stanfordnlp/CoreNLP · error · RuntimeException

Found a non-empty line in a tsurgeon section after reading…

Error message

Found a non-empty line in a tsurgeon section after reading the operation

What it means

GenerateTrees.readGrammar reads a grammar file divided into sections (TSURGEON, TERMINALS, etc.). In the TSURGEON section, exactly one operation is expected per entry; any additional non-empty line after the operation has been read means the file is malformed, so it throws with the offending line context.

Solutions

Edit the grammar file so each tsurgeon operation is separated properly (blank line / next section marker) and no extra non-empty lines follow an operation.
Split multiple tsurgeon operations into separate operation entries in the section.
Check for hidden characters or missing newlines around the tsurgeon block reported by the stack trace.

Example fix

// before
TSURGEON
delete NP-1
move NP > NP
// after
TSURGEON
delete NP-1

TSURGEON
move NP > NP

Defensive patterns

Strategy: validation

Validate before calling

// parse the grammar file first and assert each TSURGEON block has exactly one operation before calling readGrammar
assertEachTsurgeonBlockHasOneOp(grammarFile);

Try / catch

try { g.readGrammar(reader); } catch (RuntimeException e) { if (e.getMessage().contains("non-empty line in a tsurgeon section")) { throw new GrammarFormatException("fix TSURGEON section: " + e.getMessage(), e); } }

Prevention

Separate tsurgeon operations with blank lines / section markers
Lint grammar files for stray lines inside TSURGEON sections
Keep one operation per TSURGEON entry

When it happens

Trigger: A tsurgeon section in the grammar file contains more than one operation's worth of non-empty lines — e.g. two operations without a proper section/blank-line separator, or stray text after the tsurgeon command.

Common situations: Hand-edited grammar files where blank-line separators between tsurgeon operations were removed; concatenating tsurgeon scripts without the expected section formatting; copy-paste artifacts leaving extra lines in the TSURGEON block.

Understand the failure class

Background: "Invalid ... format", "must be in format X", "does not look like a ..." — invalid argument format errors across CLI tools and libraries — this error's family across 17 libraries.

Related errors

AI-assisted analysis of stanfordnlp/CoreNLP@1b7edd19c4 (2026-09-10). Data as JSON: /api/errors/32e39d72a93cccd4. Report an issue: GitHub.

Appendix: source

Thrown at src/edu/stanford/nlp/trees/GenerateTrees.java:134

        try {
          Section newSection = Section.valueOf(line.toUpperCase(Locale.ROOT));
          section = newSection;
          if (section == Section.TSURGEON) {
            // this will tregex pattern until it has eaten a blank
            // line, then read tsurgeon until it has eaten another
            // blank line.
            Pair<TregexPattern, TsurgeonPattern> operation = Tsurgeon.getOperationFromReader(bin, compiler);
            tsurgeons.add(operation);
          }
          continue;
        } catch (IllegalArgumentException e) {
          // never mind, not an enum
        }
        
        String[] pieces = line.split(" +");
        switch(section) {
        case TSURGEON: {
          throw new RuntimeException("Found a non-empty line in a tsurgeon section after reading the operation");
        }
        case TERMINALS: {
          Counter<String> productions = terminals.get(pieces[0]);
          if (productions == null) {
            productions = new ClassicCounter<>();
            terminals.put(pieces[0], productions);
          }
          for (int i = 1; i < pieces.length; ++i) {
            productions.incrementCount(pieces[i]);
          }
          break;
        }
        case NONTERMINALS: {
          Counter<List<String>> productions = nonTerminals.get(pieces[0]);
          if (productions == null) {
            productions = new ClassicCounter<>();
            nonTerminals.put(pieces[0], productions);
          }

View on GitHub (pinned to 1b7edd19c4)