Help with XPath Expression

[a.o.advena]

I'm trying to convert a book from ODT to EPUB, but I'm having problems with the XPath expression for chapter breaks. My document has specific custom formatting at each break, which (using the suggestion from the "Chapter Detection/Table of Contents Tutorial" sticky) I determined to be:

<p class="P-Section_20_Number">

AND

<p class="P-Chapter_20_Number c_odt2">

I tried a few different things from the Calibre XPath Tutorial (https://manual.calibre-ebook.com/xpa...xpath-tutorial) but nothing seems to be working.

Would appreciate any help.

[kovidgoyal]

//h:p[@class="P-Section_20_Number]|//h:p[@class="P-Chapter_20_Number c_odt2"]

[a.o.advena]

Thanks for your response. I just tried to use this (pasted the line above into "Detect chapters as (XPath expression):" but Calibre says that the expression is invalid.

In case I wasn't clear above, I want it to break at either one of the above classes.

[davidfor]

Originally Posted by a.o.advena

Thanks for your response. I just tried to use this (pasted the line above into "Detect chapters as (XPath expression):" but Calibre says that the expression is invalid.

In case I wasn't clear above, I want it to break at either one of the above classes.

There is a double quote missing before the first close square bracket. It should be:

//h:p[@class="P-Section_20_Number"]|//h:p[@class="P-Chapter_20_Number c_odt2"]

[a.o.advena]

Thanks for catching the error. I tried this new expression, but no luck. I still get all the section & chapters in one big xml file.

For kicks, I did a little more experimentation and tried some other combos based on the manual and wizard

//h:p[@class="P-Chapter-Number c_odt0"]

//*[@class="P-Chapter-Number c_odt0"]

//*[re:test(@class, "P-Chapter-Number c_odt0", "i")]

all with the same result.

[davidfor]

Is it the same class for all the paragraphs you want to split on? Or does the number in "P-Chapter_20_Number" change? Is that "20" actually the chapter number? If that is the case, the your last test is getting close. I think the following should do it:

//*[re:test(@class, "P-Chapter", "i")]

And as you had to different classes in the first post, it would be:

//*[re:test(@class, "P-Chapter|P-Section", "i")]

If that doesn't work, and my spelling is correct, I think we need to see a sample of the book to get an idea of what to do.

[a.o.advena]

Originally Posted by davidfor

Is it the same class for all the paragraphs you want to split on? Or does the number in "P-Chapter_20_Number" change? Is that "20" actually the chapter number?

The classes are all the same. The "_20_" is what Calibre or LibreOffice (not sure which) substitutes for the space that is in the style's name.

Originally Posted by davidfor

I think the following should do it:

//*[re:test(@class, "P-Chapter", "i")]

And as you had to different classes in the first post, it would be:

//*[re:test(@class, "P-Chapter|P-Section", "i")]

Yes, that's the ticket. Perticularly the second one as it captures both the chapter and section breaks. (For others who may use this later - I still had to do a little manual tweaking of the TOC, but this took care of breaking everything up into separate files that could then be arraigned in the TOC.)

Thanks for your help.