fix(ava/insight): modify parameter definition

antvis · Nov 3, 2023 · 850f8d9 · 850f8d9
1 parent 5bd88b2
commit 850f8d9
Show file tree

Hide file tree

Showing 18 changed files with 66 additions and 29 deletions.
diff --git a/packages/ava/__tests__/unit/insight/extractors/category-outlier.test.ts b/packages/ava/__tests__/unit/insight/extractors/category-outlier.test.ts
@@ -39,7 +39,7 @@ describe('extract category-outlier insight', () => {
   test('check outliers result', () => {
     const result = insightPatternsExtractor({
       data,
-      dimensions: ['type'],
+      dimensions: [{ fieldName: 'type' }],
       measures: [{ fieldName: 'sales', method: 'SUM' }],
       insightType: 'category_outlier',
       options: {

diff --git a/packages/ava/__tests__/unit/insight/extractors/change-point.test.ts b/packages/ava/__tests__/unit/insight/extractors/change-point.test.ts
@@ -16,7 +16,7 @@ describe('extract change-point insight', () => {
   test('check change-point result', () => {
     const result = insightPatternsExtractor({
       data,
-      dimensions: ['year'],
+      dimensions: [{ fieldName: 'year' }],
       measures: [{ fieldName: 'value', method: 'SUM' }],
       insightType: 'change_point',
       options: {

diff --git a/packages/ava/__tests__/unit/insight/extractors/low-variance.test.ts b/packages/ava/__tests__/unit/insight/extractors/low-variance.test.ts
@@ -39,7 +39,7 @@ describe('extract low-variance insight', () => {
   test('check low-variance result', () => {
     const result = insightPatternsExtractor({
       data,
-      dimensions: ['type'],
+      dimensions: [{ fieldName: 'type' }],
       measures: [{ fieldName: 'sales', method: 'SUM' }],
       insightType: 'low_variance',
       options: {

diff --git a/packages/ava/__tests__/unit/insight/extractors/majority.test.ts b/packages/ava/__tests__/unit/insight/extractors/majority.test.ts
@@ -39,7 +39,7 @@ describe('extract majority insight', () => {
   test('check majority result', () => {
     const result = insightPatternsExtractor({
       data,
-      dimensions: ['type'],
+      dimensions: [{ fieldName: 'type' }],
       measures: [{ fieldName: 'sales', method: 'SUM' }],
       insightType: 'majority',
       options: {

diff --git a/packages/ava/__tests__/unit/insight/extractors/time-series-outlier.test.ts b/packages/ava/__tests__/unit/insight/extractors/time-series-outlier.test.ts
@@ -16,7 +16,7 @@ describe('extract time-series-outlier insight', () => {
   test('check outliers result', () => {
     const result = insightPatternsExtractor({
       data,
-      dimensions: ['year'],
+      dimensions: [{ fieldName: 'year' }],
       measures: [{ fieldName: 'value', method: 'SUM' }],
       insightType: 'time_series_outlier',
       options: {

diff --git a/packages/ava/__tests__/unit/insight/extractors/trend.test.ts b/packages/ava/__tests__/unit/insight/extractors/trend.test.ts
@@ -16,7 +16,7 @@ describe('extract trend insight', () => {
   test('check trend result', () => {
     const result = insightPatternsExtractor({
       data,
-      dimensions: ['year'],
+      dimensions: [{ fieldName: 'year' }],
       measures: [{ fieldName: 'value', method: 'SUM' }],
       insightType: 'trend',
       options: {

diff --git a/packages/ava/src/insight/constant.ts b/packages/ava/src/insight/constant.ts
@@ -3,11 +3,11 @@ export const SIGNIFICANCE_BENCHMARK = 0.95;
 
 export const SIGNIFICANCE_LEVEL = 0.05;
 
-export const InsightScoreBenchmark = 0.01;
+export const INSIGHT_SCORE_BENCHMARK = 0.01;
 
 export const IMPACT_SCORE_WEIGHT = 0.2;
 
-export const InsightDefaultLimit = 20;
+export const INSIGHT_DEFAULT_LIMIT = 20;
 
 export const IQR_K = 1.5;
 

diff --git a/packages/ava/src/insight/insights/extractors/categoryOutlier.ts b/packages/ava/src/insight/insights/extractors/categoryOutlier.ts
@@ -1,4 +1,4 @@
-import { get, isString, orderBy } from 'lodash';
+import { get, isNil, isString, orderBy } from 'lodash';
 
 import { distinct, mean } from '../../../data';
 import { categoryOutlier } from '../../algorithms';
@@ -76,6 +76,13 @@ export const getCategoryOutlierInfo: GetPatternInfo<CategoryOutlierInfo> = (prop
   if (isString(valid))
     return getNonSignificantInsight({ detailInfo: valid, insightType, infoType: 'verificationFailure' });
   const { dimension, values, measure } = getAlgorithmCommonInput(props);
+  if (isNil(dimension) || isNil(measure))
+    return getNonSignificantInsight({
+      detailInfo: 'Measure or dimension is empty.',
+      insightType,
+      infoType: 'verificationFailure',
+    });
+
   if (distinct(values) === 1)
     return getNonSignificantInsight({
       insightType,

