stanfordnlp/CoreNLP · error · IllegalStateException

Format is not implemented: " + FORMAT

Error message

Format is not implemented: " + FORMAT

What it means

tripleToString() formats extracted RelationTriples according to the configured output format (OpenIE.FORMAT). The switch handles REVERB, QA_SRL, and DEFAULT; any other configured format is not implemented, so an IllegalStateException is thrown. This happens during output, after successful extraction.

Solutions

1. Set openie.format to one of the supported values: default, reverb, or qasrl.
2. Check spelling/case of the configured format string against OpenIE.Format enum values.
3. Upgrade CoreNLP if you need a format introduced after your current version.
4. Validate/normalize the format option at program start rather than failing mid-output.

Example fix

// before
props.setProperty("openie.format", "json");
// after
props.setProperty("openie.format", "default"); // or reverb / qasrl

String fmt = props.getProperty("openie.format", "default").toLowerCase(Locale.ROOT);
if (!fmt.equals("default") && !fmt.equals("reverb") && !fmt.equals("qasrl")) {
  throw new IllegalArgumentException("Unsupported openie.format: " + fmt);
}

try {
  String s = tripleToString(extraction, sentence);
} catch (IllegalStateException e) {
  log.error("Invalid openie.format configured: " + e.getMessage());
}

When it happens

Trigger: Setting the -openie.format property to a value other than default/reverb/qasrl (e.g. a typo or a format supported only in newer CoreNLP versions) and running OpenIE's main/CLI through processDocument.

Common situations: Typos in the openie.format CLI option; copying example configs from a different CoreNLP version where additional formats exist; case-sensitive format names.

Related errors

• Cannot output in format
• Sorry, but dependencies cannot be output as part of the…
• annotator " " requires annotation " ". The usual…
• : Entry doesn't have overwriteable types , but entry type…
• : Ignoring duplicate entry: , old type = , new type =

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/30166939747d10ad. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/naturalli/OpenIE.java:634

   *
   * @param extraction The triple to write.
   * @param docid The document ID (for the ReVerb format)
   * @param sentence The sentence the triple was extracted from (for the ReVerb format)
   *
   * @return A String representation of the triple.
   */
  public static String tripleToString(RelationTriple extraction, String docid, CoreMap sentence) {
    switch (FORMAT) {
      case REVERB:
        return extraction.toReverbString(docid, sentence);
      case OLLIE:
        return extraction.confidenceGloss() + ": (" + extraction.subjectGloss() + "; " + extraction.relationGloss() + "; " + extraction.objectGloss() + ')';
      case DEFAULT:
        return extraction.toString();
      case QA_SRL:
        return extraction.toQaSrlString(sentence);
      default:
        throw new IllegalStateException("Format is not implemented: " + FORMAT);
    }

  }

  /**
   * Process a single file or line of standard in.
   *
   * @param pipeline The annotation pipeline to run the lines of the input through.
   * @param docid The docid of the document we are extracting.
   * @param document the document to annotate.
   */
  @SuppressWarnings("SynchronizeOnNonFinalField")
  private static void processDocument(AnnotationPipeline pipeline, String docid, String document) {
    // Error checks
    if (document.trim().isEmpty()) {
      return;
    }

View on GitHub (pinned to 1b7edd19c4)