uc-cdis · ocshawn · May 5, 2023 · Apr 18, 2023 · May 1, 2023 · May 4, 2023
diff --git a/doc/queries.md b/doc/queries.md
@@ -8,6 +8,7 @@ Table of Contents
    - [Numeric Aggregation](#aggs-numeric)
    - [Nested Aggregation](#aggs-nested)
    - [Sub-aggregations](#aggs-sub)
+   - [Cardinality Count Aggregation](#aggs-cardinality)
 - [Filters](#filter)
    - [Basic Filter Unit](#filter-unit)
    - [Text Search Unit in Filter](#filter-search)
@@ -765,6 +766,48 @@ Result:
 }
 ```
 
+<a name="aggs-cardinality"></a>
+
+### 6. Cardinality Count Aggregation
+ By using `_cardinalityCount` keyword, return a cardinality count of a field.
+
+ See [Elasticsearch documentation on Cardinality](https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-metrics-cardinality-aggregation.html)
+
+> **Note**
+> curenntly does not support nested feilds
+
+ Example:
+
+ ```
+ query ($filter: JSON) {
+  _aggregation  {
+    subject(filter: $filter) {
+      submitter_id {
+        _cardinalityCount(
+          precision_threshold: 1000 //optional defaults to 3000
+        )
+      }
+    }
+  }
+}
+```
+
+Example result:
+
+```
+{
+	"data": {
+		"_aggregation": {
+			"subject": {
+				"submitter_id": {
+					"_cardinalityCount": 98
+				}
+			}
+		}
+	}
+}
+```
+
 <a name="filter"></a>
 
 ## Filters

diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@gen3/guppy",
-  "version": "0.15.5",
+  "version": "0.15.7",
   "description": "Server that support GraphQL queries on data from elasticsearch",
   "main": "src/server/server.js",
   "directories": {

diff --git a/src/components/ConnectedFilter/index.jsx b/src/components/ConnectedFilter/index.jsx
@@ -46,6 +46,7 @@ class ConnectedFilter extends React.Component {
 
     this.state = {
       allFields,
+      countFields: this.props.extraAggsFieldsCardinalityCount,
       initialAggsData: {},
       receivedAggsData: {},
       accessibility: ENUM_ACCESSIBILITY.ALL,
@@ -70,6 +71,7 @@ class ConnectedFilter extends React.Component {
       this.props.guppyConfig.path,
       this.props.guppyConfig.type,
       this.state.allFields,
+      this.state.countFields,
       this.state.accessibility,
       this.state.filter,
     )
@@ -126,6 +128,7 @@ class ConnectedFilter extends React.Component {
       this.props.guppyConfig.path,
       this.props.guppyConfig.type,
       this.state.allFields,
+      this.state.countFields,
       mergedFilterResults,
       this.state.accessibility,
     )
@@ -317,6 +320,7 @@ ConnectedFilter.propTypes = {
     })),
   }).isRequired,
   extraAggsFields: PropTypes.arrayOf(PropTypes.string),
+  extraAggsFieldsCardinalityCount: PropTypes.arrayOf(PropTypes.string),
   guppyConfig: PropTypes.shape({
     path: PropTypes.string.isRequired,
     type: PropTypes.string.isRequired,
@@ -344,6 +348,7 @@ ConnectedFilter.propTypes = {
 
 ConnectedFilter.defaultProps = {
   extraAggsFields: [],
+  extraAggsFieldsCardinalityCount: [],
   onFilterChange: () => {},
   onReceiveNewAggsData: () => {},
   className: '',

diff --git a/src/components/Utils/queries.js b/src/components/Utils/queries.js
@@ -23,31 +23,52 @@ const histogramQueryStrForEachField = (field) => {
   }`);
 };
 
-const queryGuppyForAggs = (path, type, fields, gqlFilter, acc) => {
+const cardinalityQueryStrForEachField = (field) => {
+  const splittedFieldArray = field.split('.');
+  const splittedField = splittedFieldArray.shift();
+
+  if (splittedFieldArray.length === 0) {
+    return (`
+      ${splittedField} {
+        _cardinalityCount
+      }
+      `);
+  }
+  return (`
+  ${splittedField} {
+    ${cardinalityQueryStrForEachField(splittedFieldArray.join('.'))}
+  }`);
+};
+
+const queryGuppyForAggs = (path, type, fields, cardinalityFields = [], gqlFilter, acc) => {
   let accessibility = acc;
   if (accessibility !== 'all' && accessibility !== 'accessible' && accessibility !== 'unaccessible') {
     accessibility = 'all';
   }
 
