Originally Posted by MacZap
I wish to build a TOC (in Mobi or Epub files) by creating a Xpath expression that recognize these tags.
For formats like OpenOffice documents, calibre first converts the file to XHTML. Your XPath expression needs to be in terms of the XHTML calibre generates. To look at that XHTML and figure out the expression you need, create an EPUB file, then unzip it and look at the HTML it contains. (The EPUB container is just a ZIP file, so you can unzip it with the tool of our choice.)