[ML] Support the unsigned_long type in data frame analytics

Adds support for the unsigned_long type to data frame analytics.

This type is handled in the same way as the long type.  Values
sent to the ML native processes are converted to floats and
hence will lose accuracy when outside the range where a float
can uniquely represent long values.

Relates elastic#60050

Author: droberts195

x-pack/plugin/core/src/main/java/org/elasticsearch/xpack/core/ml/dataframe/analyses/Types.java

@@ -28,12 +28,13 @@ private Types() {}
             .collect(Collectors.toUnmodifiableSet());
 
     private static final Set<String> NUMERICAL_TYPES =
-        Stream.concat(Stream.of(NumberType.values()).map(NumberType::typeName), Stream.of("scaled_float"))
+        Stream.concat(Stream.of(NumberType.values()).map(NumberType::typeName), Stream.of("scaled_float", "unsigned_long"))
             .collect(Collectors.toUnmodifiableSet());
 
     private static final Set<String> DISCRETE_NUMERICAL_TYPES =
-        Stream.of(NumberType.BYTE, NumberType.SHORT, NumberType.INTEGER, NumberType.LONG)
-            .map(NumberType::typeName)
+        Stream.concat(
+            Stream.of(NumberType.BYTE, NumberType.SHORT, NumberType.INTEGER, NumberType.LONG).map(NumberType::typeName),
+            Stream.of("unsigned_long"))
             .collect(Collectors.toUnmodifiableSet());
 
     private static final Set<String> BOOL_TYPES = Collections.singleton(BooleanFieldMapper.CONTENT_TYPE);

x-pack/plugin/ml/qa/native-multi-node-tests/src/javaRestTest/java/org/elasticsearch/xpack/ml/integration/DataFrameAnalysisCustomFeatureIT.java

@@ -167,7 +167,7 @@ private static void createIndex(String index, boolean isDatastream) {
             "          \"type\": \"double\"\n" +
             "        }," +
             "        \""+ DISCRETE_NUMERICAL_FIELD + "\": {\n" +
-            "          \"type\": \"integer\"\n" +
+            "          \"type\": \"unsigned_long\"\n" +
             "        }," +
             "        \""+ TEXT_FIELD + "\": {\n" +
             "          \"type\": \"text\"\n" +

x-pack/plugin/ml/qa/native-multi-node-tests/src/javaRestTest/java/org/elasticsearch/xpack/ml/integration/ExplainDataFrameAnalyticsIT.java

@@ -52,7 +52,7 @@ public void testSourceQueryIsApplied() throws IOException {
         client().admin().indices().prepareCreate(sourceIndex)
             .setMapping(
                 "numeric_1", "type=double",
-                "numeric_2", "type=float",
+                "numeric_2", "type=unsigned_long",
                 "categorical", "type=keyword",
                 "filtered_field", "type=keyword")
             .get();
@@ -64,7 +64,7 @@ public void testSourceQueryIsApplied() throws IOException {
             IndexRequest indexRequest = new IndexRequest(sourceIndex);
             indexRequest.source(
                 "numeric_1", 1.0,
-                "numeric_2", 2.0,
+                "numeric_2", 2,
                 "categorical", i % 2 == 0 ? "class_1" : "class_2",
                 "filtered_field", i < 2 ? "bingo" : "rest"); // We tag bingo on the first two docs to ensure we have 2 classes
             bulkRequestBuilder.add(indexRequest);

x-pack/plugin/ml/qa/native-multi-node-tests/src/javaRestTest/java/org/elasticsearch/xpack/ml/integration/RegressionIT.java

@@ -698,7 +698,7 @@ static void indexData(String sourceIndex, int numTrainingRows, int numNonTrainin
             "          \"type\": \"double\"\n" +
             "        }," +
             "        \"" + DISCRETE_NUMERICAL_FEATURE_FIELD + "\": {\n" +
-            "          \"type\": \"long\"\n" +
+            "          \"type\": \"unsigned_long\"\n" +
             "        }," +
             "        \"" + DEPENDENT_VARIABLE_FIELD + "\": {\n" +
             "          \"type\": \"double\"\n" +

x-pack/plugin/ml/qa/native-multi-node-tests/src/javaRestTest/java/org/elasticsearch/xpack/ml/integration/RunDataFrameAnalyticsIT.java

@@ -71,7 +71,7 @@ public void testOutlierDetectionWithFewDocuments() throws Exception {
         String sourceIndex = "test-outlier-detection-with-few-docs";
 
         client().admin().indices().prepareCreate(sourceIndex)
-            .setMapping("numeric_1", "type=double", "numeric_2", "type=float", "categorical_1", "type=keyword")
+            .setMapping("numeric_1", "type=double", "numeric_2", "type=unsigned_long", "categorical_1", "type=keyword")
             .get();
 
         BulkRequestBuilder bulkRequestBuilder = client().prepareBulk();
@@ -83,7 +83,7 @@ public void testOutlierDetectionWithFewDocuments() throws Exception {
             // We insert one odd value out of 5 for one feature
             String docId = i == 0 ? "outlier" : "normal" + i;
             indexRequest.id(docId);
-            indexRequest.source("numeric_1", i == 0 ? 100.0 : 1.0, "numeric_2", 1.0, "categorical_1", "foo_" + i);
+            indexRequest.source("numeric_1", i == 0 ? 100.0 : 1.0, "numeric_2", 1, "categorical_1", "foo_" + i);
             bulkRequestBuilder.add(indexRequest);
         }
         BulkResponse bulkResponse = bulkRequestBuilder.get();