From 1a27080e79e765b656e881871d5d79e939970b95 Mon Sep 17 00:00:00 2001
From: Laura Trotta <laura.trotta@elastic.co>
Date: Mon, 28 Oct 2024 09:42:27 +0100
Subject: [PATCH] ngram tokenizer backport

---
 output/openapi/elasticsearch-openapi.json     |  5 +--
 .../elasticsearch-serverless-openapi.json     |  5 +--
 output/schema/schema-serverless.json          | 37 ++++++++++---------
 output/schema/schema.json                     | 37 ++++++++++---------
 output/typescript/types.ts                    |  6 +--
 specification/_types/analysis/tokenizers.ts   |  9 +++--
 6 files changed, 49 insertions(+), 50 deletions(-)

diff --git a/output/openapi/elasticsearch-openapi.json b/output/openapi/elasticsearch-openapi.json
index 6bd96fd9d5..8074c6e58d 100644
--- a/output/openapi/elasticsearch-openapi.json
+++ b/output/openapi/elasticsearch-openapi.json
@@ -58561,10 +58561,7 @@
               }
             },
             "required": [
-              "type",
-              "max_gram",
-              "min_gram",
-              "token_chars"
+              "type"
             ]
           }
         ]
diff --git a/output/openapi/elasticsearch-serverless-openapi.json b/output/openapi/elasticsearch-serverless-openapi.json
index 08d407d333..e0b2adce5f 100644
--- a/output/openapi/elasticsearch-serverless-openapi.json
+++ b/output/openapi/elasticsearch-serverless-openapi.json
@@ -39323,10 +39323,7 @@
               }
             },
             "required": [
-              "type",
-              "max_gram",
-              "min_gram",
-              "token_chars"
+              "type"
             ]
           }
         ]
diff --git a/output/schema/schema-serverless.json b/output/schema/schema-serverless.json
index b259ce04f3..caa4532c2f 100644
--- a/output/schema/schema-serverless.json
+++ b/output/schema/schema-serverless.json
@@ -53207,7 +53207,7 @@
         "name": "NoriDecompoundMode",
         "namespace": "_types.analysis"
       },
-      "specLocation": "_types/analysis/tokenizers.ts#L74-L78"
+      "specLocation": "_types/analysis/tokenizers.ts#L77-L81"
     },
     {
       "kind": "interface",
@@ -79776,7 +79776,7 @@
         "name": "TokenChar",
         "namespace": "_types.analysis"
       },
-      "specLocation": "_types/analysis/tokenizers.ts#L46-L53"
+      "specLocation": "_types/analysis/tokenizers.ts#L49-L56"
     },
     {
       "codegenNames": [
@@ -83059,7 +83059,7 @@
         "name": "Tokenizer",
         "namespace": "_types.analysis"
       },
-      "specLocation": "_types/analysis/tokenizers.ts#L119-L121",
+      "specLocation": "_types/analysis/tokenizers.ts#L122-L124",
       "type": {
         "items": [
           {
@@ -83086,7 +83086,7 @@
         "name": "TokenizerDefinition",
         "namespace": "_types.analysis"
       },
-      "specLocation": "_types/analysis/tokenizers.ts#L123-L141",
+      "specLocation": "_types/analysis/tokenizers.ts#L126-L144",
       "type": {
         "items": [
           {
@@ -83243,7 +83243,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L55-L59"
+      "specLocation": "_types/analysis/tokenizers.ts#L58-L62"
     },
     {
       "kind": "interface",
@@ -83370,7 +83370,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L61-L64"
+      "specLocation": "_types/analysis/tokenizers.ts#L64-L67"
     },
     {
       "inherits": {
@@ -83394,7 +83394,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L66-L68"
+      "specLocation": "_types/analysis/tokenizers.ts#L69-L71"
     },
     {
       "inherits": {
@@ -83418,7 +83418,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L70-L72"
+      "specLocation": "_types/analysis/tokenizers.ts#L73-L75"
     },
     {
       "inherits": {
@@ -83454,7 +83454,7 @@
         },
         {
           "name": "max_gram",
-          "required": true,
+          "required": false,
           "type": {
             "kind": "instance_of",
             "type": {
@@ -83465,7 +83465,7 @@
         },
         {
           "name": "min_gram",
-          "required": true,
+          "required": false,
           "type": {
             "kind": "instance_of",
             "type": {
@@ -83476,7 +83476,8 @@
         },
         {
           "name": "token_chars",
-          "required": true,
+          "required": false,
+          "serverDefault": [],
           "type": {
             "kind": "array_of",
             "value": {
@@ -83489,7 +83490,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L38-L44"
+      "specLocation": "_types/analysis/tokenizers.ts#L38-L47"
     },
     {
       "inherits": {
@@ -83560,7 +83561,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L80-L86"
+      "specLocation": "_types/analysis/tokenizers.ts#L83-L89"
     },
     {
       "inherits": {
@@ -83666,7 +83667,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L88-L95"
+      "specLocation": "_types/analysis/tokenizers.ts#L91-L98"
     },
     {
       "inherits": {
@@ -83701,7 +83702,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L104-L107"
+      "specLocation": "_types/analysis/tokenizers.ts#L107-L110"
     },
     {
       "inherits": {
@@ -83736,7 +83737,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L109-L112"
+      "specLocation": "_types/analysis/tokenizers.ts#L112-L115"
     },
     {
       "inherits": {
@@ -83771,7 +83772,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L114-L117"
+      "specLocation": "_types/analysis/tokenizers.ts#L117-L120"
     },
     {
