[LUCENE-8723] Bad interaction bewteen WordDelimiterGraphFilter, StopFilter and FlattenGraphFilter

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 7.7.1, 8.0, 8.3
• Fix Version/s: 9.0, 8.10
• Component/s: modules/analysis
• Lucene Fields: New
• Assignee: Unassigned
• Reporter: Nicolás Lichtmaier
• Created: 12/Mar/19 18:54
• Updated: 28/Aug/22 15:42
• Resolved: 30/Aug/21 16:04

Description

I was debugging an issue (missing tokens after analysis) and when I enabled Java assertions I uncovered a bug when using WordDelimiterGraphFilter + StopFilter + FlattenGraphFilter.

I could reproduce the issue in a small piece of code. This code gives an assertion failure when assertions are enabled (-ea java option):

    Builder builder = CustomAnalyzer.builder();
    builder.withTokenizer(StandardTokenizerFactory.class);
    builder.addTokenFilter(WordDelimiterGraphFilterFactory.class, "preserveOriginal", "1");
    builder.addTokenFilter(StopFilterFactory.class);
    builder.addTokenFilter(FlattenGraphFilterFactory.class);
    Analyzer analyzer = builder.build();
     
    TokenStream ts = analyzer.tokenStream("*", new StringReader("x7in"));
    ts.reset();
    while(ts.incrementToken())
        ;

This gives:

Exception in thread "main" java.lang.AssertionError: 2
     at org.apache.lucene.analysis.core.FlattenGraphFilter.releaseBufferedToken(FlattenGraphFilter.java:195)
     at org.apache.lucene.analysis.core.FlattenGraphFilter.incrementToken(FlattenGraphFilter.java:258)
     at com.wolfram.textsearch.AnalyzerError.main(AnalyzerError.java:32)

Maybe removing stop words after WordDelimiterGraphFilter is wrong, I don't know. However is the only way to process stop-words generated by that filter. In any case, it should not eat tokens or produce assertions. 

Issue Links

• is related to: LUCENE-8740 AssertionError FlattenGraphFilter