[ML] Listing all categorization wizard checks (#60502)

* [ML] Listing all categorization wizard checks

* fixing translation

* changes based on review

* moving check

* adding real values to messages

* reordering checks enum

* fixing types

* updating tests

* updating id

Author: jgowdyelastic

x-pack/plugins/ml/common/constants/categorization_job.ts

@@ -0,0 +1,80 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License;
+ * you may not use this file except in compliance with the Elastic License.
+ */
+
+import { i18n } from '@kbn/i18n';
+import { VALIDATION_RESULT } from '../types/categories';
+
+export const NUMBER_OF_CATEGORY_EXAMPLES = 5;
+export const CATEGORY_EXAMPLES_SAMPLE_SIZE = 1000;
+export const CATEGORY_EXAMPLES_WARNING_LIMIT = 0.75;
+export const CATEGORY_EXAMPLES_ERROR_LIMIT = 0.02;
+
+export const VALID_TOKEN_COUNT = 3;
+export const MEDIAN_LINE_LENGTH_LIMIT = 400;
+export const NULL_COUNT_PERCENT_LIMIT = 0.75;
+
+export enum CATEGORY_EXAMPLES_VALIDATION_STATUS {
+  VALID = 'valid',
+  PARTIALLY_VALID = 'partially_valid',
+  INVALID = 'invalid',
+}
+
+export const VALIDATION_CHECK_DESCRIPTION = {
+  [VALIDATION_RESULT.NO_EXAMPLES]: i18n.translate(
+    'xpack.ml.models.jobService.categorization.messages.validNoDataFound',
+    {
+      defaultMessage: 'Examples  were successfully loaded.',
+    }
+  ),
+  [VALIDATION_RESULT.FAILED_TO_TOKENIZE]: i18n.translate(
+    'xpack.ml.models.jobService.categorization.messages.validFailureToGetTokens',
+    {
+      defaultMessage: 'The examples loaded were tokenized successfully.',
+    }
+  ),
+  [VALIDATION_RESULT.TOKEN_COUNT]: i18n.translate(
+    'xpack.ml.models.jobService.categorization.messages.validTokenLength',
+    {
+      defaultMessage:
+        'More than {tokenCount} tokens per example were found in over {percentage}% of the examples loaded.',
+      values: {
+        percentage: Math.floor(CATEGORY_EXAMPLES_WARNING_LIMIT * 100),
+        tokenCount: VALID_TOKEN_COUNT,
+      },
+    }
+  ),
+  [VALIDATION_RESULT.MEDIAN_LINE_LENGTH]: i18n.translate(
+    'xpack.ml.models.jobService.categorization.messages.validMedianLineLength',
+    {
+      defaultMessage:
+        'The median line length of the examples loaded was less than {medianCharCount} characters.',
+      values: {
+        medianCharCount: MEDIAN_LINE_LENGTH_LIMIT,
+      },
+    }
+  ),
+  [VALIDATION_RESULT.NULL_VALUES]: i18n.translate(
+    'xpack.ml.models.jobService.categorization.messages.validNullValues',
+    {
+      defaultMessage: 'Less than {percentage}% of the examples loaded were null.',
+      values: {
+        percentage: Math.floor(100 - NULL_COUNT_PERCENT_LIMIT * 100),
+      },
+    }
+  ),
+  [VALIDATION_RESULT.TOO_MANY_TOKENS]: i18n.translate(
+    'xpack.ml.models.jobService.categorization.messages.validTooManyTokens',
+    {
+      defaultMessage: 'Less than 10000 tokens were found in total in the examples loaded.',
+    }
+  ),
+  [VALIDATION_RESULT.INSUFFICIENT_PRIVILEGES]: i18n.translate(
+    'xpack.ml.models.jobService.categorization.messages.validUserPrivileges',
+    {
+      defaultMessage: 'The user has sufficient privileges to perform the checks.',
+    }
+  ),
+};

x-pack/plugins/ml/common/constants/new_job.ts

@@ -25,15 +25,3 @@ export const DEFAULT_RARE_BUCKET_SPAN = '1h';
 export const DEFAULT_QUERY_DELAY = '60s';
 
 export const SHARED_RESULTS_INDEX_NAME = 'shared';
-
-// Categorization
-export const NUMBER_OF_CATEGORY_EXAMPLES = 5;
-export const CATEGORY_EXAMPLES_SAMPLE_SIZE = 1000;
-export const CATEGORY_EXAMPLES_WARNING_LIMIT = 0.75;
-export const CATEGORY_EXAMPLES_ERROR_LIMIT = 0.02;
-
-export enum CATEGORY_EXAMPLES_VALIDATION_STATUS {
-  VALID = 'valid',
-  PARTIALLY_VALID = 'partially_valid',
-  INVALID = 'invalid',
-}

x-pack/plugins/ml/common/types/categories.ts

@@ -4,7 +4,7 @@
  * you may not use this file except in compliance with the Elastic License.
  */
 
