Use real index names in the analyze-text script (#64035)

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Copilot-Session: bf5940de-658e-4a1a-bffa-3018fb78b138

Author: heiskr

src/search/scripts/analyze-text.ts

@@ -9,6 +9,7 @@ import chalk from 'chalk'
 import dotenv from 'dotenv'
 
 import { languageKeys } from '@/languages/lib/languages-server'
+import { getElasticSearchIndex } from '@/search/lib/elasticsearch-indexes'
 import { allVersions } from '@/versions/lib/all-versions'
 
 import type { estypes } from '@elastic/elasticsearch'
@@ -110,10 +111,9 @@ async function main(opts: Options, textArgs: string[]): Promise<void> {
   }
 
   const { indexPrefix } = opts
-  const prefix = indexPrefix ? `${indexPrefix}_` : ''
 
   for (const languageKey of languagesToAnalyze) {
-    const indexName = `${prefix}github-docs-${versionKey}-${languageKey}`
+    const indexName = getIndexName(versionKey, languageKey, indexPrefix)
     console.log(chalk.yellow(`Analyzing in ${chalk.bold(indexName)}`))
     await analyzeVersion(client, texts, indexName)
   }
@@ -138,6 +138,11 @@ export function getLanguagesToAnalyze({
   return ['en']
 }
 
+// Matches the names the general search indexer creates.
+export function getIndexName(version: string, language: string, indexPrefix = ''): string {
+  return getElasticSearchIndex('generalSearch', version, language, indexPrefix).indexName
+}
+
 function safeUrlDisplay(url: string): string {
   const parsed = new URL(url)
   if (parsed.password) {

src/search/tests/analyze-text.ts

@@ -1,7 +1,8 @@
-import { describe, expect, test } from 'vitest'
+import { afterEach, describe, expect, test, vi } from 'vitest'
 
 import { languageKeys } from '@/languages/lib/languages-server'
-import { getLanguagesToAnalyze } from '@/search/scripts/analyze-text'
+import { getIndexName, getLanguagesToAnalyze } from '@/search/scripts/analyze-text'
+import { allVersions } from '@/versions/lib/all-versions'
 
 describe('getLanguagesToAnalyze', () => {
   test('defaults to English', () => {
@@ -24,3 +25,29 @@ describe('getLanguagesToAnalyze', () => {
     )
   })
 })
+
+describe('getIndexName', () => {
+  afterEach(() => {
+    vi.unstubAllEnvs()
+  })
+
+  test('maps dotcom to the fpt general search index', () => {
+    expect(getIndexName('dotcom', 'en')).toBe('tests_github-docs_general-search_fpt_en')
+  })
+
+  test('uses the ghec general search index', () => {
+    expect(getIndexName('ghec', 'en')).toBe('tests_github-docs_general-search_ghec_en')
+  })
+
+  test('uses the release in GHES general search index names', () => {
+    const ghes = Object.values(allVersions).find((info) => info.hasNumberedReleases)!
+    const versionKey = `${ghes.miscBaseName}${ghes.currentRelease}`
+    expect(getIndexName(versionKey, 'en')).toBe(`tests_github-docs_general-search_${versionKey}_en`)
+  })
+
+  test('adds one underscore after the index prefix', () => {
+    vi.stubEnv('NODE_ENV', 'development')
+    expect(getIndexName('dotcom', 'en', 'tests')).toBe('tests_github-docs_general-search_fpt_en')
+    expect(getIndexName('dotcom', 'en', 'tests_')).toBe('tests_github-docs_general-search_fpt_en')
+  })
+})