fix import, re #10780

archesproject · Aug 7, 2024 · 3d7545c · 3d7545c
1 parent 8ac98b2
commit 3d7545c
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 9 deletions.
diff --git a/arches/app/media/js/views/components/etl_modules/import-single-csv.js b/arches/app/media/js/views/components/etl_modules/import-single-csv.js
@@ -12,7 +12,7 @@ define([
     'bindings/datatable',
     'bindings/dropzone',
     'bindings/resizable-sidepanel',
-], function(ko, koMapping, $, dropzone, fuzzyset, uuid, arches, JsonErrorAlertViewModel, importSingleCSVTemplate) {
+], function(ko, koMapping, $, dropzone, FuzzySet, uuid, arches, JsonErrorAlertViewModel, importSingleCSVTemplate) {
     const viewModel = function(params) {
         const self = this;
         this.loadDetails = params.load_details || ko.observable();
@@ -63,19 +63,26 @@ define([
             let highestScore = 0;
             if (!!self.headers() && self.headers()[i] != 'resourceid') {
                 const header = normalizeText(self.headers()[i]);
+                const fuzzySet = FuzzySet();
                 self.nodes().forEach(function(node) {
                     if (node.name) {
                         const nameNorm = normalizeText(node.name);
                         const aliasNorm = normalizeText(node.alias);
-                        const nameScore = stringSimilarity.compareTwoStrings(header, nameNorm);
-                        const aliasScore = stringSimilarity.compareTwoStrings(header, aliasNorm);
-                        const bestNodeScore = Math.max(nameScore, aliasScore);
-                        if (bestNodeScore > highestScore) {
-                            highestScore = bestNodeScore;
-                            bestMatch = node;
-                        }
+                        fuzzySet.add(nameNorm);
+                        fuzzySet.add(aliasNorm);
                     }
                 });
+                const results = fuzzySet.get(header);
+                if (results) {
+                    const [bestScore, bestMatchName] = results[0]; // results are sorted by score
+                    if (bestScore > highestScore) {
+                        highestScore = bestScore;
+                        bestMatch = self.nodes().find(node =>
+                            normalizeText(node.name) === bestMatchName ||
+                            normalizeText(node.alias) === bestMatchName
+                        );
+                    }
+                }
                 if (bestMatch && highestScore > 0.8) { return bestMatch.alias; }
             }
             return null;

diff --git a/package.json b/package.json
@@ -81,7 +81,7 @@
         "requirejs-plugins": "1.0.2",
         "requirejs-text": "~2.0.16",
         "select-woo": "jadu/selectWoo",
-        "string-similarity": "4.0.4",
+        "fuzzyset": "1.0.7",
         "underscore": "^1.13.6",
         "uuidjs": "^3.3.0",
         "vue": "^3.4.21",
@@ -110,6 +110,7 @@
         "jquery-validate": "node_modules/jquery-validation/dist/jquery.validate.min",
         "d3": "node_modules/d3/src/index.js",
         "dropzone": "node_modules/dropzone/dist/min/dropzone-amd-module.min",
+        "fuzzyset": "node_modules/fuzzyset/dist/fuzzyset.esm",
         "ckeditor4": "node_modules/ckeditor4/ckeditor",
         "ckeditor-jquery": "node_modules/ckeditor4/adapters/jquery",
         "knockout": "node_modules/knockout/build/output/knockout-latest",