[ML] Adding category definition and category examples api tests (#124710

) (#125077)

* [ML] Adding category definition api tests

* adding test

* adding category examples tests

* updating test checks

* changes based on review

* removing test

* correcting mistake where wrong test was removed

* correctling text text

* fixing test

Co-authored-by: Kibana Machine <42973632+kibanamachine@users.noreply.github.com>
(cherry picked from commit 9790a7e)

Co-authored-by: James Gowdy <jgowdy@elastic.co>

x-pack/test/api_integration/apis/ml/results/get_category_definition.ts

@@ -0,0 +1,174 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0; you may not use this file except in compliance with the Elastic License
+ * 2.0.
+ */
+
+import expect from '@kbn/expect';
+import { FtrProviderContext } from '../../../ftr_provider_context';
+import { USER } from '../../../../functional/services/ml/security_common';
+import { COMMON_REQUEST_HEADERS } from '../../../../functional/services/ml/common_api';
+import { Datafeed } from '../../../../../plugins/ml/common/types/anomaly_detection_jobs';
+
+export default ({ getService }: FtrProviderContext) => {
+  const esArchiver = getService('esArchiver');
+  const supertest = getService('supertestWithoutAuth');
+  const ml = getService('ml');
+  const spacesService = getService('spaces');
+
+  const jobIdSpace1 = `sample_logs_${Date.now()}`;
+  const idSpace1 = 'space1';
+  const idSpace2 = 'space2';
+
+  const PARTITION_FIELD_NAME = 'event.dataset';
+  const testJobConfig = {
+    job_id: jobIdSpace1,
+    groups: ['sample_logs', 'bootstrap', 'categorization'],
+    description: "count by mlcategory (message) on 'sample logs' dataset with 15m bucket span",
+    analysis_config: {
+      bucket_span: '15m',
+      categorization_field_name: 'message',
+      per_partition_categorization: { enabled: true, stop_on_warn: true },
+      detectors: [
+        {
+          function: 'count',
+          by_field_name: 'mlcategory',
+          partition_field_name: PARTITION_FIELD_NAME,
+        },
+      ],
+      influencers: ['mlcategory'],
+    },
+    analysis_limits: { model_memory_limit: '26MB' },
+    data_description: { time_field: '@timestamp', time_format: 'epoch_ms' },
+    model_plot_config: { enabled: false, annotations_enabled: true },
+    model_snapshot_retention_days: 10,
+    daily_model_snapshot_retention_after_days: 1,
+    allow_lazy_open: false,
+  };
+  // @ts-expect-error not full interface
+  const testDatafeedConfig: Datafeed = {
+    datafeed_id: `datafeed-${jobIdSpace1}`,
+    indices: ['ft_module_sample_logs'],
+    job_id: jobIdSpace1,
+    query: { bool: { must: [{ match_all: {} }] } },
+  };
+
+  const expectedCategoryDefinition = {
+    categoryId: '1',
+    examplesLength: 4,
+  };
+
+  async function getCategoryDefinition(
+    jobId: string,
+    categoryId: string,
+    user: USER,
+    expectedStatusCode: number,
+    space?: string
+  ) {
+    const { body } = await supertest
+      .post(`${space ? `/s/${space}` : ''}/api/ml/results/category_definition`)
+      .auth(user, ml.securityCommon.getPasswordForUser(user))
+      .set(COMMON_REQUEST_HEADERS)
+      .send({ jobId, categoryId })
+      .expect(expectedStatusCode);
+
+    return body;
+  }
+
+  describe('get category_definition', function () {
+    before(async () => {
+      await esArchiver.loadIfNeeded('x-pack/test/functional/es_archives/ml/module_sample_logs');
+      await ml.testResources.setKibanaTimeZoneToUTC();
+      await spacesService.create({ id: idSpace1, name: 'space_one', disabledFeatures: [] });
+      await spacesService.create({ id: idSpace2, name: 'space_two', disabledFeatures: [] });
+
+      await ml.api.createAndRunAnomalyDetectionLookbackJob(
+        // @ts-expect-error not full interface
+        testJobConfig,
+        testDatafeedConfig,
+        idSpace1
+      );
+    });
+
+    after(async () => {
+      await spacesService.delete(idSpace1);
+      await spacesService.delete(idSpace2);
+      await ml.api.cleanMlIndices();
+    });
+
+    it('should produce the correct category for the job', async () => {
+      const resp = await getCategoryDefinition(
+        jobIdSpace1,
+        expectedCategoryDefinition.categoryId,
+        USER.ML_POWERUSER,
+        200,
+        idSpace1
+      );
+
+      expect(resp.categoryId).to.eql(
+        expectedCategoryDefinition.categoryId,
+        `categoryId should be ${expectedCategoryDefinition.categoryId} (got ${resp.categoryId})`
+      );
+      expect(resp.examples.length).to.eql(
+        expectedCategoryDefinition.examplesLength,
+        `examples list length should be ${expectedCategoryDefinition.examplesLength} (got ${resp.examples.length})`
+      );
+      expect(resp.terms.length).to.be.greaterThan(
+        0,
+        `terms string length should be greater than 0 (got ${resp.terms.length})`
+      );
+      expect(resp.regex.length).to.be.greaterThan(
+        0,
+        `regex string length should be greater than 0 (got ${resp.regex.length})`
+      );
+    });
+
+    it('should not produce the correct category for the job in the wrong space', async () => {
+      await getCategoryDefinition(
+        jobIdSpace1,
+        expectedCategoryDefinition.categoryId,
+        USER.ML_POWERUSER,
+        404,
+        idSpace2
+      );
+    });
+
+    it('should produce the correct category for ml viewer user', async () => {
+      const resp = await getCategoryDefinition(
+        jobIdSpace1,
+        expectedCategoryDefinition.categoryId,
+        USER.ML_VIEWER,
+        200,
+        idSpace1
+      );
+
+      expect(resp.categoryId).to.eql(
+        expectedCategoryDefinition.categoryId,
+        `categoryId should be ${expectedCategoryDefinition.categoryId} (got ${resp.categoryId})`
+      );
+      expect(resp.examples.length).to.eql(
+        expectedCategoryDefinition.examplesLength,
+        `examples list length should be ${expectedCategoryDefinition.examplesLength} (got ${resp.examples.length})`
+      );
+      expect(resp.terms.length).to.be.greaterThan(
+        0,
+        `terms string length should be greater than 0 (got ${resp.terms.length})`
+      );
+      expect(resp.regex.length).to.be.greaterThan(
+        0,
+        `regex string length should be greater than 0 (got ${resp.regex.length})`
+      );
+    });
+
+    it('should not produce the correct category for ml unauthorized user', async () => {
+      await getCategoryDefinition(
+        jobIdSpace1,
+        expectedCategoryDefinition.categoryId,
+        USER.ML_UNAUTHORIZED,
+        403,
+        idSpace1
+      );
+    });
+  });
+};