-  const query = `query {
-    _aggregation {
-      ${type} (accessibility: ${accessibility}) {
-        ${fields.map((field) => histogramQueryStrForEachField(field))}
-      }
-    }
-  }`;
-  const queryBody = { query };
+  const queryBody = {};
   if (gqlFilter) {
     const queryWithFilter = `query ($filter: JSON) {
       _aggregation {
         ${type} (filter: $filter, filterSelf: false, accessibility: ${accessibility}) {
-          ${fields.map((field) => histogramQueryStrForEachField(field))}
+          ${fields.map((field) => histogramQueryStrForEachField(field))},
+          ${cardinalityFields.map((field) => cardinalityQueryStrForEachField(field))}
         }
       }
     }`;
     queryBody.variables = { filter: gqlFilter };
     queryBody.query = queryWithFilter;
+  } else {
+    queryBody.query = `query {
+      _aggregation {
+        ${type} (accessibility: ${accessibility}) {
+          ${fields.map((field) => histogramQueryStrForEachField(field))}
+          ${cardinalityFields.map((field) => cardinalityQueryStrForEachField(field))}
+        }
+      }
+    }`;
   }
+
   return fetch(`${path}${graphqlEndpoint}`, {
     method: 'POST',
     headers: {
@@ -265,18 +286,25 @@ export const getGQLFilter = (filterObj) => {
 };
 
 export const askGuppyAboutAllFieldsAndOptions = (
-  path, type, fields, accessibility, filter,
+  path, type, fields, countFields, accessibility, filter,
 ) => {
   const gqlFilter = getGQLFilter(filter);
-  return queryGuppyForAggs(path, type, fields, gqlFilter, accessibility);
+  return queryGuppyForAggs(path, type, fields, countFields, gqlFilter, accessibility);
 };
 
 // eslint-disable-next-line max-len
 export const askGuppyAboutArrayTypes = (path) => queryGuppyForStatus(path).then((res) => res.indices);
 
-export const askGuppyForAggregationData = (path, type, fields, filter, accessibility) => {
+export const askGuppyForAggregationData = (
+  path,
+  type,
+  fields,
+  countFields,
+  filter,
+  accessibility,
+) => {
   const gqlFilter = getGQLFilter(filter);
-  return queryGuppyForAggs(path, type, fields, gqlFilter, accessibility);
+  return queryGuppyForAggs(path, type, fields, countFields, gqlFilter, accessibility);
 };
 
 export const askGuppyForSubAggregationData = (

diff --git a/src/server/__tests__/schema.test.js b/src/server/__tests__/schema.test.js
@@ -110,7 +110,7 @@ describe('Schema', () => {
 
   const expectedIndividualAggsSchemas = `
     type SubjectAggregation {
-      _totalCount: Int
+      _totalCount: Int,
       gen3_resource_path: HistogramForString,
       gender: HistogramForString,
       file_count: HistogramForNumber,
@@ -121,7 +121,7 @@ describe('Schema', () => {
       visits:NestedHistogramForVisits
     }
     type FileAggregation {
-      _totalCount: Int
+      _totalCount: Int,
       gen3_resource_path: HistogramForString,
       file_id: HistogramForString,
       file_size: HistogramForNumber,
@@ -147,12 +147,15 @@ describe('Schema', () => {
 
   const expectedHistogramSchemas = `
   type HistogramForString {
+    _cardinalityCount(precision_threshold:Int=3000): Int,
     histogram: [BucketsForNestedStringAgg]
   }
   type RegularAccessHistogramForString {
+    _cardinalityCount(precision_threshold:Int=3000): Int,
     histogram: [BucketsForNestedStringAgg]
   }
   type HistogramForNumber {
+    _cardinalityCount(precision_threshold:Int=3000): Int,
     histogram(
       rangeStart: Int,
       rangeEnd: Int,
@@ -162,6 +165,7 @@ describe('Schema', () => {
     asTextHistogram: [BucketsForNestedStringAgg]
   }
   type RegularAccessHistogramForNumber {
+    _cardinalityCount(precision_threshold:Int=3000): Int,
     histogram(
       rangeStart: Int,
       rangeEnd: Int,

diff --git a/src/server/es/filter.js b/src/server/es/filter.js
@@ -19,16 +19,16 @@ const fromPathToNode = (esInstance, esIndex, path) => {
 };
 
 const mergeRangeOperations = (a, b) => {
-  let merged = Object.assign({}, a, b);
+  const merged = { ...a, ...b };
 
