[LLM tasks] Add product documentation retrieval task (elastic#194379)

## Summary Close elastic#193473 Close elastic#193474 This PR utilize the documentation packages that are build via the tool introduced by elastic#193847, allowing to install them in Kibana and expose documentation retrieval as an LLM task that AI assistants (or other consumers) can call. Users can now decide to install the Elastic documentation from the assistant's config screen, which will expose a new tool for the assistant, `retrieve_documentation` (only implemented for the o11y assistant in the current PR, shall be done for security as a follow up). For more information, please refer to the self-review. ## General architecture <img width="1118" alt="Screenshot 2024-10-17 at 09 22 32" src="https://github.com/user-attachments/assets/3df8c30a-9ccc-49ab-92ce-c204b96d6fc4"> ## What this PR does Adds two plugin: - `productDocBase`: contains all the logic related to product documentation installation, status, and search. This is meant to be a "low level" components only responsible for this specific part. - `llmTasks`: an higher level plugin that will contain various LLM tasks to be used by assistants and genAI consumers. The intent is not to have a single place to put all llm tasks, but more to have a default place where we can introduce new tasks from. (fwiw, the `nlToEsql` task will probably be moved to that plugin). - Add a `retrieve_documentation` tool registration for the o11y assistant - Add a component on the o11y assistant configuration page to install the product doc (wiring the feature to the o11y assistant was done for testing purposes mostly, any addition / changes / enhancement should be done by the owning team - either in this PR or as a follow-up) ## What is NOT included in this PR: - Wire product base feature to the security assistant (should be done by the owning team as a follow-up) - installation - utilization as tool - FTR tests: this is somewhat blocked by the same things we need to figure out for elastic/kibana-team#1271 ## Screenshots ### Installation from o11y assistant configuration page <img width="1476" alt="Screenshot 2024-10-17 at 09 41 24" src="https://github.com/user-attachments/assets/31daa585-9fb2-400a-a2d1-5917a262367a"> ### Example of output #### Without product documentation installed <img width="739" alt="Screenshot 2024-10-10 at 09 59 41" src="https://github.com/user-attachments/assets/993fb216-6c9a-433f-bf44-f6e383d20d9d"> #### With product documentation installed <img width="718" alt="Screenshot 2024-10-10 at 09 55 38" src="https://github.com/user-attachments/assets/805ea4ca-8bc9-4355-a434-0ba81f8228a9"> --------- Co-authored-by: kibanamachine <[email protected]> Co-authored-by: Alex Szabo <[email protected]> Co-authored-by: Matthias Wilhelm <[email protected]> Co-authored-by: Elastic Machine <[email protected]> (cherry picked from commit 455c781) # Conflicts: # .github/CODEOWNERS
pgayvallet · Nov 19, 2024 · dc865f8 · dc865f8
1 parent 43fa8a5
commit dc865f8
Show file tree

Hide file tree

Showing 149 changed files with 5,659 additions and 64 deletions.
diff --git a/docs/developer/plugin-list.asciidoc b/docs/developer/plugin-list.asciidoc
@@ -690,6 +690,10 @@ the infrastructure monitoring use-case within Kibana.
 using the CURL scripts in the scripts folder.
 
 
+|{kib-repo}blob/{branch}/x-pack/plugins/ai_infra/llm_tasks/README.md[llmTasks]
+|This plugin contains various LLM tasks.
+
+
 |{kib-repo}blob/{branch}/x-pack/plugins/observability_solution/logs_data_access/README.md[logsDataAccess]
 |Exposes services to access logs data.
 
@@ -767,6 +771,10 @@ Elastic.
 |This plugin helps users learn how to use the Painless scripting language.
 
 
+|{kib-repo}blob/{branch}/x-pack/plugins/ai_infra/product_doc_base/README.md[productDocBase]
+|This plugin contains the product documentation base service.
+
+
 |{kib-repo}blob/{branch}/x-pack/plugins/observability_solution/profiling/README.md[profiling]
 |Universal Profiling provides fleet-wide, whole-system, continuous profiling with zero instrumentation. Get a comprehensive understanding of what lines of code are consuming compute resources throughout your entire fleet by visualizing your data in Kibana using the flamegraph, stacktraces, and top functions views.
 

diff --git a/docs/user/security/audit-logging.asciidoc b/docs/user/security/audit-logging.asciidoc
@@ -148,6 +148,9 @@ Refer to the corresponding {es} logs for potential write errors.
 | `success` | Creating trained model.
 | `failure` | Failed to create trained model.
 
