Update entailment classifier; tweak OpenIE demo

stanfordnlp · May 15, 2015 · 2684210 · 2684210
1 parent ffabc1e
commit 2684210
Show file tree

Hide file tree

Showing 3 changed files with 40 additions and 28 deletions.
diff --git a/src/edu/stanford/nlp/naturalli/ClauseSplitterSearchProblem.java b/src/edu/stanford/nlp/naturalli/ClauseSplitterSearchProblem.java
@@ -58,6 +58,10 @@ public class ClauseSplitterSearchProblem {
       add("clone_nsubj");
       add("simple");
     }});
+    put("vmod", new ArrayList<String>() {{
+      add("clone_nsubj");
+      add("simple");
+    }});
   }});
 
   /**

diff --git a/src/edu/stanford/nlp/naturalli/OpenIEServlet.java b/src/edu/stanford/nlp/naturalli/OpenIEServlet.java
@@ -36,6 +36,7 @@ public class OpenIEServlet extends HttpServlet {
   public void init()  throws ServletException {
     Properties commonProps = new Properties() {{
       setProperty("depparse.extradependencies", "ref_only_uncollapsed");
+      setProperty("parse.extradependencies", "ref_only_uncollapsed");
       setProperty("openie.splitter.threshold", "0.10");
       setProperty("openie.optimze_for", "GENERAL");
       setProperty("openie.ignoreaffinity", "false");
@@ -73,23 +74,22 @@ public void init()  throws ServletException {
   /**
    * Annotate a document (which is usually just a sentence).
    */
-  public void annotate(Annotation ann) {
-    pipeline.annotate(ann);
-    if (ann.get(CoreAnnotations.SentencesAnnotation.class).size() == 1) {
-      CoreMap sentence = ann.get(CoreAnnotations.SentencesAnnotation.class).get(0);
-      if (sentence.get(NaturalLogicAnnotations.RelationTriplesAnnotation.class) != null) {
-        if (sentence.get(NaturalLogicAnnotations.RelationTriplesAnnotation.class).isEmpty()) {
-          for (CoreLabel token : sentence.get(CoreAnnotations.TokensAnnotation.class)) {
-            token.remove(NaturalLogicAnnotations.OperatorAnnotation.class);
-            token.remove(NaturalLogicAnnotations.PolarityAnnotation.class);
-          }
-          sentence.remove(NaturalLogicAnnotations.RelationTriplesAnnotation.class);
-          sentence.remove(NaturalLogicAnnotations.EntailedSentencesAnnotation.class);
-          sentence.remove(SemanticGraphCoreAnnotations.BasicDependenciesAnnotation.class);
-          sentence.remove(SemanticGraphCoreAnnotations.CollapsedDependenciesAnnotation.class);
-          sentence.remove(SemanticGraphCoreAnnotations.CollapsedCCProcessedDependenciesAnnotation.class);
-          backoff.annotate(ann);
+  public void annotate(StanfordCoreNLP pipeline, Annotation ann) {
+    if (ann.get(CoreAnnotations.SentencesAnnotation.class) == null) {
+      pipeline.annotate(ann);
+    } else {
+      if (ann.get(CoreAnnotations.SentencesAnnotation.class).size() == 1) {
+        CoreMap sentence = ann.get(CoreAnnotations.SentencesAnnotation.class).get(0);
+        for (CoreLabel token : sentence.get(CoreAnnotations.TokensAnnotation.class)) {
+          token.remove(NaturalLogicAnnotations.OperatorAnnotation.class);
+          token.remove(NaturalLogicAnnotations.PolarityAnnotation.class);
         }
+        sentence.remove(NaturalLogicAnnotations.RelationTriplesAnnotation.class);
+        sentence.remove(NaturalLogicAnnotations.EntailedSentencesAnnotation.class);
+        sentence.remove(SemanticGraphCoreAnnotations.BasicDependenciesAnnotation.class);
+        sentence.remove(SemanticGraphCoreAnnotations.CollapsedDependenciesAnnotation.class);
+        sentence.remove(SemanticGraphCoreAnnotations.CollapsedCCProcessedDependenciesAnnotation.class);
+        pipeline.annotate(ann);
       }
     }
   }
@@ -153,6 +153,22 @@ public static String quote(String string) {
     return sb.toString();
   }
 
+  private void runWithPipeline(StanfordCoreNLP pipeline, Annotation ann, Set<String> triples, Set<String> entailments) {
+    // Annotate
+    annotate(pipeline, ann);
+    // Extract info
+    for (CoreMap sentence : ann.get(CoreAnnotations.SentencesAnnotation.class)) {
+      for (SentenceFragment fragment : sentence.get(NaturalLogicAnnotations.EntailedSentencesAnnotation.class)) {
+        entailments.add(quote(fragment.toString()));
+      }
+      for (RelationTriple fragment : sentence.get(NaturalLogicAnnotations.RelationTriplesAnnotation.class)) {
+        triples.add("[ " + quote(fragment.subjectGloss()) + ", " + quote(fragment.relationGloss()) + ", " + quote(fragment.objectGloss()) + " ]");
+      }
+    }
+
+  }
+
+
   /**
    * Actually perform the GET request, given all the relevant information (already sanity checked).
    * This is the meat of the servlet code.
@@ -172,19 +188,11 @@ private void doGet(PrintWriter out, String q) {
     // Annotate
     Annotation ann = new Annotation(q);
     try {
-      // Annotate
-      annotate(ann);
       // Collect results
-      List<String> entailments = new ArrayList<>();
+      Set<String> entailments = new HashSet<>();
       Set<String> triples = new LinkedHashSet<>();
-      for (CoreMap sentence : ann.get(CoreAnnotations.SentencesAnnotation.class)) {
-        for (SentenceFragment fragment : sentence.get(NaturalLogicAnnotations.EntailedSentencesAnnotation.class)) {
-          entailments.add(quote(fragment.toString()));
-        }
-        for (RelationTriple fragment : sentence.get(NaturalLogicAnnotations.RelationTriplesAnnotation.class)) {
-          triples.add("[ " + quote(fragment.subjectLemmaGloss()) + ", " + quote(fragment.relationLemmaGloss()) + ", " + quote(fragment.objectLemmaGloss()) + " ]");
-        }
-      }
+      runWithPipeline(pipeline, ann, triples, entailments);  // pipeline must come before backoff
+      runWithPipeline(backoff, ann, triples, entailments);   // backoff must come after pipeline
       // Write results
       out.println("{ " +
           "\"ok\":true, " +

diff --git a/src/edu/stanford/nlp/pipeline/AnnotatorFactories.java b/src/edu/stanford/nlp/pipeline/AnnotatorFactories.java
@@ -444,7 +444,7 @@ public String additionalSignature() {
   //
   // Parser
   //
-  public static AnnotatorFactory parse(Properties properties, final AnnotatorImplementations annotatorImplementation) {
+  public static AnnotatorFactory parse(final Properties properties, final AnnotatorImplementations annotatorImplementation) {
     return new AnnotatorFactory(properties, annotatorImplementation) {
       private static final long serialVersionUID = 1L;