KoreanTokenizer with outputUnknownUnigrams=true emits every punctuation, symbol and whitespace character inside an unknown word as its own token ('$', '@', '.', ' ', '\t', '\n') although discardPunctuation=true (the default): LUCENE-8784 (8.2.0) dropped the punctuation check from the unigram branch
Finding LFV-444Area: analysis / nori KoreanTokenizer (Viterbi.backtrace unigram branch; KoreanTokenizerFactory, KoreanAnalyzer)Priority: mediumImpact: silent wrong resultsNovelty: new
AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.
What it shows
You analyze '가격: $100' or '다음\n\n문장' with KoreanTokenizer(factory, null, DISCARD, outputUnknownUnigrams=true, discardPunctuation=true), or with the 'korean' factory given only outputUnknownUnigrams="true", and expect punctuation and whitespace to be dropped, as the discardPunctuation javadoc says ('true if punctuation tokens should be dropped from the output') and as happens with outputUnknownUnigrams=false ([가격, 100], [다음, 문장]) and in 7.4.0-8.1.1. Instead you get [가격, $, 1, 0, 0] and [다음, \n, \n, 문장], and 'test@example.com' keeps its '@' and '.': the unigram branch of nori's Viterbi.backtrace adds each single-character token of an unknown word without the shouldFilterToken check that the other branch applies. Every such character becomes an index term and pushes the following words one position further, so an exact phrase across it no longer matches.
The test shows that KoreanTokenizer with outputUnknownUnigrams=true and discardPunctuation=true keeps the $ of 가격: $100 as a token (the unigram branch of Viterbi.backtrace skips the shouldFilterToken check).
Expected vs actual
Expected, per the discardPunctuation javadoc and the outputUnknownUnigrams=false output: [가격, 1, 0, 0]. Actual: [가격, $, 1, 0, 0] (term 1 expected:<[1]> but was:<[$]>).
Files
Run it
- Tested on Lucene main e1faeba08bde (2026-11-02): still reproduces.
- Tested on Lucene main 645d4fe14a20 (2026-10-28): reproduced.
- Recorded outcome of run 1 on Lucene main 645d4fe14a20: 1 test(s) run, 1 failed.
- How to read the outcome: a failing test is the finding reproducing (the failure message states expected against actual).
- Seeds: a rerun with the same seed replays the same random choices. Assertions off (-da -dsa -Dtests.asserts=false) is the production-like setting.
# Lucene checkout at the tested commit, built with: ./gradlew --no-daemon assemble (JDK 25)
LUCENE=/path/to/lucene; V=12.0.0-SNAPSHOT
CP=$LUCENE/lucene/core/build/libs/lucene-core-$V.jar
CP=$CP:$LUCENE/lucene/codecs/build/libs/lucene-codecs-$V.jar
CP=$CP:$LUCENE/lucene/test-framework/build/libs/lucene-test-framework-$V.jar
CP=$CP:$LUCENE/lucene/analysis/nori/build/libs/lucene-analysis-nori-$V.jar
CP=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar # test-framework dependencies (Maven Central)
CP=$CP:junit-platform-console-standalone-6.1.3.jar # JUnit Platform console launcher (Maven Central)
javac -d classes -cp "$CP" $(find src -name '*.java')
# run 1: assertions on, seed DEADBEEF
java -ea -esa --add-modules jdk.incubator.vector --enable-native-access=ALL-UNNAMED -Dtests.seed=DEADBEEF \
-cp "classes:$CP" org.junit.platform.console.ConsoleLauncher \
execute --select-class org.apache.lucene.analysis.ko.TestKoreanUnigramsDiscardPunctuationThe module(s) analysis-nori also need their third-party runtime jars; ./gradlew -p lucene/<module> dependencies lists them.
Gradle alternative: copy the test into the matching module's src/test/ tree of the checkout and run ./gradlew -p lucene/<module> test --tests <class> -Ptests.seed=<seed> (add -Ptests.asserts=false for a run without assertions).
Feedback
Open an issue in lucene-ai-bug-hunt. Please do not open issues or pull requests in apache/lucene about this finding unless a Lucene committer agrees it should be filed.
Snapshot of 2026-11-03. AI-assisted; not yet triaged by a human Lucene committer. Feedback: open an issue in lucene-ai-bug-hunt.