+.1+| `product_documentation_create`
+| `unknown` | User requested to install the product documentation for use in AI Assistants.
+
 3+a|
 ====== Type: change
 
@@ -334,6 +337,9 @@ Refer to the corresponding {es} logs for potential write errors.
 | `success` | Updating trained model deployment.
 | `failure` | Failed to update trained model deployment.
 
+.1+| `product_documentation_update`
+| `unknown` | User requested to update the product documentation for use in AI Assistants.
+
 3+a|
 ====== Type: deletion
 
@@ -425,6 +431,9 @@ Refer to the corresponding {es} logs for potential write errors.
 | `success` | Deleting trained model.
 | `failure` | Failed to delete trained model.
 
+.1+| `product_documentation_delete`
+| `unknown` | User requested to delete the product documentation for use in AI Assistants.
+
 3+a|
 ====== Type: access
 

diff --git a/package.json b/package.json
@@ -616,6 +616,7 @@
     "@kbn/licensing-plugin": "link:x-pack/plugins/licensing",
     "@kbn/links-plugin": "link:src/plugins/links",
     "@kbn/lists-plugin": "link:x-pack/plugins/lists",
+    "@kbn/llm-tasks-plugin": "link:x-pack/plugins/ai_infra/llm_tasks",
     "@kbn/locator-examples-plugin": "link:examples/locator_examples",
     "@kbn/locator-explorer-plugin": "link:examples/locator_explorer",
     "@kbn/logging": "link:packages/kbn-logging",
@@ -718,6 +719,8 @@
     "@kbn/presentation-panel-plugin": "link:src/plugins/presentation_panel",
     "@kbn/presentation-publishing": "link:packages/presentation/presentation_publishing",
     "@kbn/presentation-util-plugin": "link:src/plugins/presentation_util",
+    "@kbn/product-doc-base-plugin": "link:x-pack/plugins/ai_infra/product_doc_base",
+    "@kbn/product-doc-common": "link:x-pack/packages/ai-infra/product-doc-common",
     "@kbn/profiling-data-access-plugin": "link:x-pack/plugins/observability_solution/profiling_data_access",
     "@kbn/profiling-plugin": "link:x-pack/plugins/observability_solution/profiling",
     "@kbn/profiling-utils": "link:packages/kbn-profiling-utils",

diff --git a/packages/kbn-check-mappings-update-cli/current_fields.json b/packages/kbn-check-mappings-update-cli/current_fields.json
@@ -855,6 +855,13 @@
   "policy-settings-protection-updates-note": [
     "note"
   ],
+  "product-doc-install-status": [
+    "index_name",
+    "installation_status",
+    "last_installation_date",
+    "product_name",
+    "product_version"
+  ],
   "query": [
     "description",
     "title",

diff --git a/packages/kbn-check-mappings-update-cli/current_mappings.json b/packages/kbn-check-mappings-update-cli/current_mappings.json
@@ -2841,6 +2841,26 @@
       }
     }
   },
+  "product-doc-install-status": {
+    "dynamic": false,
+    "properties": {
+      "index_name": {
+        "type": "keyword"
+      },
+      "installation_status": {
+        "type": "keyword"
+      },
+      "last_installation_date": {
+        "type": "date"
+      },
+      "product_name": {
+        "type": "keyword"
+      },
+      "product_version": {
+        "type": "keyword"
+      }
+    }
+  },
   "query": {
     "dynamic": false,
     "properties": {

diff --git a/packages/kbn-optimizer/limits.yml b/packages/kbn-optimizer/limits.yml
@@ -124,6 +124,7 @@ pageLoadAssetSize:
   painlessLab: 179748
   presentationPanel: 55463
   presentationUtil: 58834
+  productDocBase: 22500
   profiling: 36694
   remoteClusters: 51327
   reporting: 58600

