroman]

Originally Posted by exaltedwombat

It's going to be easier to clear this up if you make the code of an entire chapter available to us.

Good call.
I've uploaded chapter one (aka index_split_006.xhtml). I had to add an acceptable suffix of "zip" for the forum to allow the upload. When you download it simply remove the ".zip" part . The file really ends in ".xhtml". Again, please know this isn't really a zipped/archive file.

[thymesnewroman]

Originally Posted by thymesnewroman

3. I might be able to get my hands on the PDF from an educational institution I'm associated with.

I just spoke with the librarian. Unfortunately, they don't have an electronic copy of the book (whether epub or pdf or whatever). Just the print copy.

If it's going to be too hard to fix this mangled epub, I guess I can just ignore the superscripts and just read the notes at the end of the chapter. I'll try to remember to not tap on the links!

[exaltedwombat]

So the main problem is that in the footnote link "anchor9" should read "anchor116", "anchor10" should be "anchor117" etc. We need to locate each footnote reference and add 107 to the numeric part.

Sounds trivial. But, unfortunately, RegEx doesn't do arithmetic, at least not in any straightforward way.

An interesting challenge!

[thymesnewroman]

Hi wombat,

When you say "main problem", am I understanding you correctly to believe that solving this "main problem" doesn't completely fix the links?

[BetterRed]

Originally Posted by exaltedwombat

So the main problem is that in the footnote link "anchor9" should read "anchor116", "anchor10" should be "anchor117" etc. We need to locate each footnote reference and add 107 to the numeric part.

Sounds trivial. But, unfortunately, RegEx doesn't do arithmetic, at least not in any straightforward way.

An interesting challenge!

Try Function mode for Search & replace in the calibre Editor

BR

[exaltedwombat]

Well, the reference to split_000 would have to go too. But that's easy. I was looking for PROBLEMS!

It would be neat to just do the arithmetic. And @BetterRed's suggestion looks very promising. But there's another approach. The footnote numbering - the 1, 2 etc. visible to the reader - is correct, and is consistently styled. It should be quite possible to use RegEx to strip all the existing href stuff then to create new links deriving anchor names from the footnote numbers.

[doubleshuffle]

In the time everybody here has invested in thinking about possible solutions, the OP could have fixed all the links manually. Just sayin' (he said and headed over to the unendurable phrases thread...

[exaltedwombat]

Not really. Partly because there seem to be a LOT of them, partly because he didn't know how. Though I take your point in general, no point in developing automation for a one-off problem that COULD be fixed manually. Unless you have an enquiring mind and enjoy problem-solving. (And if you don't, I pity you!)

[doubleshuffle]

You're absolutely right. I wasn't being completely serious. But in this case, I had a look at the mess, and when I saw that the anchors didn't even match, I thought it was a lost case for automation.

But kudos if you manage to come up with a solution!

[thymesnewroman]

Originally Posted by BetterRed

Try Function mode for Search & replace in the calibre Editor

BR

Hi BetterRed.

Thanks for your reply.
After turning on Function Mode, what do I type in the Search Field and the Replace field?

[thymesnewroman]

Originally Posted by doubleshuffle

You're absolutely right. I wasn't being completely serious. But in this case, I had a look at the mess, and when I saw that the anchors didn't even match, I thought it was a lost case for automation.

But kudos if you manage to come up with a solution!

Doubleshuffle,
Are you saying that using the Function mode in Calbre editor isn't enough to solve this problem?

[Tex2002ans]

Originally Posted by thymesnewroman

I've uploaded chapter one (aka index_split_006.xhtml).

Thanks for the attachment.

I say go full nuclear. Wipe out all the crap links, substitute with good ones.

(I tested this on your "index_split_006.xhtml" and it worked perfectly on all footnotes in that chapter.)

Full Nuclear Option (For This Book)

1. In Sigil, press Ctrl+F to pop up the Find/Replace.

Make sure you set Mode: Regex.



2. Once you have it set to Regex...

These will wipe out all the busted links and substitute them with the number in brackets:

Search: <a class="[^"]+" href="[^"]+"><span class="[^"]+">(\d+)</span></a>
Replace: <a href="#fn\1" id="ft\1">[\1]</a>

and then these 3 will wipe out the broken anchors and fix the Footnotes:

Search: (<p class="[^"]+">)<span class="[^"]+" id="[^"]+"></span>
Replace: \1

Search: (<p class="[^"]+">)<a class="[^"]+" href="[^"]+"><span class="[^"]+"></span></a>
Replace: \1

Search: (<p class="[^"]+">)<a href="#fn[0-9]+" id="ft[0-9]+">\[([0-9]+)\]</a>
Replace: \1<a href="#ft\2" id="fn\2">[\2]</a>

(Completely Optional) If you want to throw away the other anchor code:

Search: <a class="[^"]+" href="[^"]+"><span class="[^"]+"></span></a>
Replace: ***[Blank] PUT NOTHING***

3. After you're done, make sure to go back to Case Sensitive Mode.