Revision: 16865
http://sourceforge.net/p/gate/code/16865
Author: ian_roberts
Date: 2013-09-09 15:19:40 +0000 (Mon, 09 Sep 2013)
Log Message:
-----------
The mapsList is read-only once initialized, so can be made @Sharable for
efficient duplication.
Modified Paths:
--------------
gate/trunk/plugins/Lang_Russian/src/com/ontotext/russie/gazetteer/InflectionalGazetteer.java
Modified:
gate/trunk/plugins/Lang_Russian/src/com/ontotext/russie/gazetteer/InflectionalGazetteer.java
===================================================================
---
gate/trunk/plugins/Lang_Russian/src/com/ontotext/russie/gazetteer/InflectionalGazetteer.java
2013-09-09 12:50:13 UTC (rev 16864)
+++
gate/trunk/plugins/Lang_Russian/src/com/ontotext/russie/gazetteer/InflectionalGazetteer.java
2013-09-09 15:19:40 UTC (rev 16865)
@@ -29,6 +29,7 @@
import gate.creole.metadata.CreoleResource;
import gate.creole.metadata.Optional;
import gate.creole.metadata.RunTime;
+import gate.creole.metadata.Sharable;
import gate.util.InvalidOffsetException;
import gate.util.LuckyException;
import gate.util.profile.Profiler;
@@ -99,7 +100,7 @@
* etc. each map's value might be an ArrayList of TYpe objects specifying
* categories tied to this word/phrase.
*/
- protected List<Map> mapsList = new ArrayList<Map>(10);
+ protected List<Map> mapsList;
/**
* size of the mapsList
@@ -141,97 +142,112 @@
* @return returns this resource
*/
public Resource init() throws ResourceInstantiationException {
- Profiler profiler = null;
- if(DEBUG) {
- profiler = new Profiler();
- profiler.enableGCCalling(false);
- profiler.initRun("Inflectional Gazetteer init()");
- profiler.checkPoint("reset");
- }
- // check if there's a list URL
- if(config == null) { throw new ResourceInstantiationException(
- "No config provided for gazetteer!"); } // if
+ if(mapsList != null) {
+ // this is a duplicate
+ mapsListSize = mapsList.size();
+ } else {
+ Profiler profiler = null;
+ if(DEBUG) {
+ profiler = new Profiler();
+ profiler.enableGCCalling(false);
+ profiler.initRun("Inflectional Gazetteer init()");
+ profiler.checkPoint("reset");
+ }
+ mapsList = new ArrayList<Map>(10);
+ // check if there's a list URL
+ if(config == null) { throw new ResourceInstantiationException(
+ "No config provided for gazetteer!"); } // if
- try {
- BufferedReader configReader =
- new BufferedReader(new InputStreamReader(config.openStream(),
encoding));
- int lemmaIdx = 0;
try {
- // each line in config is a relative path to an .infl file
+ BufferedReader configReader =
+ new BufferedReader(new InputStreamReader(config.openStream(),
encoding));
+ int lemmaIdx = 0;
+ try {
+ // each line in config is a relative path to an .infl file
- String configLine = null;
- while((configLine = configReader.readLine()) != null) {
- configLine = configLine.trim();
- // ignore blank lines and comments
- if("".equals(configLine) || configLine.startsWith("#")) continue;
+ String configLine = null;
+ while((configLine = configReader.readLine()) != null) {
+ configLine = configLine.trim();
+ // ignore blank lines and comments
+ if("".equals(configLine) || configLine.startsWith("#")) continue;
- URL fileURL = new URL(config, configLine);
+ URL fileURL = new URL(config, configLine);
- InflectionalGazetteerXMLReader gazReader = getGazReader(fileURL);
- List<Lemma> lemmas = gazReader.getLemmas();
- Iterator<Lemma> lemmaIter = lemmas.iterator();
+ InflectionalGazetteerXMLReader gazReader = getGazReader(fileURL);
+ List<Lemma> lemmas = gazReader.getLemmas();
+ Iterator<Lemma> lemmaIter = lemmas.iterator();
- int linesCnt = lemmas.size();
- // allocate the hashmap for the first words from the phrases
- mapsList.add(new HashMap(linesCnt));
- mapsListSize = mapsList.size();
+ int linesCnt = lemmas.size();
+ // allocate the hashmap for the first words from the phrases
+ mapsList.add(new HashMap(linesCnt));
+ mapsListSize = mapsList.size();
- // allocate the category Map with optimal initial capacity & load
- // factor
- msTypeSet = new HashSet<SuffixNest>();
+ // allocate the category Map with optimal initial capacity & load
