stanfordnlp/CoreNLP · error · RuntimeException

Bad pattern

Error message

Bad pattern: <pattern>

What it means

During GrammaticalRelation construction, each target pattern string is compiled with the Tregex compiler. If a pattern is not valid Tregex syntax, TregexParseException is caught and rethrown as RuntimeException('Bad pattern: ' + pattern, pe). This means one of the relation's target tregex expressions cannot be parsed.

Solutions

1. Fix the tregex pattern syntax reported in the message; verify against the Tregex grammar for your library version
2. Quote regex label values inside patterns (e.g. /^NN/ usage rules) and balance all brackets/angle brackets
3. Compile the pattern standalone with a TregexPatternCompiler to get the full TregexParseException detail (also in the cause chain)
4. Check release notes for tregex syntax changes if the relation worked on an older CoreNLP version

Example fix

// before
targetPatterns = new String[]{ "VP < (NP $-- ADJP" }; // unbalanced paren
// after
targetPatterns = new String[]{ "VP < (NP $-- ADJP)" };

TregexPatternCompiler compiler = new TregexPatternCompiler(headFinder);
for (String p : targetPatterns) {
    try { compiler.compile(p); } catch (TregexParseException e) {
        throw new IllegalArgumentException("Invalid tregex: " + p, e);
    }
}

try {
    GrammaticalRelation r = new GrammaticalRelation(lang, name, longName, src, Arrays.asList(targetPatterns), ...);
} catch (RuntimeException e) {
    if (e.getMessage().startsWith("Bad pattern:")) {
        TregexParseException cause = (TregexParseException) e.getCause(); // inspect exact syntax error
    } else throw e;
}

When it happens

Trigger: Constructing a GrammaticalRelation whose targetPatterns list contains a string that fails Tregex parsing (bad node descriptions, unbalanced brackets, unsupported operators, missing quotes around regex labels).

Common situations: Writing custom relation definitions with hand-written tregex; version changes where tregex grammar became stricter; typos such as unbalanced '>' or '[ ]' in patterns like 'NP < NN'.

Related errors

• Bad pattern
• Coordination node must have at least 2 children.
• Expected lines of the format original (tab) replacement. …
• Node cannot be both negated and optional.
• Unknown multi relation

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/f9c3318e2f7a942f. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/trees/GrammaticalRelation.java:307

      parent.addChild(this);
    }

    if (sourcePattern != null) {
      try {
        this.sourcePattern = Pattern.compile(sourcePattern);
      } catch (java.util.regex.PatternSyntaxException e) {
        throw new RuntimeException("Bad pattern: " + sourcePattern);
      }
    } else {
      this.sourcePattern = null;
    }

    for (String pattern : targetPatterns) {
      try {
        TregexPattern p = tregexCompiler.compile(pattern);
        this.targetPatterns.add(p);
      } catch (edu.stanford.nlp.trees.tregex.TregexParseException pe) {
        throw new RuntimeException("Bad pattern: " + pattern, pe);
      }
    }

    GrammaticalRelation previous;
    synchronized (stringsToRelations) {
      Map<String, GrammaticalRelation> sToR = stringsToRelations.get(language);
      if (sToR == null) {
        sToR = Generics.newHashMap();
        stringsToRelations.put(language, sToR);
      }
      previous = sToR.put(toString(), this);
    }
    if (previous != null) {
      if ( ! previous.isFromString() && ! isFromString()) {
        throw new IllegalArgumentException("There is already a relation named " + this + '!');
      } else {
        /* We get here if we previously just built a fake relation from a string
         * we previously read in from a file.

View on GitHub (pinned to 1b7edd19c4)