search: add longer timeouts to ingest examples for semantic_text

Author: TattdCodeMonkey

x-pack/solutions/search/plugins/search_indices/public/code_examples/ingest_data.ts

@@ -8,8 +8,8 @@
 import { i18n } from '@kbn/i18n';
 import { IngestDataCodeExamples } from '../types';
 
-import { JSIngestDataExample } from './javascript';
-import { PythonIngestDataExample } from './python';
+import { JSIngestDataExample, JSSemanticIngestDataExample } from './javascript';
+import { PythonIngestDataExample, PythonSemanticIngestDataExample } from './python';
 import { ConsoleIngestDataExample } from './sense';
 import { CurlIngestDataExample } from './curl';
 import { INSTALL_INSTRUCTIONS_TITLE, INSTALL_INSTRUCTIONS_DESCRIPTION } from './constants';
@@ -78,6 +78,6 @@ export const SemanticIngestDataCodeExamples: IngestDataCodeExamples = {
   },
   sense: ConsoleIngestDataExample,
   curl: CurlIngestDataExample,
-  python: PythonIngestDataExample,
-  javascript: JSIngestDataExample,
+  python: PythonSemanticIngestDataExample,
+  javascript: JSSemanticIngestDataExample,
 };

x-pack/solutions/search/plugins/search_indices/public/code_examples/javascript.ts

@@ -28,20 +28,20 @@ export const JavascriptCreateIndexExamples: CreateIndexLanguageExamples = {
       elasticsearchURL,
       apiKey,
       indexName,
-    }) => `import { Client } from "@elastic/elasticsearch"
+    }) => `const { Client } = require('@elastic/elasticsearch');
 
 const client = new Client({
   node: '${elasticsearchURL}',
   auth: {
-    apiKey: "${apiKey ?? API_KEY_PLACEHOLDER}"
+    apiKey: '${apiKey ?? API_KEY_PLACEHOLDER}'
   }
 });
 
 client.indices.create({
-  index: "${indexName ?? INDEX_PLACEHOLDER}",
+  index: '${indexName ?? INDEX_PLACEHOLDER}',
   mappings: {
     properties: {
-      text: { type: "text"}
+      text: { type: 'text'}
     },
   },
 });`,
@@ -52,21 +52,21 @@ client.indices.create({
       elasticsearchURL,
       apiKey,
       indexName,
-    }) => `import { Client } from "@elastic/elasticsearch"
+    }) => `const { Client } = require('@elastic/elasticsearch');
 
 const client = new Client({
   node: '${elasticsearchURL}',
   auth: {
-    apiKey: "${apiKey ?? API_KEY_PLACEHOLDER}"
+    apiKey: '${apiKey ?? API_KEY_PLACEHOLDER}'
   }
 });
 
 client.indices.create({
-  index: "${indexName ?? INDEX_PLACEHOLDER}",
+  index: '${indexName ?? INDEX_PLACEHOLDER}',
   mappings: {
     properties: {
-      vector: { type: "dense_vector", dims: 3 },
-      text: { type: "text"}
+      vector: { type: 'dense_vector', dims: 3 },
+      text: { type: 'text'}
     },
   },
 });`,
@@ -77,20 +77,20 @@ client.indices.create({
       elasticsearchURL,
       apiKey,
       indexName,
-    }) => `import { Client } from "@elastic/elasticsearch"
+    }) => `const { Client } = require('@elastic/elasticsearch');
 
 const client = new Client({
   node: '${elasticsearchURL}',
   auth: {
-    apiKey: "${apiKey ?? API_KEY_PLACEHOLDER}"
+    apiKey: '${apiKey ?? API_KEY_PLACEHOLDER}'
   }
 });
 
 client.indices.create({
-  index: "${indexName ?? INDEX_PLACEHOLDER}",
+  index: '${indexName ?? INDEX_PLACEHOLDER}',
   mappings: {
     properties: {
-      text: { type: "semantic_text"}
+      text: { type: 'semantic_text'}
     },
   },
 });`,
@@ -104,16 +104,16 @@ export const JSIngestDataExample: IngestDataCodeDefinition = {
     elasticsearchURL,
     sampleDocuments,
     indexName,
-  }) => `import { Client } from "@elastic/elasticsearch";
+  }) => `const { Client } = require('@elastic/elasticsearch');
 
 const client = new Client({
   node: '${elasticsearchURL}',
   auth: {
-    apiKey: "${apiKey ?? API_KEY_PLACEHOLDER}"
+    apiKey: '${apiKey ?? API_KEY_PLACEHOLDER}'
   },
 });
 