diff --git a/packages/ava/src/insight/insights/extractors/changePoint.ts b/packages/ava/src/insight/insights/extractors/changePoint.ts
@@ -1,4 +1,4 @@
-import { get, isString } from 'lodash';
+import { get, isNil, isString } from 'lodash';
 
 import { changePoint } from '../../algorithms';
 import { getAlgorithmCommonInput, getNonSignificantInsight, preValidation } from '../util';
@@ -34,6 +34,13 @@ export const getChangePointInfo: GetPatternInfo<ChangePointInfo> = (props) => {
 
   const { data } = props;
   const { dimension, values, measure } = getAlgorithmCommonInput(props);
+  if (isNil(dimension) || isNil(measure))
+    return getNonSignificantInsight({
+      detailInfo: 'Measure or dimension is empty.',
+      insightType,
+      infoType: 'verificationFailure',
+    });
+
   const changePointsParameter = get(props, 'options.algorithmParameter.changePoint');
   const changePoints = findChangePoints(values, changePointsParameter);
   if (changePoints.length === 0) {

diff --git a/packages/ava/src/insight/insights/extractors/lowVariance.ts b/packages/ava/src/insight/insights/extractors/lowVariance.ts
@@ -1,4 +1,4 @@
-import { get, isString } from 'lodash';
+import { get, isNil, isString } from 'lodash';
 
 import { coefficientOfVariance, mean } from '../../../data';
 import { getAlgorithmCommonInput, getNonSignificantInsight, preValidation } from '../util';
@@ -36,6 +36,13 @@ export const getLowVarianceInfo: GetPatternInfo<LowVarianceInfo> = (props) => {
   if (isString(valid))
     return getNonSignificantInsight({ detailInfo: valid, insightType, infoType: 'verificationFailure' });
   const { dimension, values, measure } = getAlgorithmCommonInput(props);
+  if (isNil(dimension) || isNil(measure))
+    return getNonSignificantInsight({
+      detailInfo: 'Measure or dimension is empty.',
+      insightType,
+      infoType: 'verificationFailure',
+    });
+
   const lowVarianceParameter = get(props, 'options.algorithmParameter.lowVariance');
   const lowVariance = findLowVariance(values, lowVarianceParameter);
   if (lowVariance) {

diff --git a/packages/ava/src/insight/insights/extractors/majority.ts b/packages/ava/src/insight/insights/extractors/majority.ts
@@ -1,4 +1,4 @@
-import { get, isString } from 'lodash';
+import { get, isNil, isString } from 'lodash';
 
 import { getAlgorithmCommonInput, getNonSignificantInsight, preValidation } from '../util';
 
@@ -47,6 +47,13 @@ export const getMajorityInfo: GetPatternInfo<MajorityInfo> = (props) => {
     return getNonSignificantInsight({ detailInfo: valid, insightType, infoType: 'verificationFailure' });
   const { data } = props;
   const { dimension, values, measure } = getAlgorithmCommonInput(props);
+  if (isNil(dimension) || isNil(measure))
+    return getNonSignificantInsight({
+      detailInfo: 'Measure or dimension is empty.',
+      insightType,
+      infoType: 'verificationFailure',
+    });
+
   const majorityParameter = get(props, 'options.algorithmParameter.majority');
   const majority = findMajority(values, majorityParameter);
   if (majority) {

diff --git a/packages/ava/src/insight/insights/extractors/timeSeriesOutlier.ts b/packages/ava/src/insight/insights/extractors/timeSeriesOutlier.ts
@@ -1,4 +1,4 @@
-import { get, isString } from 'lodash';
+import { get, isNil, isString } from 'lodash';
 
 import { distinct, lowess } from '../../../data';
 import { LowessOutput } from '../../../data/statistics/types';
@@ -41,6 +41,12 @@ export const getTimeSeriesOutlierInfo: GetPatternInfo<TimeSeriesOutlierInfo> = (
 
   const { data } = props;
   const { dimension, values, measure } = getAlgorithmCommonInput(props);
+  if (isNil(dimension) || isNil(measure))
+    return getNonSignificantInsight({
+      detailInfo: 'Measure or dimension is empty.',
+      insightType,
+      infoType: 'verificationFailure',
+    });
 
   if (distinct(values) === 1) return getNonSignificantInsight({ insightType, infoType: 'noInsight' });
 

diff --git a/packages/ava/src/insight/insights/extractors/trend.ts b/packages/ava/src/insight/insights/extractors/trend.ts
@@ -1,5 +1,5 @@
 import regression from 'regression';
-import { get, isString } from 'lodash';
+import { get, isNil, isString } from 'lodash';
 
 import { CommonParameter, GetPatternInfo, TrendInfo } from '../../types';
 import { trendDirection } from '../../algorithms';
@@ -37,6 +37,13 @@ export const getTrendInfo: GetPatternInfo<TrendInfo> = (props) => {
     return getNonSignificantInsight({ detailInfo: valid, insightType, infoType: 'verificationFailure' });
 
