Quote:
Originally Posted by cybmole
I have sound some books in a series where the above structure detection appraoch is not working. chapters are not being detected. ...
I am going from mobi to epub and in the epub I see
Code:
<p class="calibre9"><span class="calibre3"><span class="bold">Chapter 1</span></span></p><br class="calibre1" />
...
<p class="calibre9"><span class="calibre3"><span class="bold">Chapter 2</span></span></p><br class="calibre1" />
I will now try to put them into the bug tracker system for you, please confirm that I've have done that correctly.
UPDATE - seems to be a case of preprocess not running on mobi source, despite having reset convert options to default.I did the extra step of copy the epub created by the failed attempt, renamed it as zip, add back to library, reconverted - then I do get chapters found, files split etc..
that's still pretty slick when I think of how long I spent fixing vol 1 & 2 of this series by hand ! but check it out please - should the preprocess have worked on the .mobi source ? |
Mobi wasn't fully integrated into preprocess, as I didn't really have any mobi sources that required this sort of processing when I implemented the feature. The work that's underway now will have all file formats going through preprocess if it's enabled.
Regarding 'making things worse', the feature just uses common patterns. False positives are possible, though a lot of logic exists to prevent this, they can't be completely eliminated. It also removes all non-breaking spaces from the document to facilitate processing - if there were some in place for formatting reasons they'll be eliminated.