Expose DelimitedTermFrequencyTokenFilter

CHANGELOG.md

@@ -89,6 +89,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - [BWC and API enforcement] Define the initial set of annotations, their meaning and relations between them ([#9223](https://github.com/opensearch-project/OpenSearch/pull/9223))
 - [Segment Replication] Support realtime reads for GET requests ([#9212](https://github.com/opensearch-project/OpenSearch/pull/9212))
 - [Feature] Expose term frequency in Painless script score context ([#9081](https://github.com/opensearch-project/OpenSearch/pull/9081))
+- Expose DelimitedTermFrequencyTokenFilter to allow providing term frequencies along with terms ([#9479](https://github.com/opensearch-project/OpenSearch/pull/9479))
 
 ### Dependencies
 - Bump `org.apache.logging.log4j:log4j-core` from 2.17.1 to 2.20.0 ([#8307](https://github.com/opensearch-project/OpenSearch/pull/8307))

modules/analysis-common/src/main/java/org/opensearch/analysis/common/CommonAnalysisModulePlugin.java

@@ -89,6 +89,7 @@
 import org.apache.lucene.analysis.lt.LithuanianAnalyzer;
 import org.apache.lucene.analysis.lv.LatvianAnalyzer;
 import org.apache.lucene.analysis.miscellaneous.ASCIIFoldingFilter;
+import org.apache.lucene.analysis.miscellaneous.DelimitedTermFrequencyTokenFilter;
 import org.apache.lucene.analysis.miscellaneous.DisableGraphAttribute;
 import org.apache.lucene.analysis.miscellaneous.KeywordRepeatFilter;
 import org.apache.lucene.analysis.miscellaneous.LengthFilter;
@@ -265,6 +266,7 @@ public Map<String, AnalysisProvider<TokenFilterFactory>> getTokenFilters() {
         );
         filters.put("decimal_digit", DecimalDigitFilterFactory::new);
         filters.put("delimited_payload", DelimitedPayloadTokenFilterFactory::new);
+        filters.put("delimited_termfreq", DelimitedTermFrequencyTokenFilterFactory::new);
         filters.put("dictionary_decompounder", requiresAnalysisSettings(DictionaryCompoundWordTokenFilterFactory::new));
         filters.put("dutch_stem", DutchStemTokenFilterFactory::new);
         filters.put("edge_ngram", EdgeNGramTokenFilterFactory::new);
@@ -500,6 +502,13 @@ public List<PreConfiguredTokenFilter> getPreConfiguredTokenFilters() {
                 )
             )
         );
+        filters.add(
+            PreConfiguredTokenFilter.singleton(
+                "delimited_termfreq",
+                false,
+                input -> new DelimitedTermFrequencyTokenFilter(input, DelimitedTermFrequencyTokenFilterFactory.DEFAULT_DELIMITER)
+            )
+        );
         filters.add(PreConfiguredTokenFilter.singleton("dutch_stem", false, input -> new SnowballFilter(input, new DutchStemmer())));
         filters.add(PreConfiguredTokenFilter.singleton("edge_ngram", false, false, input -> new EdgeNGramTokenFilter(input, 1)));
         filters.add(PreConfiguredTokenFilter.openSearchVersion("edgeNGram", false, false, (reader, version) -> {

modules/analysis-common/src/main/java/org/opensearch/analysis/common/DelimitedTermFrequencyTokenFilterFactory.java

@@ -0,0 +1,45 @@
+/*
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * The OpenSearch Contributors require contributions made to
+ * this file be licensed under the Apache-2.0 license or a
+ * compatible open source license.
+ */
+
+package org.opensearch.analysis.common;
+
+import org.apache.lucene.analysis.TokenStream;
+import org.apache.lucene.analysis.miscellaneous.DelimitedTermFrequencyTokenFilter;
+import org.opensearch.common.settings.Settings;
+import org.opensearch.env.Environment;
+import org.opensearch.index.IndexSettings;
+import org.opensearch.index.analysis.AbstractTokenFilterFactory;
+
+public class DelimitedTermFrequencyTokenFilterFactory extends AbstractTokenFilterFactory {
+    public static final char DEFAULT_DELIMITER = '|';
+    private static final String DELIMITER = "delimiter";
+    private final char delimiter;
+
+    DelimitedTermFrequencyTokenFilterFactory(IndexSettings indexSettings, Environment environment, String name, Settings settings) {
+        super(indexSettings, name, settings);
+        delimiter = parseDelimiter(settings);
+    }
+
+    @Override
+    public TokenStream create(TokenStream tokenStream) {
+        return new DelimitedTermFrequencyTokenFilter(tokenStream, delimiter);
+    }
+
+    private static char parseDelimiter(Settings settings) throws IllegalArgumentException {
+        String delimiter = settings.get(DELIMITER);
+        if (delimiter == null) {
+            return DEFAULT_DELIMITER;
+        } else if (delimiter.length() == 1) {
+            return delimiter.charAt(0);
+        }
+
+        throw new IllegalArgumentException(
+            "Setting [" + DELIMITER + "] must be a single, non-null character. [" + delimiter + "] was provided."
+        );
+    }
+}

modules/analysis-common/src/test/java/org/opensearch/analysis/common/CommonAnalysisFactoryTests.java

