Sigil Indexing problem

[mjmuk]

I am in the middle of creating a multi-layer Index for the first time and have run into a problem.

Using the Sigil Index Editor here is the text I have selected to include and the Index Entries

HMS&nbsp;<em>Brighton</em> - Ships/UK
HMS&nbsp;<em>Illustrious</em> - Ships/UK
FF&nbsp;<em>Floréal</em> - Ships/Francs
SPS&nbsp;<em>Santa María</em> - Ships/Spain
SPS&nbsp;<em>Canarias</em> - Ships/Spain

The html code is used to a) keep the prefix and name together, b) <em> for local style.

The resulting rlevant lines in the generated index.html are:

<div class="sgc-index-entry">
      Ships

      <div class="sgc-index-entry">
        UK <a href="../Text/Chapter01.xhtml#sigil_index_id_7">1</a>, <a href="../Text/Chapter01.xhtml#sigil_index_id_20">2</a>
      </div>
    </div>

As can be seen, only the Words UK and Ships are being shown along with their location links. There are no names of ships.

I have other similar lists in my index which one of which is three levels and is working okay.

I've checked the code between them and don't see anything obvious.

What am I missing here? Please tell me.

[meme]

Try a "/" at the end of the entry, e.g. Ships/UK/

Without a slash it thinks you've given it a name to use as is. With a / it knows its a group and should use the name in the index field as the default name.

[mjmuk]

Hi,

Thanks for the reply, I have added a / to the end of the each of the entries, and re-run the index, but now not even the Ships UK word are appearing in the index!!

So have taken out the / I added and rerun the index and still have nothing for this entry!

Is there some database somewhere that has got corrupted?

This is very frustrating.

[meme]

Are you using Mark for Index, or the Index Editor. You can't include html code in the Index Editor as it looks for non-html words.

How are your entries marked in HTML?

[mjmuk]

Hi,

I am using only the Index Editor.

I am selecting the text from the main body of test then pasting it into the index editor.

These are the entries as they appear in the Index Editor

HMS Brighton - Ships/UK
HMS Illustrious - Ships/UK
FF Floréal - Ships/France
SPS Santa María - Ships/Spain
SPS Canarias - Ships/Spain

Regards

Mike

[meme]

If you are using "HMS Brighton" in the Index Editor, then it will look for "HMS Brighton" and not "HMS&nbsp;<em>Brighton</em>". The <em> will invalidate a match.

You can use Mark for Index around the phrase manually each time it occurs in the book and it will work. More exact, but more tedious.

I'm not sure I can adapt the Index Editor to work with included HTML tags. Will give it some thought.

A workaround could be to change every "HMS&nbsp;<em>Brighton</em>" to "HMS Brighton", run the Create Index, then change them back to "HMS&nbsp;<em>Brighton</em>". But that's probably a one shot approach.

[meme]

On second thought try it with the em tags.

[mjmuk]

I may have confused the issue by showing the coding around the Ships name in the original post. I'll try your suggestion and come back to you tomorrow.

regards

Mike

[mjmuk]

Hi, I have carried out some more tests and if I remove the &nbsp; from between the HMS and Ships name in the HTML content, then the Index is created correctly.

[mjmuk]

BTW I have tried using Entity Number * as a substitute for the &nbsp; but that gets stripped out when I run the Create Index, I guess Sigil does not support that method or am I getting too confused!

[exaltedwombat]

There's a strong argument that a simple list, along with the device's Search function is more useful in an eBook than a linked Index.

[mjmuk]

@exaltedwombat Yes that is true. But I am going to use the generated index.html source for some futher processing.

[meme]

@mjmuk

After checking, the issue turns out to be &nbsp; as you worked out, but in a slightly different way.

The way the Indexer works is to collect all the text from a paragraph, for example, and then compare your Index Editor entry against that text. So it only sees and compares against text - not HTML code. Just copy the text as you see it in Book View and paste into the Index Editor. Don't use the string "&nbsp;" or HTML codes "<>" in the Index Editor.

But, unfortunately, the Index Editor converts all &nbsp; characters into spaces. So when it goes to match the space against the &nbsp; character in your text, it fails. I will change the program to automatically convert &nbsp; to a space before comparing to work around this in the next release.

There's another issue I found in testing this - if you use Mark for Index in Code View and select HTML tags (or type tag symbols in the dialog). It doesn't convert the & < > / correctly into the title attribute so it can produce invalid HTML if you select text containing HTML tags to use with Mark for Index, so that will need to be corrected as well.

It's may be that I will have to remove the functionality of using a semicolon ";" to allow multiple entries per line on an Index entry, since its required for HTML entities. Another character isn't as suitable and may have similar issues. But I don't think its a critical feature as you can just put one per line, which is more readable anyway.