stanfordnlp/CoreNLP · error · RuntimeException

Quotes size and gold size don't match!

Error message

Quotes size and gold size don't match!

What it means

XMLToAnnotation.readXMLFormat parses a quote-attribution XML corpus into quotes and a gold author list, then verifies counts match. If document.get(CoreAnnotations.QuotationsAnnotation.class).size() != goldList.size() it throws RuntimeException("Quotes size and gold size don't match!") at src/edu/stanford/nlp/quoteattribution/XMLToAnnotation.java:307. The check exists because gold entries must map 1:1 onto extracted quotes for training/evaluation Data.

Solutions

1. Validate the XML quote markup (properly nested, well-formed <quote> elements) and re-run.
2. Log both sizes and diff which quote index diverged; fix the offending paragraph.
3. Regenerate the gold list from the same XML file version used for parsing.
4. Escape/normalize special characters and whitespace so the quote tokenizer sees the same quotes as the annotator.

Example fix

// before
Data d = new XMLToAnnotation(pipeline).readXMLFormat("book.xml"); // mismatch
// after (pre-validate)
Document dom = ...; int goldCount = dom.getElementsByTagName("author").getLength();
int quoteCount = dom.getElementsByTagName("quote").getLength();
if (goldCount != quoteCount) throw new IllegalArgumentException("fix XML: quotes=" + quoteCount + " gold=" + goldCount);
Data d = new XMLToAnnotation(pipeline).readXMLFormat("book.xml");

Document dom = DocumentBuilderFactory.newInstance().newDocumentBuilder().parse(xml);
int q = dom.getElementsByTagName("quote").getLength();
int g = dom.getElementsByTagName("author").getLength();
if (q != g) throw new IllegalArgumentException("XML quotes/gold mismatch: " + q + " vs " + g);

try {
  Data d = reader.readXMLFormat(file);
} catch (RuntimeException e) {
  if (e.getMessage().contains("don't match")) log.severe("Fix XML quote markup in " + file);
  throw e;
}

When it happens

Trigger: Calling readXMLFormat on an XML file where the number of quote elements successfully parsed into the document differs from the number of gold author entries — e.g. nested/malformed quote tags, a quote element missing required attributes, or gold entries added for text that was not recognized as a quote.

Common situations: Hand-edited or third-party XML corpora with inconsistent quote markup; XML entities or line breaks breaking quote detection; gold annotation files not regenerated after editing the text.

Understand the failure class

Background: Schema validation failed / invalid input schema: payload rejected because its shape doesn't match the expected schema — this error's family across 28 libraries.

Related errors

• LogisticClassifier is only for binary classification!
• We need at least 2 extractors for ExtractorMerger to make se
• Too many columns: <columnI>/<numColumns> (offset: <offset>)
• Too few columns: <columnI>/<numColumns> (offset: <offset>)
• Sentence.toSentence: lengths differ

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/7963847612e18d95. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/quoteattribution/XMLToAnnotation.java:307

            }
            tokenIndex = getEndIndex(tokenIndex, tokens, nodeText) + 1;
          }
        }
      }
    }
    for(Pair<Integer, String> item : mentionIdToSpeakerList) {
      Mention mention = idToMention.get(item.first);
      if(mention == null) {
        goldList.add(new GoldQuoteInfo(-1, -1, item.second, null));
      } else {
        goldList.add(new GoldQuoteInfo(mention.begin, mention.end, item.second, mention.text));
      }

    }

    //verify
    if(document.get(CoreAnnotations.QuotationsAnnotation.class).size() != goldList.size()) {
      throw new RuntimeException("Quotes size and gold size don't match!");
    }

    return new Data(goldList, personList, document);
  }
}

View on GitHub (pinned to 1b7edd19c4)