Lenient parsing for GroupConfig

Author: davidkyle

client/rest-high-level/src/main/java/org/elasticsearch/client/dataframe/transforms/pivot/GroupConfig.java

@@ -36,13 +36,23 @@ public class GroupConfig implements ToXContentObject {
 
     private final Map<String, SingleGroupSource> groups;
 
+    /**
+     * Leniently parse a {@code GroupConfig}.
+     * Parsing is lenient in that unknown fields in the root of the
+     * object are ignored. Unknown group types {@link SingleGroupSource.Type}
+     * will cause a parsing error.
+     *
+     * @param parser The XContent parser
+     * @return The parsed object
+     * @throws IOException On parsing error
+     */
     public static GroupConfig fromXContent(final XContentParser parser) throws IOException {
         LinkedHashMap<String, SingleGroupSource> groups = new LinkedHashMap<>();
 
         // be parsing friendly, whether the token needs to be advanced or not (similar to what ObjectParser does)
         XContentParser.Token token;
         if (parser.currentToken() == XContentParser.Token.START_OBJECT) {
-            token = parser.currentToken();
+            parser.currentToken();
         } else {
             token = parser.nextToken();
             if (token != XContentParser.Token.START_OBJECT) {
@@ -52,8 +62,16 @@ public static GroupConfig fromXContent(final XContentParser parser) throws IOExc
 
         while ((token = parser.nextToken()) != XContentParser.Token.END_OBJECT) {
             ensureExpectedToken(XContentParser.Token.FIELD_NAME, token, parser::getTokenLocation);
-            String destinationFieldName = parser.currentName();
             token = parser.nextToken();
+            if (token != XContentParser.Token.START_OBJECT) {
+                // leniently skip over key-value and array fields in the root of the object
+                if (token == XContentParser.Token.START_ARRAY) {
+                    parser.skipChildren();
+                }
+                continue;
+            }
+            String destinationFieldName = parser.currentName();
+
             ensureExpectedToken(XContentParser.Token.START_OBJECT, token, parser::getTokenLocation);
             token = parser.nextToken();
             ensureExpectedToken(XContentParser.Token.FIELD_NAME, token, parser::getTokenLocation);
@@ -75,7 +93,7 @@ public static GroupConfig fromXContent(final XContentParser parser) throws IOExc
                 default:
                     throw new ParsingException(parser.getTokenLocation(), "invalid grouping type: " + groupType);
             }
-
+            // destination field end_object
             parser.nextToken();
 
             groups.put(destinationFieldName, groupSource);

client/rest-high-level/src/test/java/org/elasticsearch/client/dataframe/transforms/pivot/GroupConfigTests.java

@@ -19,7 +19,11 @@
 
 package org.elasticsearch.client.dataframe.transforms.pivot;
 
+import org.elasticsearch.common.bytes.BytesArray;
+import org.elasticsearch.common.xcontent.DeprecationHandler;
+import org.elasticsearch.common.xcontent.NamedXContentRegistry;
 import org.elasticsearch.common.xcontent.XContentParser;
+import org.elasticsearch.common.xcontent.json.JsonXContent;
 import org.elasticsearch.test.AbstractXContentTestCase;
 
 import java.io.IOException;
@@ -28,14 +32,16 @@
 import java.util.Map;
 import java.util.Set;
 
+import static org.hamcrest.Matchers.instanceOf;
+
 public class GroupConfigTests extends AbstractXContentTestCase<GroupConfig> {
 
     public static GroupConfig randomGroupConfig() {
         Map<String, SingleGroupSource> groups = new LinkedHashMap<>();
 
         // ensure that the unlikely does not happen: 2 group_by's share the same name
         Set<String> names = new HashSet<>();
-        for (int i = 0; i < randomIntBetween(1, 1); ++i) {
+        for (int i = 0; i < randomIntBetween(1, 4); ++i) {
             String targetFieldName = randomAlphaOfLengthBetween(1, 20);
             if (names.add(targetFieldName)) {
                 SingleGroupSource groupBy;
@@ -72,4 +78,33 @@ protected GroupConfig doParseInstance(XContentParser parser) throws IOException
     protected boolean supportsUnknownFields() {
         return false;
     }
+
+    public void testLenientParsing() throws IOException {
+        BytesArray json = new BytesArray(
+                "{ " +
+                        "\"unknown-field\":\"foo\", " +
+                        "\"destination-field\": {" +
+                            "\"terms\": {" +
+                                "\"field\": \"term-field\"" +
+                            "}" +
+                        "}," +
+                        "\"unknown-field-2\":\"bar\"," +
+                        "\"destination-field2\": {" +
+                            "\"terms\": {" +
+                                "\"field\": \"term-field2\"" +
+                            "}" +
+                        "}," +
+                        "\"array-field\" : [1.0, 2.0]" +
+                "}");
+        XContentParser parser = JsonXContent.jsonXContent
+                .createParser(NamedXContentRegistry.EMPTY, DeprecationHandler.THROW_UNSUPPORTED_OPERATION, json.streamInput());
+
+        GroupConfig gc = GroupConfig.fromXContent(parser);
+
+        assertEquals(gc.getGroups().size(), 2);
+        assertTrue(gc.getGroups().containsKey("destination-field"));
+        SingleGroupSource groupSource = gc.getGroups().get("destination-field");
+        assertThat(groupSource, instanceOf(TermsGroupSource.class));
+        assertEquals(groupSource.getField(), "term-field");
+    }
 }