       "inherits": {
@@ -83932,7 +83933,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L97-L102"
+      "specLocation": "_types/analysis/tokenizers.ts#L100-L105"
     },
     {
       "inherits": {
diff --git a/output/schema/schema.json b/output/schema/schema.json
index bb59892038..daf4bd89bd 100644
--- a/output/schema/schema.json
+++ b/output/schema/schema.json
@@ -65376,7 +65376,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L55-L59"
+      "specLocation": "_types/analysis/tokenizers.ts#L58-L62"
     },
     {
       "inherits": {
@@ -67129,7 +67129,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L61-L64"
+      "specLocation": "_types/analysis/tokenizers.ts#L64-L67"
     },
     {
       "kind": "interface",
@@ -67704,7 +67704,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L66-L68"
+      "specLocation": "_types/analysis/tokenizers.ts#L69-L71"
     },
     {
       "inherits": {
@@ -67836,7 +67836,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L70-L72"
+      "specLocation": "_types/analysis/tokenizers.ts#L73-L75"
     },
     {
       "inherits": {
@@ -68045,7 +68045,7 @@
         },
         {
           "name": "max_gram",
-          "required": true,
+          "required": false,
           "type": {
             "kind": "instance_of",
             "type": {
@@ -68056,7 +68056,7 @@
         },
         {
           "name": "min_gram",
-          "required": true,
+          "required": false,
           "type": {
             "kind": "instance_of",
             "type": {
@@ -68067,7 +68067,8 @@
         },
         {
           "name": "token_chars",
-          "required": true,
+          "required": false,
+          "serverDefault": [],
           "type": {
             "kind": "array_of",
             "value": {
@@ -68080,7 +68081,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L38-L44"
+      "specLocation": "_types/analysis/tokenizers.ts#L38-L47"
     },
     {
       "kind": "interface",
@@ -68164,7 +68165,7 @@
         "name": "NoriDecompoundMode",
         "namespace": "_types.analysis"
       },
-      "specLocation": "_types/analysis/tokenizers.ts#L74-L78"
+      "specLocation": "_types/analysis/tokenizers.ts#L77-L81"
     },
     {
       "inherits": {
@@ -68273,7 +68274,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L80-L86"
+      "specLocation": "_types/analysis/tokenizers.ts#L83-L89"
     },
     {
       "docId": "analysis-normalizers",
@@ -68413,7 +68414,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L88-L95"
+      "specLocation": "_types/analysis/tokenizers.ts#L91-L98"
     },
     {
       "kind": "interface",
@@ -68726,7 +68727,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L97-L102"
+      "specLocation": "_types/analysis/tokenizers.ts#L100-L105"
     },
     {
       "kind": "enum",
@@ -69445,7 +69446,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L104-L107"
+      "specLocation": "_types/analysis/tokenizers.ts#L107-L110"
     },
     {
       "inherits": {
@@ -69956,7 +69957,7 @@
         "name": "TokenChar",
         "namespace": "_types.analysis"
       },
-      "specLocation": "_types/analysis/tokenizers.ts#L46-L53"
+      "specLocation": "_types/analysis/tokenizers.ts#L49-L56"
     },
     {
       "codegenNames": [
@@ -70367,7 +70368,7 @@
         "name": "Tokenizer",
         "namespace": "_types.analysis"
       },
-      "specLocation": "_types/analysis/tokenizers.ts#L119-L121",
+      "specLocation": "_types/analysis/tokenizers.ts#L122-L124",
       "type": {
         "items": [
           {
@@ -70415,7 +70416,7 @@
         "name": "TokenizerDefinition",
         "namespace": "_types.analysis"
       },
-      "specLocation": "_types/analysis/tokenizers.ts#L123-L141",
+      "specLocation": "_types/analysis/tokenizers.ts#L126-L144",
       "type": {
         "items": [
           {
@@ -70617,7 +70618,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L109-L112"
+      "specLocation": "_types/analysis/tokenizers.ts#L112-L115"
     },
     {
       "inherits": {
@@ -70740,7 +70741,7 @@
           }
         }
       ],
-      "specLocation": "_types/analysis/tokenizers.ts#L114-L117"
+      "specLocation": "_types/analysis/tokenizers.ts#L117-L120"
     },
     {
       "inherits": {
diff --git a/output/typescript/types.ts b/output/typescript/types.ts
index 73287f120a..76ddc4382c 100644
--- a/output/typescript/types.ts
+++ b/output/typescript/types.ts
@@ -4736,9 +4736,9 @@ export interface AnalysisNGramTokenFilter extends AnalysisTokenFilterBase {
 export interface AnalysisNGramTokenizer extends AnalysisTokenizerBase {
   type: 'ngram'
   custom_token_chars?: string
-  max_gram: integer
-  min_gram: integer
-  token_chars: AnalysisTokenChar[]
+  max_gram?: integer
+  min_gram?: integer
+  token_chars?: AnalysisTokenChar[]
 }
 
 export interface AnalysisNoriAnalyzer {
diff --git a/specification/_types/analysis/tokenizers.ts b/specification/_types/analysis/tokenizers.ts
index 322a0098a1..d61128b717 100644
--- a/specification/_types/analysis/tokenizers.ts
+++ b/specification/_types/analysis/tokenizers.ts
@@ -38,9 +38,12 @@ export class EdgeNGramTokenizer extends TokenizerBase {
 export class NGramTokenizer extends TokenizerBase {
   type: 'ngram'
   custom_token_chars?: string
-  max_gram: integer
-  min_gram: integer
-  token_chars: TokenChar[]
+  max_gram?: integer
+  min_gram?: integer
+  /**
+   * @server_default []
+   */
+  token_chars?: TokenChar[]
 }
 
 export enum TokenChar {