-  Object.keys(merged).forEach(function(key) {
+  Object.keys(merged).forEach((key) => {
     if (typeof merged[key] === 'object' && merged[key] !== null) {
       merged[key] = mergeRangeOperations(a[key], b[key]);
     }
-  })
+  });
 
   return merged;
-}
+};
 
 const getNumericTextType = (
   esInstance,
@@ -252,7 +252,7 @@ const getFilterObj = (
         esInstance, esIndex, filterItem, aggsField, filterSelf, defaultAuthFilter, objPath,
       );
       if (filterObj) {
-        if ("range" in filterObj) {
+        if ('range' in filterObj) {
           filterRange.push(filterObj);
         } else {
           boolItemsList.push(filterObj);

diff --git a/src/server/es/index.js b/src/server/es/index.js
@@ -424,11 +424,32 @@ class ES {
   async getCount(esIndex, esType, filter) {
     const result = await this.filterData(
       { esInstance: this, esIndex, esType },
-      { filter, fields: false,  size: 0},
+      { filter, fields: false, size: 0 },
     );
     return result.hits.total;
   }
 
+  // eslint-disable-next-line camelcase
+  async getCardinalityCount(esIndex, esType, filter, field, precision_threshold) {
+    const queryBody = {
+      size: 0,
+      aggs: {
+        cardinality_count: {
+          cardinality: {
+            field,
+            precision_threshold,
+          },
+        },
+      },
+    };
+    if (typeof filter !== 'undefined') {
+      queryBody.query = getFilterObj(this, esIndex, filter);
+    }
+
+    const result = await this.query(esIndex, esType, queryBody);
+    return result.aggregations.cardinality_count.value;
+  }
+
   async getData({
     esIndex, esType, fields, filter, sort, offset, size,
   }) {

diff --git a/src/server/resolvers.js b/src/server/resolvers.js
@@ -123,6 +123,27 @@ const textHistogramResolver = async (parent, args, context) => {
   });
 };
 
+/**
+ * This resolver is for Cardinality.
+ * It inherits arguments from its parent,
+ * and uses "field" from parent and args "precision_threshold" to get the cardinality count
+ * @param {object} parent
+ * @param {object} args
+ */
+const cardinalityResolver = async (parent, args) => {
+  log.debug('[resolver.cardinalityResolver] args', args);
+  log.debug('[resolver.cardinalityResolver] parent', parent);
+  // TODO make work with nested
+  const {
+    esInstance, esIndex, esType, filter, field,
+  } = parent;
+
+  // eslint-disable-next-line camelcase
+  const { precision_threshold } = args;
+
+  return esInstance.getCardinalityCount(esIndex, esType, filter, field, precision_threshold);
+};
+
 const getFieldAggregationResolverMappingsByField = (field) => {
   let isNumericField = false;
   if (esFieldNumericTextTypeMapping[field.type] === NumericTextTypeTypeEnum.ES_NUMERIC_TYPE) {
@@ -137,6 +158,8 @@ const getFieldAggregationResolverMappingsByField = (field) => {
   }));
 };
 
+// this spreads all fields out into individual resolvers and
+// adds "field", "isNumericField" and "nestedPath", to parent
 const getFieldAggregationResolverMappings = (esInstance, esIndex) => {
   const { fields } = esInstance.getESFields(esIndex);
   const fieldAggregationResolverMappings = {};
@@ -165,6 +188,9 @@ const getFieldAggregationResolverMappings = (esInstance, esIndex) => {
  *         }
  *       }
  *       file_count {
+ *         _cardinality (
+ *           precision_threshold: 1000 //optional
+ *         ), ---> `cardinalityResolver`
  *         histogram (rangeStart: xx, rangeEnd: xx, rangeStep: xx, binCount: xx)
  *         {  ---> `numericHistogramResolver`
  *           key
@@ -259,17 +285,21 @@ const getResolver = (esConfig, esInstance) => {
     ...typeAggregationResolvers,
     ...typeNestedAggregationResolvers,
     HistogramForNumber: {
+      _cardinalityCount: cardinalityResolver,
       histogram: numericHistogramResolver,
       asTextHistogram: textHistogramResolver,
     },
     HistogramForString: {
+      _cardinalityCount: cardinalityResolver,
       histogram: textHistogramResolver,
     },
     RegularAccessHistogramForNumber: {
+      _cardinalityCount: cardinalityResolver,
       histogram: numericHistogramResolver,
       asTextHistogram: textHistogramResolver,
     },
     RegularAccessHistogramForString: {
+      _cardinalityCount: cardinalityResolver,
       histogram: textHistogramResolver,
     },
     Mapping: {