stanfordnlp/CoreNLP · error · RuntimeException

Attempt to create ChineseSimWordAvgDepGrammar before…

Error message

Attempt to create ChineseSimWordAvgDepGrammar before Lexicon!!!

What it means

ChineseTreebankParserParams's dependency grammar extractor creates a ChineseSimWordAvgDepGrammar in formResult() and immediately requires the lexicon field to already be set. If lex is still null — meaning the lexicon was created after (or without) the dependency grammar extractor — formResult throws RuntimeException telling you the grammar was created before the Lexicon.

Solutions

1. Ensure the Lexicon is constructed and assigned (tlpParams.lex / extractor lex field) before formResult() is invoked
2. Use the standard LexicalizedParser training entry points which create the lexicon first
3. Patch your custom extractor to lazily obtain or defer dg.setLex(lex) until the lexicon exists

Example fix

// before
MLEDependencyGrammar dg = extractor.formResult(); // lex == null -> throws
// after
lex = op.tlpParams.lex(op, wordIndex, tagIndex); // create lexicon first
dg = extractor.formResult();

// ensure lexicon exists before formResult
if (tlpParams.lex == null) { tlpParams.lex = op.tlpParams.lex(op, wordIndex, tagIndex); }

boolean lexiconReady(ChineseTreebankParserParams p) { return p.lex != null; }

try { dg = extractor.formResult(); } catch (RuntimeException e) { if (e.getMessage().contains("before Lexicon")) { initializeLexiconFirst(); } }

When it happens

Trigger: Using ChineseTreebankParserParams with a similar-word dependency grammar extractor where the training pipeline calls formResult() before a Lexicon has been assigned to the params/extractor, leaving the lex field null.

Common situations: Custom training scripts that build the grammar extractor directly instead of going through the standard train path that first creates the lexicon; ordering mistakes when wiring Options/tlpParams manually.

Understand the failure class

Background: "Invalid state transition" errors: "status must be X, actually Y", "already rejected/charging/uninstalled", "cannot ... while running" — what they mean when a library rejects your call — this error's family across 31 libraries.

Related errors

• attempt to get word when sentence and lattice are null!
• Attribute already defined:
• Attribute already defined
• Attribute match already defined:
• Attribute match already defined

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/8ecd17ab2baa620a. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/parser/lexparser/ChineseTreebankParserParams.java:1198

          wfe2.setFeatureLevel(depGramFeatureLevel);
          MaxentDependencyGrammar dg = new MaxentDependencyGrammar(op.tlpParams, wfe, wfe2, true, false, false);
          dg.train(trees);
          return dg;
        }

        public Object extract(Iterator<Tree> iterator, Function<Tree, Tree> f) {
          throw new UnsupportedOperationException();
        }
      };
    } else ------- */
    if (useSimilarWordMap) {
      return new MLEDependencyGrammarExtractor(op, wordIndex, tagIndex) {
        @Override
        public MLEDependencyGrammar formResult() {
          wordIndex.addToIndex(Lexicon.UNKNOWN_WORD);
          ChineseSimWordAvgDepGrammar dg = new ChineseSimWordAvgDepGrammar(tlpParams, directional, useDistance, useCoarseDistance, op.trainOptions.basicCategoryTagsInDependencyGrammar, op, wordIndex, tagIndex);
          if (lex == null) {
            throw new RuntimeException("Attempt to create ChineseSimWordAvgDepGrammar before Lexicon!!!");
          } else {
            dg.setLex(lex);
          }
          for (IntDependency dependency : dependencyCounter.keySet()) {
            dg.addRule(dependency, dependencyCounter.getCount(dependency));
          }
          return dg;
        }

     };
    } else {
      return new MLEDependencyGrammarExtractor(op, wordIndex, tagIndex);
    }
  }

  /**
   * Return a default sentence for the language (for testing)
   */

View on GitHub (pinned to 1b7edd19c4)