-const index = "${indexName}";
+const index = '${indexName}';
 const docs = ${JSON.stringify(sampleDocuments, null, 2)};
 
 const bulkIngestResponse = await client.helpers.bulk({
@@ -131,16 +131,16 @@ console.log(bulkIngestResponse);`,
     elasticsearchURL,
     indexName,
     mappingProperties,
-  }) => `import { Client } from "@elastic/elasticsearch";
+  }) => `const { Client } = require('@elastic/elasticsearch');
 
 const client = new Client({
-node: '${elasticsearchURL}',
-auth: {
-  apiKey: "${apiKey ?? API_KEY_PLACEHOLDER}"
-}
+  node: '${elasticsearchURL}',
+  auth: {
+    apiKey: '${apiKey ?? API_KEY_PLACEHOLDER}'
+  }
 });
 
-const index = "${indexName}";
+const index = '${indexName}';
 const mapping = ${JSON.stringify(mappingProperties, null, 2)};
 
 const updateMappingResponse = await client.indices.putMapping({
@@ -149,3 +149,41 @@ const updateMappingResponse = await client.indices.putMapping({
 });
 console.log(updateMappingResponse);`,
 };
+
+export const JSSemanticIngestDataExample: IngestDataCodeDefinition = {
+  ...JSIngestDataExample,
+  ingestCommand: ({
+    apiKey,
+    elasticsearchURL,
+    sampleDocuments,
+    indexName,
+  }) => `const { Client } = require('@elastic/elasticsearch');
+
+const client = new Client({
+  node: '${elasticsearchURL}',
+  auth: {
+    apiKey: '${apiKey ?? API_KEY_PLACEHOLDER}'
+  },
+});
+
+// Set the ingestion timeout to 5 minutes, to allow for semantic ingestion
+// to complete. The initial ingest with semantic_text fields may take longer
+// while the model loads, or ML nodes are allocated.
+const timeout = '5m';
+
+const index = '${indexName}';
+const docs = ${JSON.stringify(sampleDocuments, null, 2)};
+
+const bulkIngestResponse = await client.helpers.bulk({
+  index,
+  datasource: docs,
+  timeout,
+  onDocument() {
+    return {
+      index: {},
+    };
+  }
+});
+
+console.log(bulkIngestResponse);`,
+};

x-pack/solutions/search/plugins/search_indices/public/code_examples/python.ts

@@ -133,8 +133,39 @@ mapping_response = client.indices.put_mapping(index=index_name, body=mappings)
 print(mapping_response)
 `;
 
+const semanticIngestCommand: IngestCodeSnippetFunction = ({
+  elasticsearchURL,
+  apiKey,
+  indexName,
+  sampleDocuments,
+}) => `from elasticsearch import Elasticsearch, helpers
+
+# Set the ingestion timeout to 5 minutes, to allow time for semantic ingestion
+# to complete. The initial ingest with semantic_text fields may take longer
+# while the model loads, or ML nodes are allocated.
+ingestion_timeout=300
+
+client = Elasticsearch(
+    "${elasticsearchURL}",
+    api_key="${apiKey ?? API_KEY_PLACEHOLDER}",
+    request_timeout=ingestion_timeout,
+    retry_on_timeout=True,
+)
+
+index_name = "${indexName}"
+
+docs = ${JSON.stringify(sampleDocuments, null, 4)}
+
+bulk_response = helpers.bulk(client, docs, index=index_name)
+print(bulk_response)`;
+
 export const PythonIngestDataExample: IngestDataCodeDefinition = {
   installCommand: PYTHON_INSTALL_CMD,
   ingestCommand: ingestionCommand,
   updateMappingsCommand,
 };
+
+export const PythonSemanticIngestDataExample: IngestDataCodeDefinition = {
+  ...PythonIngestDataExample,
+  ingestCommand: semanticIngestCommand,
+};