From 9c589eff4524037bdeb9d15216109cbf78c3293a Mon Sep 17 00:00:00 2001
From: Russ Cam <russ.cam@elastic.co>
Date: Mon, 1 Apr 2019 15:51:07 +1000
Subject: [PATCH] Add User Dictionary Rules to NoriTokenizer

Relates #3615, elastic/elasticsearch#3620
---
 src/Nest/Analysis/Tokenizers/NoriTokenizer.cs | 25 +++++++++++++++++--
 .../Analysis/Tokenizers/TokenizerTests.cs     | 23 +++++++++++++++++
 2 files changed, 46 insertions(+), 2 deletions(-)

diff --git a/src/Nest/Analysis/Tokenizers/NoriTokenizer.cs b/src/Nest/Analysis/Tokenizers/NoriTokenizer.cs
index 761f4e3d8c4..6ab009e2b9b 100644
--- a/src/Nest/Analysis/Tokenizers/NoriTokenizer.cs
+++ b/src/Nest/Analysis/Tokenizers/NoriTokenizer.cs
@@ -1,4 +1,5 @@
-using System.Runtime.Serialization;
+using System.Collections.Generic;
+using System.Runtime.Serialization;
 using Newtonsoft.Json;
 using Newtonsoft.Json.Converters;
 
@@ -32,10 +33,20 @@ public interface INoriTokenizer : ITokenizer
 
 		/// <summary>
 		/// The Nori tokenizer uses the mecab-ko-dic dictionary by default. A user_dictionary with custom nouns (NNG) may be appended to
-		/// the default dictionary. This property allows you to specify this file on disk
+		/// the default dictionary. This property allows you to specify a path to this file on disk
 		/// </summary>
 		[JsonProperty("user_dictionary")]
 		string UserDictionary { get; set; }
+
+		/// <summary>
+		/// The Nori tokenizer uses the mecab-ko-dic dictionary by default. A user_dictionary with custom nouns (NNG)
+		/// can be specified inline with this property
+		/// </summary>
+		/// <remarks>
+		/// Valid for Elasticsearch 6.6.0+
+		/// </remarks>
+		[JsonProperty("user_dictionary_rules")]
+		IEnumerable<string> UserDictionaryRules { get; set; }
 	}
 
 	/// <inheritdoc cref="INoriTokenizer" />
@@ -48,6 +59,9 @@ public class NoriTokenizer : TokenizerBase, INoriTokenizer
 
 		/// <inheritdoc cref="INoriTokenizer.UserDictionary" />
 		public string UserDictionary { get; set; }
+
+		/// <inheritdoc cref="INoriTokenizer.UserDictionaryRules" />
+		public IEnumerable<string> UserDictionaryRules { get; set; }
 	}
 
 	/// <inheritdoc cref="INoriTokenizer" />
@@ -58,11 +72,18 @@ public class NoriTokenizerDescriptor
 
 		NoriDecompoundMode? INoriTokenizer.DecompoundMode { get; set; }
 		string INoriTokenizer.UserDictionary { get; set; }
+		IEnumerable<string> INoriTokenizer.UserDictionaryRules { get; set; }
 
 		/// <inheritdoc cref="INoriTokenizer.DecompoundMode" />
 		public NoriTokenizerDescriptor DecompoundMode(NoriDecompoundMode? mode) => Assign(a => a.DecompoundMode = mode);
 
 		/// <inheritdoc cref="INoriTokenizer.UserDictionary" />
 		public NoriTokenizerDescriptor UserDictionary(string path) => Assign(a => a.UserDictionary = path);
+
+		/// <inheritdoc cref="INoriTokenizer.UserDictionaryRules" />
+		public NoriTokenizerDescriptor UserDictionaryRules(params string[] rules) => Assign(a => a.UserDictionaryRules = rules);
+
+		/// <inheritdoc cref="INoriTokenizer.UserDictionaryRules" />
+		public NoriTokenizerDescriptor UserDictionaryRules(IEnumerable<string> rules) => Assign(a => a.UserDictionaryRules = rules);
 	}
 }
diff --git a/src/Tests/Tests/Analysis/Tokenizers/TokenizerTests.cs b/src/Tests/Tests/Analysis/Tokenizers/TokenizerTests.cs
index bb9a8e9858c..c124b1ca055 100644
--- a/src/Tests/Tests/Analysis/Tokenizers/TokenizerTests.cs
+++ b/src/Tests/Tests/Analysis/Tokenizers/TokenizerTests.cs
@@ -222,6 +222,29 @@ public class NoriTests : TokenizerAssertionBase<NoriTests>
 			public override string Name => "nori";
 		}
 
+		[SkipVersion("<6.6.0", "inline user dictionary rules introduced in 6.6.0")]
+		public class NoriWithUserDictionaryTests : TokenizerAssertionBase<NoriWithUserDictionaryTests>
+		{
+			public override FuncTokenizer Fluent => (n, t) => t.Nori(n, e => e
+				.DecompoundMode(NoriDecompoundMode.Mixed)
+				.UserDictionaryRules("c++", "C샤프", "세종", "세종시 세종 시")
+			);
+
+			public override ITokenizer Initializer => new NoriTokenizer
+			{
+				DecompoundMode = NoriDecompoundMode.Mixed,
+				UserDictionaryRules = new [] { "c++", "C샤프", "세종", "세종시 세종 시" }
+			};
+
+			public override object Json => new
+			{
+				type = "nori_tokenizer",
+				decompound_mode = "mixed",
+				user_dictionary_rules = new [] { "c++", "C샤프", "세종", "세종시 세종 시" }
+			};
+			public override string Name => "nori_userdictionary";
+		}
+
 		[SkipVersion("<6.4.0", "char_group introduced in 6.4.0")]
 		public class CharGroupTests : TokenizerAssertionBase<CharGroupTests>
 		{