-import { CATEGORY_EXAMPLES_VALIDATION_STATUS } from '../constants/new_job';
+import { CATEGORY_EXAMPLES_VALIDATION_STATUS } from '../constants/categorization_job';
 
 export type CategoryId = number;
 
@@ -39,12 +39,12 @@ export interface CategoryFieldExample {
 }
 
 export enum VALIDATION_RESULT {
+  NO_EXAMPLES,
+  FAILED_TO_TOKENIZE,
+  TOO_MANY_TOKENS,
   TOKEN_COUNT,
   MEDIAN_LINE_LENGTH,
   NULL_VALUES,
-  NO_EXAMPLES,
-  TOO_MANY_TOKENS,
-  FAILED_TO_TOKENIZE,
   INSUFFICIENT_PRIVILEGES,
 }
 

x-pack/plugins/ml/public/application/jobs/new_job/common/job_creator/categorization_job_creator.ts

@@ -16,8 +16,8 @@ import {
   CREATED_BY_LABEL,
   DEFAULT_BUCKET_SPAN,
   DEFAULT_RARE_BUCKET_SPAN,
-  CATEGORY_EXAMPLES_VALIDATION_STATUS,
 } from '../../../../../../common/constants/new_job';
+import { CATEGORY_EXAMPLES_VALIDATION_STATUS } from '../../../../../../common/constants/categorization_job';
 import { ML_JOB_AGGREGATION } from '../../../../../../common/constants/aggregation_types';
 import {
   CategorizationAnalyzer,

x-pack/plugins/ml/public/application/jobs/new_job/common/job_validator/job_validator.ts

@@ -16,7 +16,7 @@ import { JobCreator, JobCreatorType, isCategorizationJobCreator } from '../job_c
 import { populateValidationMessages, checkForExistingJobAndGroupIds } from './util';
 import { ExistingJobsAndGroups } from '../../../../services/job_service';
 import { cardinalityValidator, CardinalityValidatorResult } from './validators';
-import { CATEGORY_EXAMPLES_VALIDATION_STATUS } from '../../../../../../common/constants/new_job';
+import { CATEGORY_EXAMPLES_VALIDATION_STATUS } from '../../../../../../common/constants/categorization_job';
 
 // delay start of validation to allow the user to make changes
 // e.g. if they are typing in a new value, try not to validate

x-pack/plugins/ml/public/application/jobs/new_job/common/results_loader/categorization_examples_loader.ts

@@ -11,7 +11,7 @@ import { ml } from '../../../../services/ml_api_service';
 import {
   NUMBER_OF_CATEGORY_EXAMPLES,
   CATEGORY_EXAMPLES_VALIDATION_STATUS,
-} from '../../../../../../common/constants/new_job';
+} from '../../../../../../common/constants/categorization_job';
 
 export class CategorizationExamplesLoader {
   private _jobCreator: CategorizationJobCreator;

x-pack/plugins/ml/public/application/jobs/new_job/pages/components/job_details_step/components/additional_section/additional_section.tsx

@@ -5,11 +5,17 @@
  */
 
 import React, { FC, Fragment } from 'react';
+import { i18n } from '@kbn/i18n';
 import { EuiAccordion, EuiFlexGroup, EuiFlexItem, EuiSpacer } from '@elastic/eui';
 import { CalendarsSelection } from './components/calendars';
 import { CustomUrlsSelection } from './components/custom_urls';
 
-const ButtonContent = <Fragment>Additional settings</Fragment>;
+const buttonContent = i18n.translate(
+  'xpack.ml.newJob.wizard.jobDetailsStep.additionalSectionButton',
+  {
+    defaultMessage: 'Additional settings',
+  }
+);
 
 interface Props {
   additionalExpanded: boolean;
@@ -22,7 +28,7 @@ export const AdditionalSection: FC<Props> = ({ additionalExpanded, setAdditional
       <EuiSpacer />
       <EuiAccordion
         id="advanced-section"
-        buttonContent={ButtonContent}
+        buttonContent={buttonContent}
         onToggle={setAdditionalExpanded}
         initialIsOpen={additionalExpanded}
         data-test-subj="mlJobWizardToggleAdditionalSettingsSection"

x-pack/plugins/ml/public/application/jobs/new_job/pages/components/job_details_step/components/advanced_section/advanced_section.tsx

@@ -19,7 +19,7 @@ import { ModelMemoryLimitInput } from '../../../common/model_memory_limit';
 import { JobCreatorContext } from '../../../job_creator_context';
 import { JOB_TYPE } from '../../../../../../../../../common/constants/new_job';
 
