Revision: 17460
http://sourceforge.net/p/gate/code/17460
Author: adamfunk
Date: 2014-02-26 21:29:40 +0000 (Wed, 26 Feb 2014)
Log Message:
-----------
Deprecate & TODO everything, when in doubt.
Modified Paths:
--------------
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/apply/TermScoreCopier.java
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/AbstractTermbank.java
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/AnnotationTermbank.java
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/DocumentFrequencyBank.java
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/HyponymyTermbank.java
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/TfIdfTermbank.java
Modified:
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/apply/TermScoreCopier.java
===================================================================
---
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/apply/TermScoreCopier.java
2014-02-26 21:09:57 UTC (rev 17459)
+++
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/apply/TermScoreCopier.java
2014-02-26 21:29:40 UTC (rev 17460)
@@ -69,7 +69,6 @@
String termFeature = termbank.getInputAnnotationFeature();
String languageFeature = termbank.getLanguageFeature();
String scoreFeature = termbank.getScoreProperty();
- String rawScoreFeature = scoreFeature + ".raw";
AnnotationSet candidates =
document.getAnnotations(annotationSetName).get(annotationTypes);
checkInterruption();
@@ -82,15 +81,14 @@
fm.put(scoreFeature, score);
}
- Double rawScore = termbank.getRawScore(term);
- if (rawScore != null) {
- fm.put(rawScoreFeature, rawScore);
- }
-
if (useFeature(frequencyFeature)) {
fm.put(frequencyFeature, termbank.getTermFrequency(term));
}
+ // TODO
+ // Figure out whether to deal with non-main scores (term freq, doc freq,
+ // ref doc freq, &c.; & if so, how
+
if (useFeature(docFrequencyFeature)) {
fm.put(docFrequencyFeature, termbank.getDocFrequency(term));
}
Modified:
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/AbstractTermbank.java
===================================================================
---
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/AbstractTermbank.java
2014-02-26 21:09:57 UTC (rev 17459)
+++
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/AbstractTermbank.java
2014-02-26 21:29:40 UTC (rev 17460)
@@ -26,8 +26,10 @@
import javax.swing.Action;
+import org.apache.commons.lang.StringEscapeUtils;
+
public abstract class AbstractTermbank extends AbstractBank
implements ActionsPublisher {
private static final long serialVersionUID = -2809051430169834059L;
@@ -288,6 +290,7 @@
}
+ @Deprecated
public int getTermFrequency(Term term) {
if (termFrequencies.containsKey(term)) {
return termFrequencies.get(term);
@@ -297,6 +300,7 @@
}
+ @Deprecated
public int getDocFrequency(Term term) {
if (docFrequencies.containsKey(term)) {
return docFrequencies.get(term);
@@ -305,13 +309,48 @@
return 0;
}
+
+ public String getCsvHeader() {
+ StringBuilder sb = new StringBuilder();
+ sb.append(StringEscapeUtils.escapeCsv("Term"));
+ sb.append(',').append(StringEscapeUtils.escapeCsv("Lang"));
+ sb.append(',').append(StringEscapeUtils.escapeCsv("Type"));
+ for (ScoreType type : this.scoreTypes) {
+ sb.append(',').append(StringEscapeUtils.escapeCsv(type.toString()));
+ }
+ sb.append(getCsvSubheader());
+ return sb.toString();
+ }
- public abstract String getCsvHeader();
+
+ /**
+ * TODO: This is not right (columns).
+ * Should be overridden as necessary, for totals etc.
+ * Must start with a newline.
+ * @return
+ */
+ protected String getCsvSubheader() {
+ StringBuilder sb = new StringBuilder();
+ sb.append('\n');
+ sb.append(',').append(StringEscapeUtils.escapeCsv("_TOTAL_DOCS_"));
+ sb.append(',').append(StringEscapeUtils.escapeCsv(""));
+ sb.append(',').append(StringEscapeUtils.escapeCsv(""));
+
sb.append(',').append(StringEscapeUtils.escapeCsv(Integer.toString(this.getDocumentCount())));
+ return sb.toString();
+ }
- public abstract String getCsvLine(Term term);
+ public String getCsvLine(Term term) {
+ StringBuilder sb = new StringBuilder();
+ sb.append(StringEscapeUtils.escapeCsv(term.getTermString()));
+
sb.append(',').append(StringEscapeUtils.escapeCsv(term.getLanguageCode()));
+ sb.append(',').append(StringEscapeUtils.escapeCsv(term.getType()));
+ for (ScoreType type : this.scoreTypes) {
+ sb.append(',').append(StringEscapeUtils.escapeCsv(this.getScore(type,
term).toString()));
+ }
+ return sb.toString();
+ }
-
/***** CREOLE PARAMETERS *****/
Modified:
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/AnnotationTermbank.java
===================================================================
---
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/AnnotationTermbank.java
2014-02-26 21:09:57 UTC (rev 17459)
+++
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/AnnotationTermbank.java
2014-02-26 21:29:40 UTC (rev 17460)
@@ -91,38 +91,6 @@
}
- public String getCsvHeader() {
- StringBuilder sb = new StringBuilder();
- sb.append(StringEscapeUtils.escapeCsv("Term"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Lang"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Type"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("ScoreType"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Score"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Document_Count"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Term_Frequency"));
- return sb.toString();
- }
-
-
- public String getCsvLine(Term term) {
- StringBuilder sb = new StringBuilder();
- sb.append(StringEscapeUtils.escapeCsv(term.getTermString()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(term.getLanguageCode()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(term.getType()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(this.getScoreProperty()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(this.getScore(term).toString()));
