Implement URI parts ingest processor (#5304)

Author: stevejgordon

src/Nest/Ingest/ProcessorFormatter.cs

@@ -46,6 +46,7 @@ internal class ProcessorFormatter : IJsonFormatter<IProcessor>
 			{ "circle", 30 },
 			{ "enrich", 31 },
 			{ "csv", 32 },
+			{ "uri_parts", 33 }
 		};
 
 		public IProcessor Deserialize(ref JsonReader reader, IJsonFormatterResolver formatterResolver)
@@ -165,6 +166,9 @@ public IProcessor Deserialize(ref JsonReader reader, IJsonFormatterResolver form
 					case 32:
 						processor = Deserialize<CsvProcessor>(ref reader, formatterResolver);
 						break;
+					case 33:
+						processor = Deserialize<UriPartsProcessor>(ref reader, formatterResolver);
+						break;
 				}
 			}
 			else
@@ -285,6 +289,9 @@ public void Serialize(ref JsonWriter writer, IProcessor value, IJsonFormatterRes
 				case "circle":
 					Serialize<ICircleProcessor>(ref writer, value, formatterResolver);
 					break;
+				case "uri_parts":
+					Serialize<IUriPartsProcessor>(ref writer, value, formatterResolver);
+					break;
 				default:
 					var formatter = DynamicObjectResolver.ExcludeNullCamelCase.GetFormatter<IProcessor>();
 					formatter.Serialize(ref writer, value, formatterResolver);

src/Nest/Ingest/Processors/UriPartsProcessor.cs

@@ -0,0 +1,97 @@
+// Licensed to Elasticsearch B.V under one or more agreements.
+// Elasticsearch B.V licenses this file to you under the Apache 2.0 License.
+// See the LICENSE file in the project root for more information
+
+using System;
+using System.Linq.Expressions;
+using System.Runtime.Serialization;
+using Elasticsearch.Net.Utf8Json;
+
+namespace Nest
+{
+	[InterfaceDataContract]
+	public interface IUriPartsProcessor : IProcessor
+	{
+		/// <summary>
+		/// The field to decode
+		/// </summary>
+		[DataMember(Name = "field")]
+		Field Field { get; set; }
+
+		/// <summary>
+		/// If <c>true</c> the processor copies the unparsed URI to <target_field>.original.
+		/// </summary>
+		[DataMember(Name = "keep_original")]
+		bool? KeepOriginal { get; set; }
+
+		/// <summary>
+		/// The field to assign the converted value to, by default <see cref="Field" /> is updated in-place
+		/// </summary>
+		[DataMember(Name = "target_field")]
+		Field TargetField { get; set; }
+
+		/// <summary>
+		/// If <c>true</c> the processor removes the <see cref="Field" /> after parsing the URI string.
+		/// If parsing fails, the processor does not remove the <see cref="Field" />.
+		/// </summary>
+		[DataMember(Name = "remove_if_successful")]
+		bool? RemoveIfSuccessful { get; set; }
+	}
+
+	/// <inheritdoc cref="IUriPartsProcessor" />
+	public class UriPartsProcessor : ProcessorBase, IUriPartsProcessor
+	{
+		/// <inheritdoc />
+		[DataMember(Name = "field")]
+		public Field Field { get; set; }
+
+		/// <inheritdoc />
+		[DataMember(Name = "keep_original")]
+		public bool? KeepOriginal { get; set; }
+
+		/// <inheritdoc />
+		[DataMember(Name = "target_field")]
+		public Field TargetField { get; set; }
+
+		/// <inheritdoc />
+		[DataMember(Name = "remove_if_successful")]
+		public bool? RemoveIfSuccessful { get; set; }
+
+		protected override string Name => "uri_parts";
+	}
+
+	/// <inheritdoc cref="IUriPartsProcessor" />
+	public class UriPartsProcessorDescriptor<T>
+		: ProcessorDescriptorBase<UriPartsProcessorDescriptor<T>, IUriPartsProcessor>, IUriPartsProcessor
+		where T : class
+	{
+		protected override string Name => "uri_parts";
+
+		Field IUriPartsProcessor.Field { get; set; }
+		bool? IUriPartsProcessor.KeepOriginal { get; set; }
+		Field IUriPartsProcessor.TargetField { get; set; }
+		bool? IUriPartsProcessor.RemoveIfSuccessful { get; set; }
+
+		/// <inheritdoc cref="IUriPartsProcessor.Field" />
+		public UriPartsProcessorDescriptor<T> Field(Field field) => Assign(field, (a, v) => a.Field = v);
+
+		/// <inheritdoc cref="IUriPartsProcessor.Field" />
+		public UriPartsProcessorDescriptor<T> Field<TValue>(Expression<Func<T, TValue>> objectPath) =>
+			Assign(objectPath, (a, v) => a.Field = v);
+
+		/// <inheritdoc cref="IUriPartsProcessor.TargetField" />
+		public UriPartsProcessorDescriptor<T> TargetField(Field field) => Assign(field, (a, v) => a.TargetField = v);
+
+		/// <inheritdoc cref="IUriPartsProcessor.TargetField" />
+		public UriPartsProcessorDescriptor<T> TargetField<TValue>(Expression<Func<T, TValue>> objectPath) =>
+			Assign(objectPath, (a, v) => a.TargetField = v);
+
+		/// <inheritdoc cref="IUriPartsProcessor.KeepOriginal" />
+		public UriPartsProcessorDescriptor<T> KeepOriginal(bool? keepOriginal = true) =>
+			Assign(keepOriginal, (a, v) => a.KeepOriginal = v);
+
+		/// <inheritdoc cref="IUriPartsProcessor.RemoveIfSuccessful" />
+		public UriPartsProcessorDescriptor<T> RemoveIfSuccessful(bool? removeIfSuccessful = true) =>
+			Assign(removeIfSuccessful, (a, v) => a.RemoveIfSuccessful = v);
+	}
+}