x-pack/test/api_integration/apis/ml/results/get_category_examples.ts

@@ -0,0 +1,178 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0; you may not use this file except in compliance with the Elastic License
+ * 2.0.
+ */
+
+import expect from '@kbn/expect';
+import { FtrProviderContext } from '../../../ftr_provider_context';
+import { USER } from '../../../../functional/services/ml/security_common';
+import { COMMON_REQUEST_HEADERS } from '../../../../functional/services/ml/common_api';
+import { Datafeed } from '../../../../../plugins/ml/common/types/anomaly_detection_jobs';
+
+export default ({ getService }: FtrProviderContext) => {
+  const esArchiver = getService('esArchiver');
+  const supertest = getService('supertestWithoutAuth');
+  const ml = getService('ml');
+  const spacesService = getService('spaces');
+
+  const jobIdSpace1 = `sample_logs_${Date.now()}`;
+  const idSpace1 = 'space1';
+  const idSpace2 = 'space2';
+
+  const PARTITION_FIELD_NAME = 'event.dataset';
+  const testJobConfig = {
+    job_id: jobIdSpace1,
+    groups: ['sample_logs', 'bootstrap', 'categorization'],
+    description: "count by mlcategory (message) on 'sample logs' dataset with 15m bucket span",
+    analysis_config: {
+      bucket_span: '15m',
+      categorization_field_name: 'message',
+      per_partition_categorization: { enabled: true, stop_on_warn: true },
+      detectors: [
+        {
+          function: 'count',
+          by_field_name: 'mlcategory',
+          partition_field_name: PARTITION_FIELD_NAME,
+        },
+      ],
+      influencers: ['mlcategory'],
+    },
+    analysis_limits: { model_memory_limit: '26MB' },
+    data_description: { time_field: '@timestamp', time_format: 'epoch_ms' },
+    model_plot_config: { enabled: false, annotations_enabled: true },
+    model_snapshot_retention_days: 10,
+    daily_model_snapshot_retention_after_days: 1,
+    allow_lazy_open: false,
+  };
+  // @ts-expect-error not full interface
+  const testDatafeedConfig: Datafeed = {
+    datafeed_id: `datafeed-${jobIdSpace1}`,
+    indices: ['ft_module_sample_logs'],
+    job_id: jobIdSpace1,
+    query: { bool: { must: [{ match_all: {} }] } },
+  };
+
+  const expectedCategoryExamples = {
+    categoryId: '1',
+    examplesLength: 3,
+  };
+
+  async function getCategoryExamples(
+    jobId: string,
+    categoryIds: string[],
+    maxExamples: number,
+    user: USER,
+    expectedStatusCode: number,
+    space?: string
+  ) {
+    const { body } = await supertest
+      .post(`${space ? `/s/${space}` : ''}/api/ml/results/category_examples`)
+      .auth(user, ml.securityCommon.getPasswordForUser(user))
+      .set(COMMON_REQUEST_HEADERS)
+      .send({ jobId, categoryIds, maxExamples })
+      .expect(expectedStatusCode);
+
+    return body;
+  }
+
+  describe('get category_examples', function () {
+    before(async () => {
+      await esArchiver.loadIfNeeded('x-pack/test/functional/es_archives/ml/module_sample_logs');
+      await ml.testResources.setKibanaTimeZoneToUTC();
+      await spacesService.create({ id: idSpace1, name: 'space_one', disabledFeatures: [] });
+      await spacesService.create({ id: idSpace2, name: 'space_two', disabledFeatures: [] });
+
+      await ml.api.createAndRunAnomalyDetectionLookbackJob(
+        // @ts-expect-error not full interface
+        testJobConfig,
+        testDatafeedConfig,
+        idSpace1
+      );
+    });
+
+    after(async () => {
+      await spacesService.delete(idSpace1);
+      await spacesService.delete(idSpace2);
+      await ml.api.cleanMlIndices();
+    });
+
+    it('should produce the correct 1 example for the job', async () => {
+      const maxExamples = 1;
+      const resp = await getCategoryExamples(