- sb.append(',');
-
sb.append(StringEscapeUtils.escapeCsv(Integer.toString(this.getDocFrequency(term))));
- sb.append(',');
-
sb.append(StringEscapeUtils.escapeCsv(Integer.toString(this.getTermFrequency(term))));
- return sb.toString();
- }
-
-
protected void initializeScoreTypes() {
this.scoreTypes = new ArrayList<ScoreType>();
this.scoreTypes.add(new ScoreType(scoreProperty));
Modified:
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/DocumentFrequencyBank.java
===================================================================
---
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/DocumentFrequencyBank.java
2014-02-26 21:09:57 UTC (rev 17459)
+++
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/DocumentFrequencyBank.java
2014-02-26 21:29:40 UTC (rev 17460)
@@ -319,12 +319,8 @@
}
- public String getCsvHeader() {
+ public String getCsvSubheader() {
StringBuilder sb = new StringBuilder();
- sb.append(StringEscapeUtils.escapeCsv("Term"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Lang"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Type"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("DocFrequency"));
sb.append('\n');
sb.append(',').append(StringEscapeUtils.escapeCsv("_TOTAL_DOCS_"));
sb.append(',').append(StringEscapeUtils.escapeCsv(""));
Modified:
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/HyponymyTermbank.java
===================================================================
---
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/HyponymyTermbank.java
2014-02-26 21:09:57 UTC (rev 17459)
+++
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/HyponymyTermbank.java
2014-02-26 21:29:40 UTC (rev 17460)
@@ -161,36 +161,6 @@
- public String getCsvHeader() {
- StringBuilder sb = new StringBuilder();
- sb.append(StringEscapeUtils.escapeCsv("Term"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Lang"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Type"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("ScoreType"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Score"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Document_Count"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Term_Frequency"));
- return sb.toString();
- }
-
- public String getCsvLine(Term term) {
- StringBuilder sb = new StringBuilder();
- sb.append(StringEscapeUtils.escapeCsv(term.getTermString()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(term.getLanguageCode()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(term.getType()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(this.getScoreProperty()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(this.getScore(term).toString()));
- sb.append(',');
-
sb.append(StringEscapeUtils.escapeCsv(Integer.toString(this.getDocFrequency(term))));
- sb.append(',');
-
sb.append(StringEscapeUtils.escapeCsv(Integer.toString(this.getTermFrequency(term))));
- return sb.toString();
- }
-
/***** CREOLE PARAMETERS *****/
@CreoleParameter(comment = "Annotation features (in order) to be scanned as
terms' heads")
Modified:
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/TfIdfTermbank.java
===================================================================
---
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/TfIdfTermbank.java
2014-02-26 21:09:57 UTC (rev 17459)
+++
gate/branches/termraider-refactoring/plugins/TermRaider/src/gate/termraider/bank/TfIdfTermbank.java
2014-02-26 21:29:40 UTC (rev 17460)
@@ -149,41 +149,6 @@
}
- public String getCsvHeader() {
- StringBuilder sb = new StringBuilder();
- sb.append(StringEscapeUtils.escapeCsv("Term"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Lang"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Type"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("ScoreType"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Score"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Document_Count"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Ref_Doc_Frequency"));
- sb.append(',').append(StringEscapeUtils.escapeCsv("Term_Frequency"));
- return sb.toString();
- }
-
-
- public String getCsvLine(Term term) {
- StringBuilder sb = new StringBuilder();
- sb.append(StringEscapeUtils.escapeCsv(term.getTermString()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(term.getLanguageCode()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(term.getType()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(this.getScoreProperty()));
- sb.append(',');
- sb.append(StringEscapeUtils.escapeCsv(this.getScore(term).toString()));
- sb.append(',');
-
sb.append(StringEscapeUtils.escapeCsv(Integer.toString(this.getDocFrequency(term))));
- sb.append(',');
-
sb.append(StringEscapeUtils.escapeCsv(Integer.toString(this.docFreqSource.getDocFrequency(term))));
- sb.append(',');
-
sb.append(StringEscapeUtils.escapeCsv(Integer.toString(this.getTermFrequency(term))));
- return sb.toString();
- }
-
-
protected void prepare() throws ResourceInstantiationException {
if ( (corpora == null) || (corpora.size() == 0) ) {
throw new ResourceInstantiationException("No corpora given");
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
Flow-based real-time traffic analytics software. Cisco certified tool.
Monitor traffic, SLAs, QoS, Medianet, WAAS etc. with NetFlow Analyzer
Customize your own dashboards, set traffic alerts and generate reports.
Network behavioral analysis & security monitoring. All-in-one tool.
http://pubads.g.doubleclick.net/gampad/clk?id=126839071&iu=/4140/ostg.clktrk
_______________________________________________
GATE-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gate-cvs