Remove sort and plot type options from categorical attributes

packages/core/src/aggregate.ts

@@ -4,6 +4,7 @@ import {
   Aggregate,
   AggregateBy,
   Aggregates,
+  ColumnTypes,
   Items,
   Row,
   Rows,
@@ -48,12 +49,23 @@ export function getItems(row: Row): string[] {
  * @param aggs - The aggregates object containing the order and values of the aggregates.
  * @param items - The items object containing the data items.
  * @param attributeColumns - The array of attribute columns to calculate the attributes from.
+ * @param columnTypes - The column type definitions.
  */
-function updateAggValues(aggs: Aggregates, items: Items, attributeColumns: string[]) {
+function updateAggValues(
+  aggs: Aggregates,
+  items: Items,
+  attributeColumns: string[],
+  columnTypes: ColumnTypes,
+) {
   aggs.order.forEach((aggId) => {
     aggs.values[aggId].atts = {
       dataset: {
-        ...getSixNumberSummary(items, getItems(aggs.values[aggId]), attributeColumns),
+        ...getSixNumberSummary(
+          items,
+          getItems(aggs.values[aggId]),
+          attributeColumns,
+          columnTypes,
+        ),
       },
       derived: { deviation: aggs.values[aggId].atts.derived.deviation },
     };
@@ -87,6 +99,7 @@ export const getAggSize = (row: Row) => {
  * @param level - The level of the aggregation.
  * @param items - The items to be aggregated.
  * @param attributeColumns - The attribute columns to be considered in the aggregation.
+ * @param columnTypes - The column type definitions.
  * @param parentPrefix - The parent prefix for the aggregated subsets.
  * @returns The aggregated subsets.
  */
@@ -95,6 +108,7 @@ function aggregateByDegree(
   level: number,
   items: Items,
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
   parentPrefix: string,
 ) {
   if (subsets.order.length === 0) return subsets;
@@ -152,7 +166,7 @@ function aggregateByDegree(
     relevantAggregate.atts.derived.deviation += subset.atts.derived.deviation;
   });
 
-  updateAggValues(aggs, items, attributeColumns);
+  updateAggValues(aggs, items, attributeColumns, columnTypes);
 
   return aggs;
 }
@@ -165,6 +179,7 @@ function aggregateByDegree(
  * @param level - The level of aggregation.
  * @param items - The items to be aggregated.
  * @param attributeColumns - The attribute columns used for aggregation.
+ * @param columnTypes - The column type definitions.
  * @param parentPrefix - The parent prefix used for aggregation.
  * @returns The aggregated subsets.
  */
@@ -174,6 +189,7 @@ function aggregateBySets(
   level: number,
   items: Items,
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
   parentPrefix: string,
 ) {
   if (subsets.order.length === 0) return subsets;
@@ -259,7 +275,7 @@ function aggregateBySets(
     });
   });
 
-  updateAggValues(aggs, items, attributeColumns);
+  updateAggValues(aggs, items, attributeColumns, columnTypes);
 
   return aggs;
 }
@@ -271,6 +287,7 @@ function aggregateBySets(
  * @param level - The level of the aggregation.
  * @param items - The items to be aggregated.
  * @param attributeColumns - The attribute columns to be considered.
+ * @param columnTypes - The column type definitions.
  * @param parentPrefix - The parent prefix for the aggregation.
  * @returns The aggregated subsets.
  */
@@ -279,6 +296,7 @@ function aggregateByDeviation(
   level: number,
   items: Items,
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
   parentPrefix: string,
 ) {
   if (subsets.order.length === 0) return subsets;
@@ -347,7 +365,7 @@ function aggregateByDeviation(
     relevantAggregate.atts.derived.deviation += subset.atts.derived.deviation;
   });
 
-  updateAggValues(aggs, items, attributeColumns);
+  updateAggValues(aggs, items, attributeColumns, columnTypes);
 
   return aggs;
 }
@@ -361,6 +379,7 @@ function aggregateByDeviation(
  * @param level - The level of the aggregation.
  * @param items - The items associated with the subsets.
  * @param attributeColumns - The attribute columns to be considered for aggregation.
+ * @param columnTypes - The column type definitions.
  * @param parentPrefix - The prefix for the parent identifier.
  * @returns The aggregated subsets.
  */
@@ -371,6 +390,7 @@ function aggregateByOverlaps(
   level: number,
   items: Items,
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
   parentPrefix: string,
 ) {
   if (subsets.order.length === 0) return subsets;
@@ -451,7 +471,7 @@ function aggregateByOverlaps(
     });
   });
 
