More doc and test

apache · Nov 9, 2023 · 75eb903 · 75eb903
1 parent 61f92fd
commit 75eb903
Show file tree

Hide file tree

Showing 4 changed files with 22 additions and 2 deletions.
diff --git a/lucene/CHANGES.txt b/lucene/CHANGES.txt
@@ -72,6 +72,8 @@ New Features
 
 * LUCENE-10010 Introduce NFARunAutomaton to run NFA directly. (Patrick Zhai)
 
+* GITHUB-12767: Add a flag to enable executing using NFA in RegexpQuery. (Patrick Zhai)
+
 * LUCENE-10626 Hunspell: add tools to aid dictionary editing:
   analysis introspection, stem expansion and stem/flag suggestion (Peter Gromov)
 

diff --git a/lucene/core/src/java/org/apache/lucene/search/RegexpQuery.java b/lucene/core/src/java/org/apache/lucene/search/RegexpQuery.java
@@ -158,7 +158,10 @@ public RegexpQuery(
    * @param doDeterminization whether do determinization to force the query to use DFA as
    *     runAutomaton, if false, the query will not try to determinize the generated automaton from
    *     regexp such that it might or might not be a DFA. In case it is an NFA, the query will
-   *     eventually use {@link org.apache.lucene.util.automaton.NFARunAutomaton} to execute.
+   *     eventually use {@link org.apache.lucene.util.automaton.NFARunAutomaton} to execute. Notice
+   *     that {@link org.apache.lucene.util.automaton.NFARunAutomaton} is not thread-safe, so better
+   *     to avoid rewritten method like {@link #CONSTANT_SCORE_BLENDED_REWRITE} when searcher is
+   *     configured with an executor service
    */
   public RegexpQuery(
       Term term,

diff --git a/lucene/core/src/java/org/apache/lucene/util/automaton/NFARunAutomaton.java b/lucene/core/src/java/org/apache/lucene/util/automaton/NFARunAutomaton.java
@@ -25,9 +25,11 @@
 
 /**
  * A RunAutomaton that does not require DFA. It will lazily determinize on-demand, memorizing the
- * generated DFA states that has been explored
+ * generated DFA states that has been explored. Note: the current implementation is NOT thread-safe
  *
  * <p>implemented based on: https://swtch.com/~rsc/regexp/regexp1.html
+ *
+ * @lucene.internal
  */
 public class NFARunAutomaton implements ByteRunnable, TransitionAccessor {
 

diff --git a/lucene/core/src/test/org/apache/lucene/search/TestRegexpRandom2.java b/lucene/core/src/test/org/apache/lucene/search/TestRegexpRandom2.java
@@ -52,6 +52,7 @@
 public class TestRegexpRandom2 extends LuceneTestCase {
   protected IndexSearcher searcher1;
   protected IndexSearcher searcher2;
+  protected IndexSearcher searcher3;
   private IndexReader reader;
   private Directory dir;
   protected String fieldName;
@@ -95,6 +96,7 @@ public void setUp() throws Exception {
     reader = writer.getReader();
     searcher1 = newSearcher(reader);
     searcher2 = newSearcher(reader);
+    searcher3 = newSearcher(reader);
     writer.close();
   }
 
@@ -172,11 +174,22 @@ public void testRegexps() throws Exception {
   /** check that the # of hits is the same as from a very simple regexpquery implementation. */
   protected void assertSame(String regexp) throws IOException {
     RegexpQuery smart = new RegexpQuery(new Term(fieldName, regexp), RegExp.NONE);
+    RegexpQuery nfaQuery =
+        new RegexpQuery(
+            new Term(fieldName, regexp),
+            RegExp.NONE,
+            0,
+            RegexpQuery.DEFAULT_PROVIDER,
+            0,
+            MultiTermQuery.CONSTANT_SCORE_BLENDED_REWRITE,
+            false);
     DumbRegexpQuery dumb = new DumbRegexpQuery(new Term(fieldName, regexp), RegExp.NONE);
 
     TopDocs smartDocs = searcher1.search(smart, 25);
     TopDocs dumbDocs = searcher2.search(dumb, 25);
+    TopDocs nfaDocs = searcher3.search(nfaQuery, 25);
 
     CheckHits.checkEqual(smart, smartDocs.scoreDocs, dumbDocs.scoreDocs);
+    CheckHits.checkEqual(nfaQuery, nfaDocs.scoreDocs, dumbDocs.scoreDocs);
   }
 }