@@ -145,6 +145,7 @@ protected Map<String, Class<?>> getTokenFilters() {
         filters.put("cjkwidth", CJKWidthFilterFactory.class);
         filters.put("cjkbigram", CJKBigramFilterFactory.class);
         filters.put("delimitedpayload", DelimitedPayloadTokenFilterFactory.class);
+        filters.put("delimitedtermfrequency", DelimitedTermFrequencyTokenFilterFactory.class);
         filters.put("keepword", KeepWordFilterFactory.class);
         filters.put("type", KeepTypesFilterFactory.class);
         filters.put("classic", ClassicFilterFactory.class);
@@ -202,6 +203,7 @@ protected Map<String, Class<?>> getPreConfiguredTokenFilters() {
         filters.put("decimal_digit", null);
         filters.put("delimited_payload_filter", org.apache.lucene.analysis.payloads.DelimitedPayloadTokenFilterFactory.class);
         filters.put("delimited_payload", org.apache.lucene.analysis.payloads.DelimitedPayloadTokenFilterFactory.class);
+        filters.put("delimited_termfreq", org.apache.lucene.analysis.miscellaneous.DelimitedTermFrequencyTokenFilterFactory.class);
         filters.put("dutch_stem", SnowballPorterFilterFactory.class);
         filters.put("edge_ngram", null);
         filters.put("edgeNGram", null);

modules/analysis-common/src/yamlRestTest/resources/rest-api-spec/test/analysis-common/40_token_filters.yml

@@ -1198,6 +1198,43 @@
     - match:  { tokens.0.token: foo }
 
 ---
+"delimited_termfreq":
+  - do:
+      indices.create:
+        index: test
+        body:
+          settings:
+            analysis:
+              filter:
+                my_delimited_termfreq:
+                  type: delimited_termfreq
+                  delimiter: ^
+  - do:
+      indices.analyze:
+        index: test
+        body:
+          text:       foo^3
+          tokenizer:  keyword
+          filter:     [my_delimited_termfreq]
+          attributes: termFrequency
+          explain:    true
+  - length: { detail.tokenfilters: 1 }
+  - match:  { detail.tokenfilters.0.tokens.0.token: foo }
+  - match:  { detail.tokenfilters.0.tokens.0.termFrequency: 3 }
+
+  # Test pre-configured token filter too:
+  - do:
+      indices.analyze:
+        body:
+          text:       foo|100
+          tokenizer:  keyword
+          filter:     [delimited_termfreq]
+          attributes: termFrequency
+          explain:    true
+  - length: { detail.tokenfilters: 1 }
+  - match:  { detail.tokenfilters.0.tokens.0.token: foo }
+  - match:  { detail.tokenfilters.0.tokens.0.termFrequency: 100 }
+---
 "keep_filter":
     - do:
         indices.create:

test/framework/src/main/java/org/opensearch/indices/analysis/AnalysisFactoryTestCase.java

@@ -98,6 +98,7 @@ public abstract class AnalysisFactoryTestCase extends OpenSearchTestCase {
         .put("czechstem", MovedToAnalysisCommon.class)
         .put("decimaldigit", MovedToAnalysisCommon.class)
         .put("delimitedpayload", MovedToAnalysisCommon.class)
+        .put("delimitedtermfrequency", MovedToAnalysisCommon.class)
         .put("dictionarycompoundword", MovedToAnalysisCommon.class)
         .put("edgengram", MovedToAnalysisCommon.class)
         .put("elision", MovedToAnalysisCommon.class)
@@ -201,9 +202,6 @@ public abstract class AnalysisFactoryTestCase extends OpenSearchTestCase {
         .put("daterecognizer", Void.class)
         // for token filters that generate bad offsets, which are now rejected since Lucene 7
         .put("fixbrokenoffsets", Void.class)
-        // should we expose it, or maybe think about higher level integration of the
-        // fake term frequency feature (LUCENE-7854)
-        .put("delimitedtermfrequency", Void.class)
         // LUCENE-8273: ProtectedTermFilterFactory allows analysis chains to skip
         // particular token filters based on the attributes of the current token.
         .put("protectedterm", Void.class)