-  updateAggValues(aggs, items, attributeColumns);
+  updateAggValues(aggs, items, attributeColumns, columnTypes);
 
   return aggs;
 }
@@ -465,6 +485,7 @@ function aggregateByOverlaps(
  * @param sets - The sets associated with the subsets.
  * @param items - The items associated with the subsets.
  * @param attributeColumns - The attribute columns to consider when aggregating by deviations.
+ * @param columnTypes - The column type definitions.
  * @param level - The level of aggregation (default: 1).
  * @param parentPrefix - The parent prefix for the aggregated subsets (default: '').
  * @returns The aggregated subsets.
@@ -476,15 +497,38 @@ function aggregateSubsets(
   sets: Sets,
   items: Items,
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
   level: number = 1,
   parentPrefix: string = '',
 ) {
   if (aggregateBy === 'Degree')
-    return aggregateByDegree(subsets, level, items, attributeColumns, parentPrefix);
+    return aggregateByDegree(
+      subsets,
+      level,
+      items,
+      attributeColumns,
+      columnTypes,
+      parentPrefix,
+    );
   if (aggregateBy === 'Sets')
-    return aggregateBySets(subsets, sets, level, items, attributeColumns, parentPrefix);
+    return aggregateBySets(
+      subsets,
+      sets,
+      level,
+      items,
+      attributeColumns,
+      columnTypes,
+      parentPrefix,
+    );
   if (aggregateBy === 'Deviations')
-    return aggregateByDeviation(subsets, level, items, attributeColumns, parentPrefix);
+    return aggregateByDeviation(
+      subsets,
+      level,
+      items,
+      attributeColumns,
+      columnTypes,
+      parentPrefix,
+    );
   if (aggregateBy === 'Overlaps') {
     return aggregateByOverlaps(
       subsets,
@@ -493,6 +537,7 @@ function aggregateSubsets(
       level,
       items,
       attributeColumns,
+      columnTypes,
       parentPrefix,
     );
   }
@@ -508,6 +553,7 @@ function aggregateSubsets(
  * @param sets - The sets associated with the subsets.
  * @param items - The items associated with the subsets.
  * @param attributeColumns - The attribute columns to be considered during aggregation.
+ * @param columnTypes - The column type definitions.
  * @returns The aggregated rows.
  */
 export function firstAggregation(
@@ -517,6 +563,7 @@ export function firstAggregation(
   sets: Sets,
   items: Items,
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
 ): Rows {
   return aggregateSubsets(
     subsets,
@@ -525,6 +572,7 @@ export function firstAggregation(
     sets,
     items,
     attributeColumns,
+    columnTypes,
   );
 }
 
@@ -537,6 +585,7 @@ export function firstAggregation(
  * @param sets - The sets data.
  * @param items - The items data.
  * @param attributeColumns - The attribute columns to consider.
+ * @param columnTypes - The column type definitions.
  * @returns The aggregated result.
  */
 export function secondAggregation(
@@ -546,6 +595,7 @@ export function secondAggregation(
   sets: Sets,
   items: Items,
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
 ) {
   const aggs: Aggregates = {
     values: {},
@@ -562,6 +612,7 @@ export function secondAggregation(
         sets,
         items,
         attributeColumns,
+        columnTypes,
         2,
         `${agg.id}-`,
       );

packages/core/src/process.ts

@@ -162,21 +162,29 @@ function processRawData(data: TableRow[], columns: ColumnTypes) {
 }
 
 /**
- * Calculates the six-number summary for each attribute in the given items.
+ * Calculates the six-number summary for each numerical attribute in the given items.
+ * Categorical attributes are automatically filtered out.
  *
  * @param items - The items to calculate the summary for.
  * @param memberItems - The member items to consider.
  * @param attributeColumns - The attribute columns to calculate the summary for.
- * @returns An object containing the six-number summary for each attribute.
+ * @param columnTypes - The column type definitions to filter numerical attributes.
+ * @returns An object containing the six-number summary for each numerical attribute.
  */
 export function getSixNumberSummary(
   items: Items,
   memberItems: string[],
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
 ): AttributeList {
   const attributes: AttributeList = {};
 