diff --git a/src/core/server/integration_tests/ci_checks/saved_objects/check_registered_types.test.ts b/src/core/server/integration_tests/ci_checks/saved_objects/check_registered_types.test.ts
@@ -145,6 +145,7 @@ describe('checking migration metadata changes on all registered SO types', () =>
         "osquery-pack-asset": "cd140bc2e4b092e93692b587bf6e38051ef94c75",
         "osquery-saved-query": "6095e288750aa3164dfe186c74bc5195c2bf2bd4",
         "policy-settings-protection-updates-note": "33924bb246f9e5bcb876109cc83e3c7a28308352",
+        "product-doc-install-status": "ca6e96840228e4cc2f11bae24a0797f4f7238c8c",
         "query": "501bece68f26fe561286a488eabb1a8ab12f1137",
         "risk-engine-configuration": "aea0c371a462e6d07c3ceb3aff11891b47feb09d",
         "rules-settings": "ba57ef1881b3dcbf48fbfb28902d8f74442190b2",

diff --git a/src/core/server/integration_tests/saved_objects/migrations/group3/type_registrations.test.ts b/src/core/server/integration_tests/saved_objects/migrations/group3/type_registrations.test.ts
@@ -115,6 +115,7 @@ const previouslyRegisteredTypes = [
   'osquery-usage-metric',
   'osquery-manager-usage-metric',
   'policy-settings-protection-updates-note',
+  'product-doc-install-status',
   'query',
   'rules-settings',
   'sample-data-telemetry',

diff --git a/tsconfig.base.json b/tsconfig.base.json
@@ -1146,6 +1146,8 @@
       "@kbn/lint-ts-projects-cli/*": ["packages/kbn-lint-ts-projects-cli/*"],
       "@kbn/lists-plugin": ["x-pack/plugins/lists"],
       "@kbn/lists-plugin/*": ["x-pack/plugins/lists/*"],
+      "@kbn/llm-tasks-plugin": ["x-pack/plugins/ai_infra/llm_tasks"],
+      "@kbn/llm-tasks-plugin/*": ["x-pack/plugins/ai_infra/llm_tasks/*"],
       "@kbn/locator-examples-plugin": ["examples/locator_examples"],
       "@kbn/locator-examples-plugin/*": ["examples/locator_examples/*"],
       "@kbn/locator-explorer-plugin": ["examples/locator_explorer"],
@@ -1384,6 +1386,10 @@
       "@kbn/presentation-util-plugin/*": ["src/plugins/presentation_util/*"],
       "@kbn/product-doc-artifact-builder": ["x-pack/packages/ai-infra/product-doc-artifact-builder"],
       "@kbn/product-doc-artifact-builder/*": ["x-pack/packages/ai-infra/product-doc-artifact-builder/*"],
+      "@kbn/product-doc-base-plugin": ["x-pack/plugins/ai_infra/product_doc_base"],
+      "@kbn/product-doc-base-plugin/*": ["x-pack/plugins/ai_infra/product_doc_base/*"],
+      "@kbn/product-doc-common": ["x-pack/packages/ai-infra/product-doc-common"],
+      "@kbn/product-doc-common/*": ["x-pack/packages/ai-infra/product-doc-common/*"],
       "@kbn/profiling-data-access-plugin": ["x-pack/plugins/observability_solution/profiling_data_access"],
       "@kbn/profiling-data-access-plugin/*": ["x-pack/plugins/observability_solution/profiling_data_access/*"],
       "@kbn/profiling-plugin": ["x-pack/plugins/observability_solution/profiling"],

diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/README.md b/x-pack/packages/ai-infra/product-doc-artifact-builder/README.md
@@ -1,3 +1,49 @@
 # @kbn/product-doc-artifact-builder
 
