Re: [xsl] citation processing

["Andrew Welch" <andrew.j.welch@xxxxxxxxx>]

On 10/20/06, David Carlisle <davidc@xxxxxxxxx> wrote:
> > If you think its not really feasible to parse a plain text citation
> > into a marked up version then that's good feedback
>
> well never say never of course and it depends if the input always
> follows the rules.
>
> You could parse the example you gave with a couple (or half a dozen:-)
> incantations with xsl:analyze-string,
>
>
> get up to the first "." as list of authors
> recursively spit that up on , to get each author,
> etc
>
> trouble is if the citations have been entered by hand  some of them are
> going to use a : where you expect a , or a . instead of ;, or microsoft
> code page "smart quote" characters instead of the real thing.
> and a simple regexp replace mechanism isn't usually very good at
> recovering from variable input like that.
>
> If on the other hand the text was originally in a citation system and
> was generated but has been cut and pasted across a few generations of
> html pages and the original source is no longer available, you mightbe
> OK to assume that the text itself is regular in its use of punctuation
> and
> ([^\.]*).([^\.]*).([^;]*);(.*)
> will for example give you the author list, the title, the journal title,
> and volumepage info as $1 ..$4 each of which could be further split up.

Yes that's the plan - I've been told there are four variations (so
far) on format, so test each string against each variation and mark it
up if one matches - any that drop out the bottom can be done by hand
(by someone else!) or if that list gets too large try and infer some
more rules.  If they are based on based on a known format then it
should be simple enough to work backwards.