Hi, there is a nice, although slightly dated, Lucene-based index for parse trees called Fangorn [1]. That's probably what you are referring to as Steven Bird's work.
I started playing around with that once, but didn't quite finish it. What's there is a CAS consumer (DKPro writer) and a test case which illustrates how to query the Fangorn index produced by the writer. Both are available from the DKPro Core ASL SVN repository [2]. [1] https://code.google.com/p/fangorn/ [2] https://dkpro-core-asl.googlecode.com/svn/de.tudarmstadt.ukp.dkpro.core-asl/trunk/de.tudarmstadt.ukp.dkpro.core.io.fangorn-asl Am 11.06.2013 um 16:44 schrieb Georg Fette <[email protected]>: > Hello, > I am looking for information how to use the UIMA Lucene indexer to index a > dependency parsed corpus. I read the documentation for the Lucene index on > the UIMA sandbox site but I still do not really feel ready to start > implementing something with it. Is there a link to a good hands-on-tutorial > about how to use the index to store tree structures like dependency trees out > there ? I have found the work of Steven Bird who uses a Lucene to index parse > trees but I do not really get it how he queries for structure information > instead of the basic "temporal" order in the token streams. > I would be grateful for useful tutorials. > Or can anyone give some good advice on alternative approaches for indexing > structure information like dependency parses ? I thought about using a Neo4J > graph database. > Greetings > Georg