-Script to build the knowledge base artifacts
+Script to build the knowledge base artifacts.
+
+## How to run
+
+```
+node scripts/build_product_doc_artifacts.js --stack-version {version} --product-name {product}
+```
+
+### parameters
+
+#### `stack-version`: 
+
+the stack version to generate the artifacts for.
+
+#### `product-name`: 
+
+(multi-value) the list of products to generate artifacts for.
+
+possible values: 
+- "kibana"
+- "elasticsearch"
+- "observability"
+- "security"
+
+#### `target-folder`:
+
+The folder to generate the artifacts in.
+
+Defaults to `{REPO_ROOT}/build-kb-artifacts`.
+
+#### `build-folder`:
+
+The folder to use for temporary files.
+
+Defaults to `{REPO_ROOT}/build/temp-kb-artifacts`
+
+#### Cluster infos
+
+- params for the source cluster:
+`sourceClusterUrl` / env.KIBANA_SOURCE_CLUSTER_URL
+`sourceClusterUsername` / env.KIBANA_SOURCE_CLUSTER_USERNAME
+`sourceClusterPassword` / env.KIBANA_SOURCE_CLUSTER_PASSWORD
+
+- params for the embedding cluster:
+`embeddingClusterUrl` / env.KIBANA_EMBEDDING_CLUSTER_URL
+`embeddingClusterUsername` / env.KIBANA_EMBEDDING_CLUSTER_USERNAME
+`embeddingClusterPassword` / env.KIBANA_EMBEDDING_CLUSTER_PASSWORD
diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/src/artifact/manifest.ts b/x-pack/packages/ai-infra/product-doc-artifact-builder/src/artifact/manifest.ts
@@ -5,17 +5,13 @@
  * 2.0.
  */
 
-export interface ArtifactManifest {
-  formatVersion: string;
-  productName: string;
-  productVersion: string;
-}
+import type { ArtifactManifest, ProductName } from '@kbn/product-doc-common';
 
 export const getArtifactManifest = ({
   productName,
   stackVersion,
 }: {
-  productName: string;
+  productName: ProductName;
   stackVersion: string;
 }): ArtifactManifest => {
   return {

diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/src/artifact/mappings.ts b/x-pack/packages/ai-infra/product-doc-artifact-builder/src/artifact/mappings.ts
@@ -21,10 +21,7 @@ export const getArtifactMappings = (inferenceEndpoint: string): MappingTypeMappi
       slug: { type: 'keyword' },
       url: { type: 'keyword' },
       version: { type: 'version' },
-      ai_subtitle: {
-        type: 'semantic_text',
-        inference_id: inferenceEndpoint,
-      },
+      ai_subtitle: { type: 'text' },
       ai_summary: {
         type: 'semantic_text',
         inference_id: inferenceEndpoint,

diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/src/artifact/product_name.ts b/x-pack/packages/ai-infra/product-doc-artifact-builder/src/artifact/product_name.ts
@@ -5,7 +5,34 @@
  * 2.0.
  */
 
-/**
- * The allowed product names, as found in the source's cluster
- */
-export const sourceProductNames = ['Kibana', 'Elasticsearch', 'Security', 'Observability'];
+import type { ProductName } from '@kbn/product-doc-common';
+
+const productNameToSourceNamesMap: Record<ProductName, string[]> = {
+  kibana: ['Kibana'],
+  elasticsearch: ['Elasticsearch'],
+  security: ['Security'],
+  observability: ['Observability'],
+};
+
+const sourceNameToProductName = Object.entries(productNameToSourceNamesMap).reduce<
+  Record<string, ProductName>
+>((map, [productName, sourceNames]) => {
+  sourceNames.forEach((sourceName) => {
+    map[sourceName] = productName as ProductName;
+  });
+  return map;
+}, {});
+
+export const getSourceNamesFromProductName = (productName: ProductName): string[] => {
+  if (!productNameToSourceNamesMap[productName]) {
+    throw new Error(`Unknown product name: ${productName}`);
+  }
+  return productNameToSourceNamesMap[productName];
+};
+
+export const getProductNameFromSource = (source: string): ProductName => {
+  if (!sourceNameToProductName[source]) {
+    throw new Error(`Unknown source name: ${source}`);
+  }
+  return sourceNameToProductName[source];
+};
diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/src/build_artifacts.ts b/x-pack/packages/ai-infra/product-doc-artifact-builder/src/build_artifacts.ts
@@ -8,6 +8,7 @@
 import Path from 'path';
 import { Client } from '@elastic/elasticsearch';
 import { ToolingLog } from '@kbn/tooling-log';
+import type { ProductName } from '@kbn/product-doc-common';
 import {
   // checkConnectivity,
   createTargetIndex,
@@ -18,6 +19,7 @@ import {
   createArtifact,
   cleanupFolders,
   deleteIndex,
+  processDocuments,
 } from './tasks';
 import type { TaskConfig } from './types';
 