-const ButtonContent = i18n.translate(
+const buttonContent = i18n.translate(
   'xpack.ml.newJob.wizard.jobDetailsStep.advancedSectionButton',
   {
     defaultMessage: 'Advanced',
@@ -55,7 +55,7 @@ export const AdvancedSection: FC<Props> = ({ advancedExpanded, setAdvancedExpand
       <EuiSpacer size="xl" />
       <EuiAccordion
         id="advanced-section"
-        buttonContent={ButtonContent}
+        buttonContent={buttonContent}
         onToggle={setAdvancedExpanded}
         initialIsOpen={advancedExpanded}
         data-test-subj="mlJobWizardToggleAdvancedSection"

x-pack/plugins/ml/public/application/jobs/new_job/pages/components/pick_fields_step/components/categorization_view/examples_valid_callout.tsx

@@ -5,7 +5,14 @@
  */
 
 import React, { FC } from 'react';
-import { EuiCallOut, EuiSpacer, EuiCallOutProps } from '@elastic/eui';
+import {
+  EuiCallOut,
+  EuiSpacer,
+  EuiCallOutProps,
+  EuiAccordion,
+  EuiListGroup,
+  EuiListGroupItemProps,
+} from '@elastic/eui';
 import { i18n } from '@kbn/i18n';
 import { FormattedMessage } from '@kbn/i18n/react';
 
@@ -14,14 +21,25 @@ import {
   FieldExampleCheck,
 } from '../../../../../../../../../common/types/categories';
 import { EditCategorizationAnalyzerFlyout } from '../../../common/edit_categorization_analyzer_flyout';
-import { CATEGORY_EXAMPLES_VALIDATION_STATUS } from '../../../../../../../../../common/constants/new_job';
+import {
+  CATEGORY_EXAMPLES_VALIDATION_STATUS,
+  VALIDATION_CHECK_DESCRIPTION,
+} from '../../../../../../../../../common/constants/categorization_job';
+import { VALIDATION_RESULT } from '../../../../../../../../../common/types/categories';
 
 interface Props {
   validationChecks: FieldExampleCheck[];
   overallValidStatus: CATEGORY_EXAMPLES_VALIDATION_STATUS;
   categorizationAnalyzer: CategorizationAnalyzer;
 }
 
+const allChecksButtonContent = i18n.translate(
+  'xpack.ml.newJob.wizard.jobDetailsStep.allChecksButton',
+  {
+    defaultMessage: 'View all checks performed',
+  }
+);
+
 export const ExamplesValidCallout: FC<Props> = ({
   overallValidStatus,
   validationChecks,
@@ -66,6 +84,10 @@ export const ExamplesValidCallout: FC<Props> = ({
       ))}
       <EuiSpacer size="s" />
       {analyzerUsed}
+      <EuiSpacer size="s" />
+      <EuiAccordion id="all-checks" buttonContent={allChecksButtonContent}>
+        <AllValidationChecks validationChecks={validationChecks} />
+      </EuiAccordion>
     </EuiCallOut>
   );
 };
@@ -96,3 +118,28 @@ const AnalyzerUsed: FC<{ categorizationAnalyzer: CategorizationAnalyzer }> = ({
     </>
   );
 };
+
+const AllValidationChecks: FC<{ validationChecks: FieldExampleCheck[] }> = ({
+  validationChecks,
+}) => {
+  const list: EuiListGroupItemProps[] = Object.keys(VALIDATION_CHECK_DESCRIPTION).map((k, i) => {
+    const failedCheck = validationChecks.find(vc => vc.id === i);
+    if (
+      failedCheck !== undefined &&
+      failedCheck?.valid !== CATEGORY_EXAMPLES_VALIDATION_STATUS.VALID
+    ) {
+      return {
+        iconType: 'cross',
+        label: failedCheck.message,
+        size: 's',
+      };
+    }
+    return {
+      iconType: 'check',
+      label: VALIDATION_CHECK_DESCRIPTION[i as VALIDATION_RESULT],
+      size: 's',
+    };
+  });
+
+  return <EuiListGroup listItems={list} maxWidth={false} />;
+};

x-pack/plugins/ml/public/application/jobs/new_job/pages/components/pick_fields_step/components/categorization_view/metric_selection.tsx

@@ -18,7 +18,7 @@ import {
   CategoryFieldExample,
   FieldExampleCheck,
 } from '../../../../../../../../../common/types/categories';
-import { CATEGORY_EXAMPLES_VALIDATION_STATUS } from '../../../../../../../../../common/constants/new_job';
+import { CATEGORY_EXAMPLES_VALIDATION_STATUS } from '../../../../../../../../../common/constants/categorization_job';
 import { LoadingWrapper } from '../../../charts/loading_wrapper';
 
 interface Props {