src/Nest/Ingest/ProcessorsDescriptor.cs

@@ -177,5 +177,9 @@ public ProcessorsDescriptor Pipeline(Func<PipelineProcessorDescriptor, IPipeline
 		/// <inheritdoc cref="IPipelineProcessor" />
 		public ProcessorsDescriptor Circle<T>(Func<CircleProcessorDescriptor<T>, ICircleProcessor> selector) where T : class  =>
 			Assign(selector, (a, v) => a.AddIfNotNull(v?.Invoke(new CircleProcessorDescriptor<T>())));
+
+		/// <inheritdoc cref="IUriPartsProcessor"/>
+		public ProcessorsDescriptor UriParts<T>(Func<UriPartsProcessorDescriptor<T>, IUriPartsProcessor> selector) where T : class =>
+			Assign(selector, (a, v) => a.AddIfNotNull(v?.Invoke(new UriPartsProcessorDescriptor<T>())));
 	}
 }

tests/Tests/Ingest/ProcessorAssertions.cs

@@ -702,5 +702,19 @@ public class Pipeline : ProcessorAssertion
 
 			public override string Key => "pipeline";
 		}
+
+		[SkipVersion("<7.11.0", "Uses URI parts which was introduced in 7.11.0")]
+		public class UriParts : ProcessorAssertion
+		{
+			public override Func<ProcessorsDescriptor, IPromise<IList<IProcessor>>> Fluent => d => d
+				.UriParts<Project>(ud => ud
+				.Field(p => p.Description)
+				.KeepOriginal()
+				.RemoveIfSuccessful());
+
+			public override IProcessor Initializer => new UriPartsProcessor { Field = "description", KeepOriginal = true, RemoveIfSuccessful = true };
+			public override object Json => new { field = "description", keep_original = true, remove_if_successful = true };
+			public override string Key => "uri_parts";
+		}
 	}
 }