Include {} as punct words, as some parsers don't produce LCB/RCB

stanfordnlp · Jul 7, 2023 · a606afa · a606afa
1 parent 56cd6bb
commit a606afa
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/src/edu/stanford/nlp/trees/PennTreebankLanguagePack.java b/src/edu/stanford/nlp/trees/PennTreebankLanguagePack.java
@@ -29,7 +29,7 @@ public PennTreebankLanguagePack() {
   private static final String[] collinsPunctTags = {"''", "``", ".", ":", ","};
 
   // new tokenizers should return (), old tokenizers return -LRB- -RRB-.  so we anticipate both
-  private static final String[] pennPunctWords = {"''", "'", "``", "`", "-LRB-", "-RRB-", "(", ")", "-LCB-", "-RCB-", ".", "?", "!", ",", ":", "-", "--", "...", ";"};
+  private static final String[] pennPunctWords = {"{", "}", "''", "'", "``", "`", "-LRB-", "-RRB-", "(", ")", "-LCB-", "-RCB-", ".", "?", "!", ",", ":", "-", "--", "...", ";"};
 
   private static final String[] pennSFPunctWords = {".", "!", "?"};