I actually had the same issue as you. I did the following:
Use a program called PDFInfo to remove the metadata (you may need to use a scripting tool like AutoIT to automate it) then I use the following regular expression:
Edit: (?P<authors>[^-]+)\s*-+\s*((?P<series>[^-]+)\s*-+\s*)*(?P<title>.+) works without the error of the series_index not being there all the time
It has support for the Authors name, followed by the - separator then an optional Series name & optional Series_Index, followed by an optional separator (tied to the Series option) and finally the Title.
I am trying to work out why sometimes it fails when there is no series_index, but I suspect that once you define a ?P<id> it MUST be used and not be make optional using a *. I may force it picking up a space (\s), unfortunately not all filenames contain a space next to the separator (-) and I much prefer using an optional space between any of the fields (\s*)
If this doesn't end up working, I may end up using Pyton/Perl/VBScript to split the file name up and use PDFInfo to populate the Metadata fields instead, unfortunately there is no MetaData field for Series or Series_Index (at least using PDFInfo) so I may need to find another tool that allows me to add fields and hope Calibre can map these fields to Series & Series_Index.
P.S. I realise you may not be using PDF's, but thats all that applicable for me, but the regular expression may still end up working with a bit of tweaking.
Last edited by Alby; 06-29-2008 at 10:02 PM.
Reason: Typos & Extra Information