On Friday 23 May 2003 14:12, Ron King wrote:
> There are over 200 .jape files in the distribution, and I haven't
> figured out from the documentation how they are actually invoked! I sort
> of understand the docs, and what you described in your email, but not
> how the JAPE grammars are used in the code. Would I make a new .jape
> file, or add it to an existing one? How can I tell which of the .jape
> files are used by the programs? The docs seem to describe the grammars,
> but not how they fit in to the overall application.
You can do either. It depends if you expect the rule to interact with another
one or not (ie if there's likely to be overlap with another type of entity).
The simplest way would probably be to add the rule to the file final.jape.
If you don't expect any interaction with other rules then it really doesn't
matter where you add it.
You can tell which .jape files are actually used by looking at the file
main.jape, which is the one that's specified on loading the JAPE grammars.
Look at the parameters for the ANNIE NE recognition grammar (think that's
what it's called) when you've loaded it and you'll see.
The main.jape file just lists all the .jape files used.
Hope that helps