stanfordnlp/CoreNLP · error · IllegalArgumentException

Invalid metric type for

Error message

Invalid metric type for ${Constants.OPTIMIZE_SIEVES_SCORE_PROP} property: ${optimizeScoreType}

What it means

The SieveCoreferenceSystem constructor validates the metric named by the Constants.OPTIMIZE_SIEVES_SCORE_PROP property against the list of valid scorer metric types. If the configured optimizeScoreType (or its metric part) is not a recognized coref scoring metric such as muc, bcub, ceafe or ceafm, it throws an IllegalArgumentException before the system is built.

Solutions

1. Set the property to a valid metric name exactly as accepted by the system (muc, bcub, ceafe, ceafm, etc.)
2. Check spelling and case against the validMetricType list used in the validation loop
3. If using a subscore form, ensure the part after the separator is a valid CorefScorer.SubScoreType enum constant
4. Consult the Constants class / CorefScorer source for the exact accepted metric strings in your CoreNLP version

Example fix

// before
coref.props: dcoref.optimize.sievesScore = bcube
// after
coref.props: dcoref.optimize.sievesScore = bcub

Set<String> valid = new HashSet<>(Arrays.asList("muc","bcub","ceafe","ceafm"));
String score = props.getProperty("dcoref.optimize.sievesScore");
if (score != null && !valid.contains(score.split("\\.")[0])) {
  throw new IllegalArgumentException("Unknown metric: " + score);
}

try {
  SieveCoreferenceSystem corefSystem = new SieveCoreferenceSystem(props);
} catch (IllegalArgumentException e) {
  logger.severe("Bad optimize metric config: " + e.getMessage());
  props.setProperty(Constants.OPTIMIZE_SIEVES_SCORE_PROP, "bcub"); // fall back to default
}

When it happens

Trigger: Setting the property dcoref.optimize.sievesScore (OPTIMIZE_SIEVES_SCORE_PROP) to a metric string not in the valid set — e.g. a misspelled name, wrong case, or a subscore format like 'pairwise.foo' whose first part doesn't match a valid metric — while sieve optimization (optimizeSieves) is enabled.

Common situations: Typos in a coref tuning properties file (e.g. 'bcube' instead of 'bcub'); copying config between CoreNLP versions where the accepted metric list changed; forgetting that combined SubScoreType values must parse via CorefScorer.SubScoreType.valueOf(parts[1]).

Understand the failure class

Background: "Invalid value" and "allowed values are" config errors: what your library rejected and how to fix it — this error's family across 41 libraries.

Related errors

• Invalid ordering constraint
• Cannot have these two ordering constraints
• Cannot have these two ordering constraints
• No input file specified!
• Invalid sieve name

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/d94c11133d0bfdad. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/dcoref/SieveCoreferenceSystem.java:239

    // flag for optimizing sieve ordering
    optimizeSieves = Boolean.parseBoolean(props.getProperty(Constants.OPTIMIZE_SIEVES_PROP, "false"));
    optimizeScoreType = props.getProperty(Constants.OPTIMIZE_SIEVES_SCORE_PROP, "pairwise.Precision");

    // Break down of the optimize score type
    String[] validMetricTypes = { "muc", "pairwise", "bcub", "ceafe", "ceafm", "combined" };
    String[] parts = optimizeScoreType.split("\\.");
    optimizeConllScore = parts.length > 2 && "conll".equalsIgnoreCase(parts[2]);
    optimizeMetricType = parts[0];
    boolean optimizeMetricTypeOk = false;
    for (String validMetricType : validMetricTypes) {
      if (validMetricType.equalsIgnoreCase(optimizeMetricType)) {
        optimizeMetricTypeOk = true;
        break;
      }
    }
    if (!optimizeMetricTypeOk) {
      throw new IllegalArgumentException("Invalid metric type for " +
              Constants.OPTIMIZE_SIEVES_SCORE_PROP + " property: " + optimizeScoreType);
    }
    optimizeSubScoreType = CorefScorer.SubScoreType.valueOf(parts[1]);

    if (optimizeSieves) {
      String keepSieveOrder = props.getProperty(Constants.OPTIMIZE_SIEVES_KEEP_ORDER_PROP);
      if (keepSieveOrder != null) {
        String[] orderings = keepSieveOrder.split("\\s*,\\s*");
        sievesKeepOrder = new ArrayList<>();
        String firstSieveConstraint = null;
        String lastSieveConstraint = null;
        for (String ordering:orderings) {
          // Convert ordering constraints from string
          Pair<Integer,Integer> p = fromSieveOrderConstraintString(ordering, sieveClassNames);
          // Do initial check of sieves order, can only have one where the first is ANY (< 0), and one where second is ANY (< 0)
          if (p.first() < 0 && p.second() < 0) {
            throw new IllegalArgumentException("Invalid ordering constraint: " + ordering);
          } else if (p.first() < 0) {

View on GitHub (pinned to 1b7edd19c4)