get numeric domain for display in chat

Author: Dev-Lan

package.json

@@ -16,6 +16,7 @@
   "dependencies": {
     "@quasar/extras": "^1.16.4",
     "@types/lodash-es": "^4.17.12",
+    "arquero": "^8.0.3",
     "lodash-es": "^4.17.21",
     "ollama": "^0.5.10",
     "openai": "^5.5.1",

src/components/IntervalFilterComponent.vue

@@ -2,6 +2,8 @@
 import { on } from 'events';
 import { range } from 'lodash-es';
 import { ref, computed, watch, onMounted, reactive } from 'vue';
+import { useDataPackageStore } from 'src/stores/dataPackageStore';
+const dataPackageStore = useDataPackageStore();
 
 interface IntervalFilterComponentProps {
   entity: string;
@@ -30,18 +32,59 @@ const rangeModel = ref<{ min: number; max: number }>({
   min: props.minInitial,
   max: props.maxInitial,
 });
+
+const rangeMinMax = computed<{ min: number; max: number }>(() => {
+  const domain = dataPackageStore.getDomainForField(props.entity, props.field);
+  if (domain && domain.type === 'interval') {
+    return {
+      min: domain.domain.min,
+      max: domain.domain.max,
+    };
+  }
+  return {
+    min: 0,
+    max: 100,
+  };
+});
+
+const minDisplayText = computed(() => {
+  if (rangeModel.value.min <= rangeMinMax.value.min) {
+    return 'min';
+  }
+  return `${rangeModel.value.min}`;
+});
+
+const maxDisplayText = computed(() => {
+  if (rangeModel.value.max >= rangeMinMax.value.max) {
+    return 'max';
+  }
+  return `${rangeModel.value.max}`;
+});
 </script>
 
 <template>
   <div class="q-mx-sm">
     Filtered <span class="emphasized q-mr-xs">{{ props.entity }}</span>
     <span class="emphasized">{{ props.field }}</span
-    >, <span class="emphasized">{{ rangeModel.min }}</span> to
-    <span class="emphasized">{{ rangeModel.max }}</span
+    >, <span class="emphasized">{{ minDisplayText }}</span> to
+    <span class="emphasized">{{ maxDisplayText }}</span
     >:
   </div>
-  <div class="q-mx-sm">
-    <q-range v-model="rangeModel" :min="0" :max="100" />
+  <div
+    v-if="
+      dataPackageStore.isValidIntervalFilter(
+        props.entity,
+        props.field,
+        rangeModel.min,
+        rangeModel.max,
+      ).isValid === 'yes'
+    "
+    class="q-mx-sm"
+  >
+    <q-range v-model="rangeModel" :min="rangeMinMax.min" :max="rangeMinMax.max" />
+  </div>
+  <div v-else class="q-mx-sm">
+    <span class="text-negative">Error: Invalid filter. </span>
   </div>
 </template>
 

src/stores/dataPackageStore.ts

@@ -1,5 +1,8 @@
 import { ref, computed } from 'vue';
 import { defineStore } from 'pinia';
+// import { parse, type ParseResult } from 'papaparse';
+import { loadCSV, agg, op, from, bin, rolling, escape, desc, type ColumnTable } from 'arquero';
+import { data } from 'autoprefixer';
 
 export const useDataPackageStore = defineStore('dataPackageStore', () => {
   const dataPackagePath = './data/hubmap_2025-05-05/datapackage_udi.json';
@@ -10,11 +13,102 @@ export const useDataPackageStore = defineStore('dataPackageStore', () => {
     .then((response) => response.json())
     .then((json) => {
       dataPackage.value = json;
+      initializeDataFieldDomains();
     })
     .catch((error) => {
       console.error('Error loading data package:', error);
     });
 
+  interface DataFieldDomain {
+    entity: string;
+    field: string;
+    domain: IntervalDomain | CategoricalDomain;
+  }
+
+  interface IntervalDomain {
+    type: 'interval';
+    min: number;
+    max: number;
+  }
+
+  interface CategoricalDomain {
+    values: string[];
+  }
+
+  const dataFieldDomains = ref<DataFieldDomain[]>([]);
+
+  function initializeDataFieldDomains() {
+    console.log('init thingy');
+    if (!dataPackage.value || !dataPackage.value.resources) return;
+    const folderPath = dataPackage.value['udi:path'];
+    for (const resource of dataPackage.value.resources) {
+      const entityName = resource.name;
+      const dataPath = resource.path;
+      const fullPath = `${folderPath}/${dataPath}`;
+      addDataDomains(fullPath, entityName);
+    }
+    return;
+  }
+
+  function getDomainForField(entity: string, field: string): DataFieldDomain | undefined {
+    return dataFieldDomains.value.find(
+      (domain) => domain.entity === entity && domain.field === field,
+    );
+  }
+
+  interface ValidStatus {
+    isValid: 'yes' | 'no' | 'unknown';
+  }
+
+  function isValidIntervalFilter(
+    entity: string,
+    field: string,
+    min: number,
+    max: number,
+  ): ValidStatus {
+    if (!dataPackage.value || !dataPackage.value.resources) {
+      return { isValid: 'unknown' };
+    }
+    const domain = getDomainForField(entity, field);
+    if (!domain) {
+      return { isValid: 'no' };
+    }
+    return { isValid: 'yes' };
+  }
+
+  async function addDataDomains(path: string, entity: string): Promise<void> {
+    const table = await loadCSV(path);
+
+    // Get column names
+    const cols = table.columnNames();
+
+    // Build domain info
+    const domains: DataFieldDomain[] = [];
+
+    for (const col of cols) {
+      const series = table.array(col); // raw column values
+      const isNumeric = series.every((v) => v == null || !isNaN(+v));
+
+      if (isNumeric) {
+        const stats = table
+          .rollup({
+            min: `(d) => op.min(d["${col}"])`,
+            max: `(d) => op.max(d["${col}"])`,
+          })
+          .objects()[0];
+        domains.push({
+          entity,
+          field: col,
+          type: 'interval',
+          domain: { min: stats.min, max: stats.max },
+        });
+      }
+      // TODO: handle categorical data
+    }
+    dataFieldDomains.value.push(...domains);
+    return;
+  }
+
   function removeVestigialInfo(data: object): object {
     // remove udi:overlapping_fields
     // for each resource.schema
@@ -65,5 +159,5 @@ export const useDataPackageStore = defineStore('dataPackageStore', () => {
     return fieldsMap;
   });
 
-  return { dataPackage, dataPackageString, sourceFields };
+  return { dataPackage, dataPackageString, sourceFields, isValidIntervalFilter, getDomainForField };
 });

yarn.lock

@@ -1245,7 +1245,7 @@ argparse@^2.0.1:
   resolved "https://registry.yarnpkg.com/argparse/-/argparse-2.0.1.tgz#246f50f3ca78a3240f6c997e8a9bd1eac49e4b38"
   integrity sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q==
 
-arquero@^8.0.1:
+arquero@^8.0.1, arquero@^8.0.3:
   version "8.0.3"
   resolved "https://registry.yarnpkg.com/arquero/-/arquero-8.0.3.tgz#ac72842df9143cbb3b9d215792344d3c6b14cef6"
   integrity sha512-7YQwe/GPVBUiahaPwEwgvu6VHyuhX0Ut61JZlIJYsAobOH5unLBwTmh43BobhX/N4dW1sb8WyKlQ8GjFq2whzQ==