stanfordnlp/CoreNLP · warning

Error extracting annotation from

Error message

Error extracting annotation from ${te}

What it means

CoreMapExpressionExtractor.annotateExpressions runs TokensRegex extraction over expressions. When an expression fails to extract (a rule match produced nothing usable, or an exception was thrown), the expression is discarded and a warning naming the offending CoreMap is logged. Extraction continues with remaining expressions.

Solutions

  1. Enable verbose mode to also see the underlying exception for failed expressions.
  2. Inspect the failing expression (te) and the rule that processed it; validate the pattern against the actual CoreMap annotations.
  3. Fix or guard the rule (e.g. check required annotations exist) so extraction succeeds or skips intentionally.
  4. Test rules incrementally with TokensRegex debugging tools.

Example fix

// before
props.setProperty("verbose", "false"); // root exception hidden
// after
props.setProperty("verbose", "true"); // logs the ex for the failing expression
Defensive patterns

Strategy: try-catch

Validate before calling

// verify required annotations exist before extraction
for (CoreMap te : expressions) {
  if (te.get(CoreAnnotations.TokensAnnotation.class) == null) {
    throw new IllegalStateException("expression missing tokens: " + te);
  }
}

Try / catch

try {
  extractor.extractExpressions(doc);
} catch (Exception ex) {
  log.warn("expression extraction failed for document", ex);
  // continue with un-enriched annotations
}

Prevention

When it happens

Trigger: applyCompositeRule/extractExpressions encountering an expression where a rule match fails or a sub-expression throws; exceptions are logged (with the same message) only when verbose=true.

Common situations: Custom TokensRegex rules with patterns that don't match the actual annotation structure; rules referencing missing annotation keys; malformed rule files producing null results.

Related errors

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/1aa19371baf05044. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/ling/tokensregex/CoreMapExpressionExtractor.java:528

  private void annotateExpressions(CoreMap annotation, List<T> expressions) {
    // TODO: Logging can be excessive
    List<T> toDiscard = new ArrayList<>();
    for (T te : expressions) {
      // Add attributes and all
      if (te.annotation == null) {
        try {
          boolean extractOkay = te.extractAnnotation(env, annotation);
          if (verbose && extractOkay) {
            log.info("annotateExpressions() matched " + te + " from " + annotation);
          }
          if (!extractOkay) {
            // Things didn't turn out so well
            toDiscard.add(te);
            log.warn("Error extracting annotation from " + te /*+ ", " + te.getExtractErrorMessage() */);
          }
        } catch (Exception ex) {
          if (verbose) {
            log.warn("Error extracting annotation from " + te);
            log.warn(ex);
          }
        }
      }
    }
    expressions.removeAll(toDiscard);
  }

  private void annotateExpressions(List<? extends CoreMap> chunks, List<T> expressions) {
    // TODO: Logging can be excessive
    List<T> toDiscard = new ArrayList<>();
    for (T te : expressions) {
      // Add attributes and all
      try {
        boolean extractOkay = te.extractAnnotation(env, chunks);
        if (verbose && extractOkay) {
          log.info("annotateExpressions() matched " + te + " from " + chunks);
        }

View on GitHub (pinned to 1b7edd19c4)