-  attributeColumns.forEach((attribute) => {
+  // Only process numerical attributes; categorical attributes don't have meaningful numerical statistics
+  const numericalAttributes = attributeColumns.filter(
+    (attr) => columnTypes[attr] === 'number',
+  );
+
+  numericalAttributes.forEach((attribute) => {
     const values = memberItems
       .map((d) => items[d].atts[attribute] as number)
       .filter((val) => !Number.isNaN(val));
@@ -201,13 +209,15 @@ export function getSixNumberSummary(
  * @param setColumns - The array of column names representing the set columns.
  * @param items - The items object containing the data items.
  * @param attributeColumns - The array of column names representing the attribute columns.
+ * @param columnTypes - The column type definitions.
  * @returns The sets object containing the retrieved sets.
  */
 function getSets(
   setMembership: { [col: string]: string[] },
   setColumns: ColumnName[],
   items: Items,
   attributeColumns: ColumnName[],
+  columnTypes: ColumnTypes,
 ) {
   const setMembershipStatus: { [col: string]: SetMembershipStatus } = {};
 
@@ -225,7 +235,12 @@ function getSets(
       size: setMembership[col].length,
       setMembership: { ...setMembershipStatus, [col]: 'Yes' },
       atts: {
-        dataset: getSixNumberSummary(items, setMembership[col], attributeColumns),
+        dataset: getSixNumberSummary(
+          items,
+          setMembership[col],
+          attributeColumns,
+          columnTypes,
+        ),
         derived: { deviation: 0 },
       },
     };
@@ -246,7 +261,7 @@ function getSets(
 export function process(data: TableRow[], columns: ColumnTypes): CoreUpsetData {
   const { items, setMembership, labelColumn, setColumns, attributeColumns } =
     processRawData(data, columns);
-  const sets = getSets(setMembership, setColumns, items, attributeColumns);
+  const sets = getSets(setMembership, setColumns, items, attributeColumns, columns);
 
   return {
     label: labelColumn,
@@ -265,13 +280,15 @@ export function process(data: TableRow[], columns: ColumnTypes): CoreUpsetData {
  * @param sets - The sets used to calculate subsets.
  * @param vSets - The vSets used to calculate subsets.
  * @param attributeColumns - The attribute columns used to calculate subsets.
+ * @param columnTypes - The column type definitions.
  * @returns The calculated subsets.
  */
 export function getSubsets(
   dataItems: { [k: string]: Item },
   sets: Sets,
   vSets: string[],
   attributeColumns: string[],
+  columnTypes: ColumnTypes,
 ): Subsets {
   if (vSets.length === 0) {
     return {
@@ -349,7 +366,7 @@ export function getSubsets(
       setMembership: setMembershipStatus,
       atts: {
         derived: { deviation: subsetDeviation },
-        dataset: getSixNumberSummary(dataItems, itm, attributeColumns),
+        dataset: getSixNumberSummary(dataItems, itm, attributeColumns, columnTypes),
       },
     };
 

packages/core/src/render.ts

@@ -67,6 +67,7 @@ const firstAggRR = (data: CoreUpsetData, state: UpsetConfig) => {
     data.sets,
     state.visibleSets,
     data.attributeColumns,
+    data.columnTypes,
   );
   return firstAggregation(
     subsets,
@@ -75,6 +76,7 @@ const firstAggRR = (data: CoreUpsetData, state: UpsetConfig) => {
     data.sets,
     data.items,
     data.attributeColumns,
+    data.columnTypes,
   );
 };
 
@@ -98,6 +100,7 @@ const secondAggRR = (data: CoreUpsetData, state: UpsetConfig) => {
       data.sets,
       data.items,
       data.attributeColumns,
+      data.columnTypes,
     );
 
     return secondAgg;
@@ -165,6 +168,7 @@ const sortByRR = (data: CoreUpsetData, state: UpsetConfig, ignoreQuery = false)
       data.sets,
       state.visibleSets,
       data.attributeColumns,
+      data.columnTypes,
     );
     renderRows = getQueryResult(subsets, state.setQuery.query);
   } else {

packages/core/src/sort.ts

@@ -197,11 +197,13 @@ function sortByAttribute(rows: Intersections, sortBy: string, sortByOrder?: Sort
       meanA = values[a].atts.derived.deviation;
       meanB = values[b].atts.derived.deviation;
     } else {
-      meanA = values[a].atts.dataset[sortBy].mean;
-      meanB = values[b].atts.dataset[sortBy].mean;
+      // For categorical attributes, dataset[sortBy] will be undefined since
+      // getSixNumberSummary filters them out; they will be sorted to the bottom
+      meanA = values[a].atts.dataset[sortBy]?.mean;
+      meanB = values[b].atts.dataset[sortBy]?.mean;
     }
 
-    // If one of the values is undefined (empty subset), sort it to the bottom
+    // If one of the values is undefined (empty subset or categorical attribute), sort it to the bottom
     if (!meanA) {
       return 1;
     }