+ // factor
+ msTypeSet = new HashSet<SuffixNest>();
- Lemma lemma;
- fireStatusChanged(READING + configLine);
- while(lemmaIter.hasNext()) {
- lemma = lemmaIter.next();
- fireProgressChanged(++lemmaIdx * 100 / linesCnt);
- this.add(lemma);
- } // while
+ Lemma lemma;
+ fireStatusChanged(READING + configLine);
+ while(lemmaIter.hasNext()) {
+ lemma = lemmaIter.next();
+ fireProgressChanged(++lemmaIdx * 100 / linesCnt);
+ this.add(lemma);
+ } // while
+ }
+
+ } finally {
+ configReader.close();
}
- } finally {
- configReader.close();
- }
+ fireProcessFinished();
- fireProcessFinished();
+ if(DEBUG) {
+ System.out.println("Infl. Gaz. Is GC Enabled ? = " +
+ profiler.isGCCallingEnabled());
+ profiler.checkPoint("Infl. Gaz. init completed.");
+ } // DEBUG
- if(DEBUG) {
- System.out.println("Infl. Gaz. Is GC Enabled ? = " +
- profiler.isGCCallingEnabled());
- profiler.checkPoint("Infl. Gaz. init completed.");
- } // DEBUG
+ if(DEBUG) {
+ System.out.println("Starting Garbage Collection ...");
+ System.gc();
+ profiler.checkPoint("Garbage Collection finished");
+ }
- if(DEBUG) {
- System.out.println("Starting Garbage Collection ...");
- System.gc();
- profiler.checkPoint("Garbage Collection finished");
- }
+ if(DEBUG) {
+ System.out.println("\nLEMMAs COUNT -> ");
+ System.out.println(lemmaIdx);
+ System.out.println("\nTYPE POOL SIZE -> ");
+ System.out.println(TypePool.size());
+ System.out.println("\nSUFFIX POOL SIZE -> ");
+ System.out.println(SuffixPool.size());
+ }
+ if(DETAILED_DEBUG) {
+ // dump distinct types
+ System.out.println("\nTYPE POOL ->");
+ System.out.println(TypePool.getString());
+ System.out.println("\nSUFFIX POOL ->");
+ System.out.println(SuffixPool.getString());
- if(DEBUG) {
- System.out.println("\nLEMMAs COUNT -> ");
- System.out.println(lemmaIdx);
- System.out.println("\nTYPE POOL SIZE -> ");
- System.out.println(TypePool.size());
- System.out.println("\nSUFFIX POOL SIZE -> ");
- System.out.println(SuffixPool.size());
- }
- if(DETAILED_DEBUG) {
- // dump distinct types
- System.out.println("\nTYPE POOL ->");
- System.out.println(TypePool.getString());
- System.out.println("\nSUFFIX POOL ->");
- System.out.println(SuffixPool.getString());
+ }
- }
+ } catch(Exception x) {
+ throw new ResourceInstantiationException(x);
+ } // catch
+ }
- } catch(Exception x) {
- throw new ResourceInstantiationException(x);
- } // catch
-
return this;
} // Resource init() throws ResourceInstantiationException
/**
+ * Re-initialize this gazetteer by re-loading the configuration.
+ */
+ public void reInit() throws ResourceInstantiationException {
+ mapsList = null;
+ msTypeSet = null;
+ init();
+ }
+
+ /**
* This method runs the gazetteer. It parses the document and looks-up the
* parsed phrases from the maps, in which the phrases vs. annotations are
set,
* in order to generate an annotation set. It assumes that all the needed
@@ -764,4 +780,34 @@
return caseSensitive;
}
+
+ /**
+ * For internal use by the duplication mechanism.
+ */
+ @Sharable
+ public void setMapsList(List<Map> mapsList) {
+ this.mapsList = mapsList;
+ }
+
+ /**
+ * For internal use by the duplication mechanism.
+ */
+ public List<Map> getMapsList() {
+ return mapsList;
+ }
+
+ /**
+ * For internal use by the duplication mechanism.
+ */
+ @Sharable
+ public void setMsTypeSet(Set<SuffixNest> msTypeSet) {
+ this.msTypeSet = msTypeSet;
+ }
+
+ /**
+ * For internal use by the duplication mechanism.
+ */
+ public Set<SuffixNest> getMsTypeSet() {
+ return msTypeSet;
+ }
}
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
Learn the latest--Visual Studio 2012, SharePoint 2013, SQL 2012, more!
Discover the easy way to master current and previous Microsoft technologies
and advance your career. Get an incredible 1,500+ hours of step-by-step
tutorial videos with LearnDevNow. Subscribe today and save!
http://pubads.g.doubleclick.net/gampad/clk?id=58041391&iu=/4140/ostg.clktrk
_______________________________________________
GATE-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gate-cvs