@@ -93,7 +95,7 @@ const buildArtifact = async ({
   sourceClient,
   log,
 }: {
-  productName: string;
+  productName: ProductName;
   stackVersion: string;
   buildFolder: string;
   targetFolder: string;
@@ -105,14 +107,16 @@ const buildArtifact = async ({
 
   const targetIndex = getTargetIndexName({ productName, stackVersion });
 
-  const documents = await extractDocumentation({
+  let documents = await extractDocumentation({
     client: sourceClient,
     index: 'search-docs-1',
     log,
     productName,
     stackVersion,
   });
 
+  documents = await processDocuments({ documents, log });
+
   await createTargetIndex({
     client: embeddingClient,
     indexName: targetIndex,

diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/src/command.ts b/x-pack/packages/ai-infra/product-doc-artifact-builder/src/command.ts
@@ -6,19 +6,19 @@
  */
 
 import Path from 'path';
-import { REPO_ROOT } from '@kbn/repo-info';
 import yargs from 'yargs';
+import { REPO_ROOT } from '@kbn/repo-info';
+import { DocumentationProduct } from '@kbn/product-doc-common';
 import type { TaskConfig } from './types';
 import { buildArtifacts } from './build_artifacts';
-import { sourceProductNames } from './artifact/product_name';
 
 function options(y: yargs.Argv) {
   return y
     .option('productName', {
       describe: 'name of products to generate documentation for',
       array: true,
-      choices: sourceProductNames,
-      default: ['Kibana'],
+      choices: Object.values(DocumentationProduct),
+      default: [DocumentationProduct.kibana],
     })
     .option('stackVersion', {
       describe: 'The stack version to generate documentation for',

diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/src/tasks/create_artifact.ts b/x-pack/packages/ai-infra/product-doc-artifact-builder/src/tasks/create_artifact.ts
@@ -8,9 +8,9 @@
 import Path from 'path';
 import AdmZip from 'adm-zip';
 import type { ToolingLog } from '@kbn/tooling-log';
+import { getArtifactName, type ProductName } from '@kbn/product-doc-common';
 import { getArtifactMappings } from '../artifact/mappings';
 import { getArtifactManifest } from '../artifact/manifest';
-import { getArtifactName } from '../artifact/artifact_name';
 
 export const createArtifact = async ({
   productName,
@@ -21,7 +21,7 @@ export const createArtifact = async ({
 }: {
   buildFolder: string;
   targetFolder: string;
-  productName: string;
+  productName: ProductName;
   stackVersion: string;
   log: ToolingLog;
 }) => {

diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/src/tasks/create_chunk_files.ts b/x-pack/packages/ai-infra/product-doc-artifact-builder/src/tasks/create_chunk_files.ts
@@ -10,7 +10,7 @@ import Fs from 'fs/promises';
 import type { Client } from '@elastic/elasticsearch';
 import type { ToolingLog } from '@kbn/tooling-log';
 
-const fileSizeLimit = 250_000;
+const fileSizeLimit = 500_000;
 
 export const createChunkFiles = async ({
   index,

diff --git a/x-pack/packages/ai-infra/product-doc-artifact-builder/src/tasks/create_index.ts b/x-pack/packages/ai-infra/product-doc-artifact-builder/src/tasks/create_index.ts
@@ -21,10 +21,7 @@ const mappings: MappingTypeMapping = {
     slug: { type: 'keyword' },
     url: { type: 'keyword' },
     version: { type: 'version' },
-    ai_subtitle: {
-      type: 'semantic_text',
-      inference_id: 'kibana-elser2',
-    },
+    ai_subtitle: { type: 'text' },
     ai_summary: {
       type: 'semantic_text',
       inference_id: 'kibana-elser2',