FilePhraser
-----------

This distribution contains the FilePhraser, a component that recognises
phrases in a file.  It combines distribution 008 (Phraser) with all
necessary components to operate directly on text files.  

Usage: construct a FilePhraser with the name of the file you want to have
processed, register any WordEvent listeners using the addWordListener()
method, and kick off the whole process by calling the FilePhraser's
start() method.

The registered listener(s) will receive each word as it is read, including
any mark-up that signifies phrase boundaries.  You can extract the data
from the event by calling its getValue() method with the parameters
"token" and "pos"; this will return null for "token" when the end of the
input file has been reached.  Markup (such as sentence boundary markers)
will also be returned; these are normally enclosed in pointed brackets
(eg <s> for sentence boundaries, and <phrase type="XXX"> for phrases).

This distribution contains all the required classes, including those which
tie together the separate components.  You can also run the FilePhraser
as a stand-alone component.  Here you will need to supply the filenames of
the files to be processed as command-line parameters, and each token will
be printed on a separate line to System.out with any associated tag.


(C) 2000 Phrasys (www.phrasys.com / www.phrasys.co.uk)

The Phraser listens to incoming WordEvents, and adds mark-up tags
around any phrases it can find.  These tags looke like <phrase type="NP">
or similar, depending on what phrases are to be recognised.  The mark-up
creates extra WordEvents which are added to the stream of incoming
events.

Usage: construct a Phraser, and register it to a component that produces
WordEvents (such as the Tokeniser or the tagger).  If you are using part
of speech tags in the resource file, you will have to use a tagger as
source in order to have access to the "pos" value of the incoming tokens.
If you want to use a different resource file from the one provided you
can also specify which one you want to use.

Then register with the Phraser a component that will listen to WordEvents
containing phrase mark-up.  The end of the input will be marked through
an event which has null as its "token" value.

For more information please consult the javadoc generated documentation on
the website.


(C) 2000 Phrasys (www.phrasys.com / www.phrasys.co.uk)