stanfordnlp/CoreNLP · error · NullPointerException

fileName , ftbID

Error message

fileName , ftbID 

What it means

FTBDataset.getCanditoTreeID builds the Candito treebank tree ID from a tree's metadata: the source fileName (from CoreLabel docID) and ftbID (SentenceIDAnnotation). It throws NullPointerException('fileName ..., ftbID ...') when either piece of metadata is null after substring extraction, since a valid ID cannot be constructed.

Solutions

1. Load FTB trees with the pipeline's readers so docID and SentenceIDAnnotation are populated
2. Verify every tree leaf has a non-null docID ending in a proper file extension before conversion
3. Null-check the annotations in calling code and skip/log trees missing FTB metadata
4. Set the missing annotations manually if the trees come from another source

Example fix

// before
String id = dataset.canditoTreeID(tree); // NPE when metadata absent
// after
CoreLabel first = (CoreLabel) tree.firstChild().label();
if (first.docID() == null || first.get(CoreAnnotations.SentenceIDAnnotation.class) == null) {
  return; // skip or log tree lacking FTB metadata
}
String id = dataset.canditoTreeID(tree);

CoreLabel leaf = (CoreLabel) tree.firstChild().label();
boolean ready = leaf != null && leaf.docID() != null
  && leaf.get(CoreAnnotations.SentenceIDAnnotation.class) != null;
if (!ready) skipTree(tree); // log and continue

boolean hasFtbMetadata(Tree t){ Label l = t.firstChild().label(); return l instanceof CoreLabel && ((CoreLabel) l).docID() != null && ((CoreLabel) l).get(CoreAnnotations.SentenceIDAnnotation.class) != null; }

try { String id = ds.canditoTreeID(tree); } catch (NullPointerException e) { /* tree lacks FTB metadata; skip */ }

When it happens

Trigger: Calling canditoTreeID(tree) on a tree read from a source whose leaves lack CoreLabel docID or SentenceIDAnnotation — e.g. trees not loaded via the FTB reading pipeline, or a file name without the expected '.' extension making lastIndexOf('.')->substring produce unexpected values.

Common situations: Running the French treebank conversion pipeline on trees created programmatically or loaded from a generic reader that does not populate FTB document/sentence annotations.

Related errors

• Trees constructed without CoreLabels! Can't extract…
• French does not support feature type:
• Unknown minimizer
• Unknown clique: " + clique
• Bad number put into wordToNumber. Word is: \"" + input +…

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/4653041d4b103174. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/international/french/pipeline/FTBDataset.java:74

    //stats for MWE pre-processing
    // The treebank may be reset if setOptions changes CC_TAGSET
    treebank = new MemoryTreebank(new FrenchXMLTreeReaderFactory(CC_TAGSET), FrenchTreebankLanguagePack.FTB_ENCODING);
    treeFileExtension = "xml";
  }

  /**
   * Return the ID of this tree according to the Candito split files.
   */
  private String getCanditoTreeID(Tree t) {
    String canditoName = null;
    if (t.label() instanceof CoreLabel) {
      String fileName = ((CoreLabel) t.label()).docID();
      fileName = fileName.substring(0, fileName.lastIndexOf('.'));
      String ftbID = ((CoreLabel) t.label()).get(CoreAnnotations.SentenceIDAnnotation.class);
      if (fileName != null && ftbID != null) {
        canditoName = fileName + "-" + ftbID;
      } else {
        throw new NullPointerException("fileName " + fileName + ", ftbID " + ftbID);
      }
    } else {
      throw new IllegalArgumentException("Trees constructed without CoreLabels! Can't extract metadata!");
    }
    return canditoName;
  }

  @Override
  public void build() {
    for(File path : pathsToData) {
      treebank.loadPath(path,treeFileExtension,false);
    }

    PrintWriter outfile = null;
    PrintWriter flatFile = null;
    try {
      outfile = new PrintWriter(new BufferedWriter(new OutputStreamWriter(new FileOutputStream(outFileName),"UTF-8")));
      flatFile = (makeFlatFile) ? new PrintWriter(new BufferedWriter(new OutputStreamWriter(new FileOutputStream(flatFileName),"UTF-8"))) : null;

View on GitHub (pinned to 1b7edd19c4)