_analyzer api inconsistency without index name (elastic#115930)

Set positionIncrementGap to default (100) when no index is provided for the _analyze API
jimczi · Nov 18, 2024 · 3253c2a · 3253c2a
1 parent c32b300
commit 3253c2a
Show file tree

Hide file tree

Showing 3 changed files with 46 additions and 8 deletions.
diff --git a/docs/changelog/115930.yaml b/docs/changelog/115930.yaml
@@ -0,0 +1,5 @@
+pr: 115930
+summary:  Inconsistency in the _analyzer api when the index is not included 
+area: Search
+type: bug
+issues: []
diff --git a/server/src/main/java/org/elasticsearch/index/analysis/AnalysisRegistry.java b/server/src/main/java/org/elasticsearch/index/analysis/AnalysisRegistry.java
@@ -189,7 +189,11 @@ public Analyzer getAnalyzer(String analyzer) throws IOException {
                 }
             });
         }
-        return analyzerProvider.get(environment, analyzer).get();
+
+        return overridePositionIncrementGap(
+            (NamedAnalyzer) analyzerProvider.get(environment, analyzer).get(),
+            TextFieldMapper.Defaults.POSITION_INCREMENT_GAP
+        );
     }
 
     @Override
@@ -720,20 +724,22 @@ private static NamedAnalyzer produceAnalyzer(
             throw new IllegalArgumentException("analyzer [" + analyzerFactory.name() + "] created null analyzer");
         }
         NamedAnalyzer analyzer;
-        if (analyzerF instanceof NamedAnalyzer) {
-            // if we got a named analyzer back, use it...
-            analyzer = (NamedAnalyzer) analyzerF;
-            if (overridePositionIncrementGap >= 0 && analyzer.getPositionIncrementGap(analyzer.name()) != overridePositionIncrementGap) {
-                // unless the positionIncrementGap needs to be overridden
-                analyzer = new NamedAnalyzer(analyzer, overridePositionIncrementGap);
-            }
+        if (analyzerF instanceof NamedAnalyzer namedAnalyzer) {
+            analyzer = overridePositionIncrementGap(namedAnalyzer, overridePositionIncrementGap);
         } else {
             analyzer = new NamedAnalyzer(name, analyzerFactory.scope(), analyzerF, overridePositionIncrementGap);
         }
         checkVersions(analyzer);
         return analyzer;
     }
 
+    private static NamedAnalyzer overridePositionIncrementGap(NamedAnalyzer analyzer, int overridePositionIncrementGap) {
+        if (overridePositionIncrementGap >= 0 && analyzer.getPositionIncrementGap(analyzer.name()) != overridePositionIncrementGap) {
+            analyzer = new NamedAnalyzer(analyzer, overridePositionIncrementGap);
+        }
+        return analyzer;
+    }
+
     private static void processNormalizerFactory(
         String name,
         AnalyzerProvider<?> normalizerFactory,

diff --git a/server/src/test/java/org/elasticsearch/action/admin/indices/TransportAnalyzeActionTests.java b/server/src/test/java/org/elasticsearch/action/admin/indices/TransportAnalyzeActionTests.java
@@ -42,6 +42,7 @@
 
 import java.io.IOException;
 import java.io.Reader;
+import java.util.Arrays;
 import java.util.List;
 import java.util.Map;
 
@@ -250,6 +251,32 @@ public void testFillsAttributes() throws IOException {
         assertEquals("<ALPHANUM>", tokens.get(3).getType());
     }
 
+    public void testAnalyzerWithTwoTextsAndNoIndexName() throws IOException {
+        AnalyzeAction.Request request = new AnalyzeAction.Request();
+
+        for (String analyzer : Arrays.asList("standard", "simple", "stop", "keyword", "whitespace", "classic")) {
+            request.analyzer(analyzer);
+            request.text("a a", "b b");
+
+            AnalyzeAction.Response analyzeIndex = TransportAnalyzeAction.analyze(request, registry, mockIndexService(), maxTokenCount);
+            List<AnalyzeAction.AnalyzeToken> tokensIndex = analyzeIndex.getTokens();
+
+            AnalyzeAction.Response analyzeNoIndex = TransportAnalyzeAction.analyze(request, registry, null, maxTokenCount);
+            List<AnalyzeAction.AnalyzeToken> tokensNoIndex = analyzeNoIndex.getTokens();
+
+            assertEquals(tokensIndex.size(), tokensNoIndex.size());
+            for (int i = 0; i < tokensIndex.size(); i++) {
+                AnalyzeAction.AnalyzeToken withIndex = tokensIndex.get(i);
+                AnalyzeAction.AnalyzeToken withNoIndex = tokensNoIndex.get(i);
+
+                assertEquals(withIndex.getStartOffset(), withNoIndex.getStartOffset());
+                assertEquals(withIndex.getEndOffset(), withNoIndex.getEndOffset());
+                assertEquals(withIndex.getPosition(), withNoIndex.getPosition());
+                assertEquals(withIndex.getType(), withNoIndex.getType());
+            }
+        }
+    }
+
     public void testWithIndexAnalyzers() throws IOException {
         AnalyzeAction.Request request = new AnalyzeAction.Request();
         request.text("the quick brown fox");