stanfordnlp/CoreNLP · error · RuntimeIOException

Failed to find node " + ie.source + "-" + ie.sourceCopy

Error message

Failed to find node " + ie.source + "-" + ie.sourceCopy

What it means

AnnotationSerializer.convertIntermediateGraph rebuilds a SemanticGraph from serialized intermediate nodes and edges; it throws RuntimeIOException when an edge's source (nodeId-sourceCopy) is absent from the nodeMap, meaning the serialized dependency graph is internally inconsistent.

Solutions

1. Check the chained cause and the serialized file for missing/renumbered nodes
2. Regenerate the serialized annotations with the same (matching) Stanford CoreNLP version used for reading
3. Ensure no custom Annotator/serialization hook removes nodes from the dependency graph without dropping dependent edges
4. If copying documents between serializers, keep nodes and edges together

Example fix

// before
// custom filter removed node 5 but kept its incoming edges
// after
// remove edges whose source/target nodes are removed, or skip node removal entirely

// ensure writer and reader CoreNLP versions match
String expected = "4.5.0";
if (!Version.version().equals(expected))
  log.warn("Version mismatch vs. serialized data: " + Version.version());

try {
  deps = serializer.readAnnotations(...);
} catch (RuntimeIOException e) {
  if (e.getMessage().startsWith("Failed to find node")) {
    log.error("Inconsistent serialized graph: " + e.getMessage());
  } else throw e;
}

When it happens

Trigger: Deserializing annotations where an IntermediateEdge references a source node id/copy that was never added to the node map — e.g. a corrupted or hand-modified serialized file, or nodes dropped by custom serialization filters.

Common situations: Version mismatch between the writing and reading CoreNLP versions changing node id assignment, manual post-processing of serialized output, dependency-conversion filters removing nodes still referenced by edges.

Understand the failure class

Background: "This is a bug, please report it": internal invariant violations, unreachable panics, and SNH errors explained — this error's family across 47 libraries.

Related errors

• Failed to find node " + ie.target + "-" + ie.targetCopy
• ERROR: Invalid dependency node line
• ERROR: Invalid format for dependency graph
• ERROR: Serialized data does not contain an Annotation!
• Unknown general relation

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/8318a1b439587e04. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/pipeline/AnnotationSerializer.java:165

        if (word.sentIndex() < 0 && in.sentIndex >= 0) {
          word.setSentIndex(in.sentIndex);
        }
        if (word.index() < 0 && in.index >= 0) {
          word.setIndex(in.index);
        }

        nodeMap.put(word.index(), word.copyCount(), word);
        graph.addVertex(word);
        if (in.isRoot) {
          graph.addRoot(word);
        }
      }

      // add all edges to the actual graph
      for (IntermediateEdge ie: edges) {
        IndexedWord source = nodeMap.get(ie.source, ie.sourceCopy);
        if (source == null) {
          throw new RuntimeIOException("Failed to find node " + ie.source + "-" + ie.sourceCopy);
        }
        IndexedWord target = nodeMap.get(ie.target, ie.targetCopy);
        if (target == null) {
          throw new RuntimeIOException("Failed to find node " + ie.target + "-" + ie.targetCopy);
        }
        // assert(target != null);
        synchronized (LOCK) {
          // this is not thread-safe: there are static fields in GrammaticalRelation
          GrammaticalRelation rel = GrammaticalRelation.valueOf(ie.dep);
          graph.addEdge(source, target, rel, 1.0, ie.isExtra);
        }
      }

      // compute root nodes if they weren't stored in the graph
      if ( ! graph.isEmpty() && graph.getRoots().isEmpty()) {
        graph.resetRoots();
      }

View on GitHub (pinned to 1b7edd19c4)