Revision: 16271
http://gate.svn.sourceforge.net/gate/?rev=16271&view=rev
Author: ian_roberts
Date: 2012-11-13 17:13:01 +0000 (Tue, 13 Nov 2012)
Log Message:
-----------
JAPE to transform the IOB chunk tags into spanning annotations.
Added Paths:
-----------
gate/trunk/plugins/OpenNLP/resources/tag-to-chunk.jape
Added: gate/trunk/plugins/OpenNLP/resources/tag-to-chunk.jape
===================================================================
--- gate/trunk/plugins/OpenNLP/resources/tag-to-chunk.jape
(rev 0)
+++ gate/trunk/plugins/OpenNLP/resources/tag-to-chunk.jape 2012-11-13
17:13:01 UTC (rev 16271)
@@ -0,0 +1,66 @@
+Imports: {
+ import static gate.Utils.*;
+}
+
+Phase: TagToChunk
+Options: control = once
+Input: Token
+
+// check there is at least one chunk to process
+Rule: TagToChunk
+({Token.chunk != "O"}):tok
+-->
+{
+ try {
+ List<Annotation> tokens = inDocumentOrder(inputAS.get("Token"));
+ String currentChunkType = null;
+ Long chunkStart = null;
+ for(int i = 0; i < tokens.size(); i++) {
+ String chunkTag = (String)tokens.get(i).getFeatures().get("chunk");
+ if(chunkTag == null) {
+ throw new NonFatalJapeException("Found token (" + start(tokens.get(i))
+ + ":" + end(tokens.get(i)) + ") with no chunk tag");
+ }
+
+ if(currentChunkType == null) {
+ // case 1: not currently in a chunk
+ if("O".equals(chunkTag)) {
+ continue;
+ } else if(chunkTag.startsWith("B-")) {
+ chunkStart = start(tokens.get(i));
+ currentChunkType = chunkTag.substring(2);
+ } else {
+ throw new NonFatalJapeException("Found chunk tag " + chunkTag
+ + " outside a chunk");
+ }
+ } else {
+ // case 2: currently looking for the end of a chunk
+ if("O".equals(chunkTag)) {
+ outputAS.add(chunkStart, end(tokens.get(i-1)), currentChunkType,
+ featureMap());
+ currentChunkType = null;
+ chunkStart = null;
+ } else if(chunkTag.startsWith("B-")) {
+ // finish one chunk and start another
+ outputAS.add(chunkStart, end(tokens.get(i-1)), currentChunkType,
+ featureMap());
+ chunkStart = start(tokens.get(i));
+ currentChunkType = chunkTag.substring(2);
+ } else if(chunkTag.equals("I-" + currentChunkType)) {
+ // still inside the current chunk, do nothing
+ } else {
+ throw new NonFatalJapeException("Token (" + start(tokens.get(i))
+ + ":" + end(tokens.get(i)) + ") has invalid chunk tag " +
chunkTag);
+ }
+ }
+ }
+
+ // check whether the last token was part of a chunk, and close it off
+ if(currentChunkType != null) {
+ outputAS.add(chunkStart, end(tokens.get(tokens.size()-1)),
currentChunkType,
+ featureMap());
+ }
+ } catch(InvalidOffsetException e) {
+ throw new JapeException(e);
+ }
+}
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
Monitor your physical, virtual and cloud infrastructure from a single
web console. Get in-depth insight into apps, servers, databases, vmware,
SAP, cloud infrastructure, etc. Download 30-day Free Trial.
Pricing starts from $795 for 25 servers or applications!
http://p.sf.net/sfu/zoho_dev2dev_nov
_______________________________________________
GATE-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gate-cvs
