Warning: ChineseNumberSequenceClassifier does not have…

Error message

Warning: ChineseNumberSequenceClassifier does not have SUTime implementation.

What it means

classifyWithGlobalInformation is the hook where document-level (SUTime) information would be applied. Since no Chinese SUTime implementation exists, the class warns if useSUTime was requested and then classifies normally without temporal information.

Solutions

Rebuild the classifier/pipeline with useSUTime=false to eliminate the per-call warning.
Run temporal extraction separately with a Chinese-capable tool if needed.
Filter/downgrade this known warning in logging configuration if SUTime must remain enabled.

Example fix

// before
props.setProperty("ner.useSUTime", "true");
// after
props.setProperty("ner.useSUTime", "false");

Defensive patterns

Strategy: validation

Validate before calling

if (classifier instanceof ChineseNumberSequenceClassifier
    && Boolean.parseBoolean(props.getProperty("ner.useSUTime"))) {
  props.setProperty("ner.useSUTime", "false"); // avoid per-call warnings
}

Prevention

Disable useSUTime for Chinese pipelines.
Suppress this known warning in log filtering if unavoidable.
Handle temporal tagging with a separate Chinese tool.

When it happens

Trigger: Calling classifyWithGlobalInformation on a ChineseNumberSequenceClassifier constructed with useSUTime=true — i.e. every document classification in a pipeline that enabled SUTime for Chinese.

Common situations: Chinese NER pipeline configured with ner.useSUTime=true (copied from English defaults); the warning repeats per document, polluting logs.

Related errors

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/bcf11aa2d3791c8b. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/ie/regexp/ChineseNumberSequenceClassifier.java:203

    int sz = pl.size();
    while (j < sz && pl.get(j).getString(CoreAnnotations.PartOfSpeechAnnotation.class).equals("CD")) {
      j++;
    }
    if (j >= sz) {
      return false;
    }
    String tag = pl.get(j).getString(CoreAnnotations.PartOfSpeechAnnotation.class);
    String word = pl.get(j).word();
    if (DEBUG) {
      log.info("rightScan testing: " + word + '/' + tag + "; answer is: " + Boolean.toString((tag.equals("NN") || tag.equals("NNS")) && CURRENCY_WORD_PATTERN.matcher(word).matches()));
    }
    return (tag.equals("M") || tag.equals("NN") || tag.equals("NNS")) && CURRENCY_WORD_PATTERN.matcher(word).matches();
  }

  @Override
  public List<CoreLabel> classifyWithGlobalInformation(List<CoreLabel> tokenSequence, CoreMap document, CoreMap sentence) {
    if(useSUTime) {
      log.warn("Warning: ChineseNumberSequenceClassifier does not have SUTime implementation.");
    }
    return classify(tokenSequence);
  }

  @Override
  public void train(Collection<List<CoreLabel>> docs, DocumentReaderAndWriter<CoreLabel> readerAndWriter) {
    // Train is not needed for this rule-based classifier
  }

  @Override
  public void serializeClassifier(String serializePath) {
  }

  @Override
  public void serializeClassifier(ObjectOutputStream oos) {
  }

  @Override

View on GitHub (pinned to 1b7edd19c4)