Add to Ssurgeon the ability to update features without overwriting the whole feature map

Allow EditNode with morpho updates and no other attributes changed

Author: AngledLuffa

src/edu/stanford/nlp/semgraph/semgrex/ssurgeon/EditNode.java

@@ -1,12 +1,16 @@
 package edu.stanford.nlp.semgraph.semgrex.ssurgeon;
 
+import java.util.Collections;
+import java.util.HashMap;
 import java.util.Map;
 import java.util.TreeMap;
 
+import edu.stanford.nlp.ling.CoreAnnotations;
 import edu.stanford.nlp.ling.CoreLabel;
 import edu.stanford.nlp.ling.IndexedWord;
 import edu.stanford.nlp.semgraph.SemanticGraph;
 import edu.stanford.nlp.semgraph.semgrex.SemgrexMatcher;
+import edu.stanford.nlp.trees.ud.CoNLLUUtils;
 
 /**
  * Edit an existing node to have new attributes.
@@ -18,17 +22,23 @@ public class EditNode extends SsurgeonEdit {
 
   final String nodeName;
   final Map<String, String> attributes;
+  final Map<String, String> updateMorphoFeatures;
 
-  public EditNode(String nodeName, Map<String, String> attributes) {
+  public EditNode(String nodeName, Map<String, String> attributes, String updateMorphoFeatures) {
     if (nodeName == null) {
       throw new SsurgeonParseException("Cannot make an EditNode with no nodeName");
     }
-    if (attributes.size() == 0) {
-      throw new SsurgeonParseException("Cannot make an EditNode with no attributes");
+    if (attributes.size() == 0 && updateMorphoFeatures == null) {
+      throw new SsurgeonParseException("Cannot make an EditNode with no attributes or updated morphological features");
     }
     AddDep.checkIllegalAttributes(attributes);
     this.nodeName = nodeName;
     this.attributes = new TreeMap<>(attributes);
+    if (updateMorphoFeatures != null) {
+      this.updateMorphoFeatures = CoNLLUUtils.parseFeatures(updateMorphoFeatures);
+    } else {
+      this.updateMorphoFeatures = Collections.emptyMap();
+    }
   }
 
 
@@ -47,9 +57,16 @@ public String toEditString() {
       buf.append(key);
       buf.append(" ");
       buf.append(attributes.get(key));
+      // TODO: why the stray quote characters?
       buf.append("\"\t");
     }
 
+    if (this.updateMorphoFeatures.size() > 0) {
+      buf.append(Ssurgeon.UPDATE_MORPHO_FEATURES);
+      buf.append(" ");
+      buf.append(CoNLLUUtils.toFeatureString(this.updateMorphoFeatures));
+    }
+
     return buf.toString();
   }
 
@@ -76,6 +93,21 @@ public boolean evaluate(SemanticGraph sg, SemgrexMatcher sm) {
       }
     }
 
+    for (String key : updateMorphoFeatures.keySet()) {
+      HashMap<String, String> features = word.get(CoreAnnotations.CoNLLUFeats.class);
+      if (features == null) {
+        changed = true;
+        features = new HashMap<>();
+        word.set(CoreAnnotations.CoNLLUFeats.class, features);
+      }
+
+      // this test will catch null, eg not yet assigned, features as well
+      if (!updateMorphoFeatures.get(key).equals(features.get(key))) {
+        changed = true;
+        features.put(key, updateMorphoFeatures.get(key));
+      }
+    }
+
     return changed;
   }
 }

src/edu/stanford/nlp/semgraph/semgrex/ssurgeon/Ssurgeon.java

@@ -397,6 +397,7 @@ public Collection<SsurgeonWordlist> getResources() {
   public static final String WEIGHT_ARG = "-weight";
   public static final String NAME_ARG = "-name";
   public static final String POSITION_ARG = "-position";
+  public static final String UPDATE_MORPHO_FEATURES = "-updateMorphoFeatures";
 
 
   // args for Ssurgeon edits, allowing us to not
@@ -422,6 +423,8 @@ protected static class SsurgeonArgs {
 
     public String position = null;
 
+    public String updateMorphoFeatures = null;
+
     public Map<String, String> annotations = new TreeMap<>();
   }
 
@@ -492,6 +495,9 @@ private static SsurgeonArgs parseArgsBox(String args, Map<String, String> additi
         case POSITION_ARG:
           argsBox.position = argsValue;
           break;
+        case UPDATE_MORPHO_FEATURES:
+          argsBox.updateMorphoFeatures = argsValue;
+          break;
         default:
           String key = argsKey.substring(1);
           Class<? extends CoreAnnotation<?>> annotation = AnnotationLookup.toCoreKey(key);
@@ -557,7 +563,7 @@ public static SsurgeonEdit parseEditLine(String editLine, Map<String, String> at
         if (argsBox.nodes.size() != 1) {
           throw new SsurgeonParseException("Cannot make an EditNode out of " + argsBox.nodes.size() + " nodes.  Please use exactly one -node");
         }
-        return new EditNode(argsBox.nodes.get(0), argsBox.annotations);
+        return new EditNode(argsBox.nodes.get(0), argsBox.annotations, argsBox.updateMorphoFeatures);
       } else if (command.equalsIgnoreCase(MergeNodes.LABEL)) {
         if (argsBox.nodes.size() < 2) {
           throw new SsurgeonParseException("Cannot make a MergeNodes out of fewer than 2 nodes (got " + argsBox.nodes.size() + ")");

src/edu/stanford/nlp/trees/ud/CoNLLUUtils.java

@@ -35,7 +35,7 @@ public static HashMap<String,String> parseFeatures(String featureString) {
      *
      * @return The feature string.
      */
-    public static String toFeatureString(HashMap<String,String> features) {
+    public static String toFeatureString(Map<String,String> features) {
         StringBuilder sb = new StringBuilder();
         boolean first = true;
         if (features != null) {