+        jobIdSpace1,
+        [expectedCategoryExamples.categoryId],
+        maxExamples,
+        USER.ML_POWERUSER,
+        200,
+        idSpace1
+      );
+
+      expect(resp[expectedCategoryExamples.categoryId].length).to.eql(
+        maxExamples,
+        `response examples length should be ${maxExamples} (got ${
+          resp[expectedCategoryExamples.categoryId].length
+        })`
+      );
+    });
+
+    it('should produce the correct 3 examples for the job', async () => {
+      const resp = await getCategoryExamples(
+        jobIdSpace1,
+        [expectedCategoryExamples.categoryId],
+        expectedCategoryExamples.examplesLength,
+        USER.ML_POWERUSER,
+        200,
+        idSpace1
+      );
+
+      expect(resp[expectedCategoryExamples.categoryId].length).to.eql(
+        expectedCategoryExamples.examplesLength,
+        `response examples length should be ${expectedCategoryExamples.examplesLength} (got ${
+          resp[expectedCategoryExamples.categoryId].length
+        })`
+      );
+    });
+
+    it('should not produce the correct examples for the job in the wrong space', async () => {
+      await getCategoryExamples(
+        jobIdSpace1,
+        [expectedCategoryExamples.categoryId],
+        expectedCategoryExamples.examplesLength,
+        USER.ML_POWERUSER,
+        404,
+        idSpace2
+      );
+    });
+
+    it('should produce the correct example for the job for the ml viewer user', async () => {
+      const resp = await getCategoryExamples(
+        jobIdSpace1,
+        [expectedCategoryExamples.categoryId],
+        expectedCategoryExamples.examplesLength,
+        USER.ML_VIEWER,
+        200,
+        idSpace1
+      );
+
+      expect(resp[expectedCategoryExamples.categoryId].length).to.eql(
+        expectedCategoryExamples.examplesLength,
+        `response examples length should be ${expectedCategoryExamples.examplesLength} (got ${
+          resp[expectedCategoryExamples.categoryId].length
+        })`
+      );
+    });
+
+    it('should not produce the correct example for the job for the ml unauthorized user', async () => {
+      await getCategoryExamples(
+        jobIdSpace1,
+        [expectedCategoryExamples.categoryId],
+        expectedCategoryExamples.examplesLength,
+        USER.ML_UNAUTHORIZED,
+        403,
+        idSpace1
+      );
+    });
+  });
+};

x-pack/test/api_integration/apis/ml/results/index.ts

@@ -12,5 +12,7 @@ export default function ({ loadTestFile }: FtrProviderContext) {
     loadTestFile(require.resolve('./get_anomalies_table_data'));
     loadTestFile(require.resolve('./get_categorizer_stats'));
     loadTestFile(require.resolve('./get_stopped_partitions'));
+    loadTestFile(require.resolve('./get_category_definition'));
+    loadTestFile(require.resolve('./get_category_examples'));
   });
 }

x-pack/test/functional/services/ml/api.ts

@@ -681,9 +681,13 @@ export function MachineLearningAPIProvider({ getService }: FtrProviderContext) {
       log.debug('> Datafeed stopped.');
     },
 
-    async createAndRunAnomalyDetectionLookbackJob(jobConfig: Job, datafeedConfig: Datafeed) {
-      await this.createAnomalyDetectionJob(jobConfig);
-      await this.createDatafeed(datafeedConfig);
+    async createAndRunAnomalyDetectionLookbackJob(
+      jobConfig: Job,
+      datafeedConfig: Datafeed,
+      space?: string
+    ) {
+      await this.createAnomalyDetectionJob(jobConfig, space);
+      await this.createDatafeed(datafeedConfig, space);
       await this.openAnomalyDetectionJob(jobConfig.job_id);
       await this.startDatafeed(datafeedConfig.datafeed_id, { start: '0', end: `${Date.now()}` });
       await this.waitForDatafeedState(datafeedConfig.datafeed_id, DATAFEED_STATE.STOPPED);