   const { dimension, values, measure } = getAlgorithmCommonInput(props);
+  if (isNil(dimension) || isNil(measure))
+    return getNonSignificantInsight({
+      detailInfo: 'Measure or dimension is empty.',
+      insightType,
+      infoType: 'verificationFailure',
+    });
+
   const trendParameter = get(props, 'options.algorithmParameter.trend');
   const result: TrendResult = findTimeSeriesTrend(values, trendParameter);
   return [

diff --git a/packages/ava/src/insight/insights/util.ts b/packages/ava/src/insight/insights/util.ts
@@ -49,8 +49,8 @@ export const getAlgorithmCommonInput = ({
   dimensions,
   measures,
 }: InsightExtractorProps): AlgorithmStandardInput => {
-  const dimension = dimensions[0];
-  const measure = measures[0].fieldName;
+  const dimension = dimensions[0]?.fieldName;
+  const measure = measures[0]?.fieldName;
   const values = data.map((item) => item?.[measure] as number);
   return { dimension, measure, values };
 };
@@ -75,7 +75,7 @@ export const preValidation = ({
   if (!checker) return true;
   const result = checker({
     data,
-    subjectInfo: { dimensions, measures, subspace: [] },
+    subjectInfo: { dimensions: dimensions?.map((dim) => dim.fieldName), measures, subspace: [] },
     fieldPropsMap,
   });
   return result;

diff --git a/packages/ava/src/insight/pipeline/extract.ts b/packages/ava/src/insight/pipeline/extract.ts
@@ -1,7 +1,7 @@
 import { groupBy, uniq, flatten, isString } from 'lodash';
 import Heap from 'heap-js';
 
-import { PATTERN_TYPES, InsightScoreBenchmark, IMPACT_SCORE_WEIGHT } from '../constant';
+import { PATTERN_TYPES, INSIGHT_SCORE_BENCHMARK, IMPACT_SCORE_WEIGHT } from '../constant';
 import { insightPatternsExtractor, ExtractorCheckers } from '../insights';
 import { aggregate } from '../utils/aggregate';
 import {
@@ -83,7 +83,7 @@ function extractPatternsFromSubject(
       };
       const extractedPatterns = insightPatternsExtractor({
         data,
-        dimensions,
+        dimensions: dimensions.map((dim) => ({ fieldName: dim })),
         measures,
         insightType,
         options: extractorOptions,
@@ -209,7 +209,7 @@ export function extractInsightsFromSubspace(
   const subspaceImpact = computeSubspaceImpact(data, subspace, impactMeasureReferences, options?.impactMeasures);
 
   // pruning1: check the subpace impact limit
-  if (subspaceImpact < InsightScoreBenchmark) {
+  if (subspaceImpact < INSIGHT_SCORE_BENCHMARK) {
     return [];
   }
 

diff --git a/packages/ava/src/insight/pipeline/insight.ts b/packages/ava/src/insight/pipeline/insight.ts
@@ -1,6 +1,6 @@
 import Heap from 'heap-js';
 
-import { InsightDefaultLimit } from '../constant';
+import { INSIGHT_DEFAULT_LIMIT } from '../constant';
 import { aggregateWithSeries, aggregateWithMeasures } from '../utils/aggregate';
 
 import { enumerateInsights } from './extract';
@@ -61,7 +61,7 @@ export function extractInsights(sourceData: Datum[], options?: InsightOptions):
   // init insights storage
   const insightsHeap = new Heap(insightPriorityComparator);
   const homogeneousInsightsHeap = new Heap(homogeneousInsightPriorityComparator);
-  const insightsLimit = options?.limit || InsightDefaultLimit;
+  const insightsLimit = options?.limit || INSIGHT_DEFAULT_LIMIT;
   insightsHeap.limit = insightsLimit;
   insightsHeap.init([]);
   homogeneousInsightsHeap.init([]);

diff --git a/packages/ava/src/insight/pipeline/specificInsight.ts b/packages/ava/src/insight/pipeline/specificInsight.ts
@@ -12,11 +12,7 @@ export const patternInfo2InsightInfo = (props: PatternInfo2InsightInfoProps) =>
   const score = patternInfos[0] ? patternInfos[0].significance * (1 - IMPACT_SCORE_WEIGHT) + IMPACT_SCORE_WEIGHT : 0;
   return {
     subspace: [],
-    dimensions: [
-      {
-        fieldName: dimensions[0],
-      },
-    ],
+    dimensions,
     measures,
     patterns: patternInfos,
     data,

diff --git a/packages/ava/src/insight/types.ts b/packages/ava/src/insight/types.ts
@@ -175,7 +175,7 @@ export type InsightExtractorOptions = {
 
 export type InsightExtractorProps = {
   data: Datum[];
-  dimensions: string[];
+  dimensions: Dimension[];
   measures: Measure[];
   insightType: InsightType;
   options?: InsightExtractorOptions;
@@ -288,7 +288,7 @@ export type InsightsResult = {
 
 export type SpecificInsightProps = {
   data: Datum[];
-  dimensions: string[];
+  dimensions: Dimension[];
   measures: Measure[];
   insightType: InsightType;
   /** 包括自定义的算法参数，是否进行数据校验等内容 */