Merge branch 'master' into storyq-2

src/stores/feature_store.ts

@@ -70,8 +70,7 @@ export class FeatureStore {
 		if (iFeature.info.kind === 'search') {
 			const tDetails = iFeature.info.details as SearchDetails,
 				tFirstPart = namingAbbreviations[tDetails.where],
-				tMatch = tDetails.freeFormText.match( /\w+/),
-				tSecondPart = tDetails.freeFormText !== '' ? `"${tMatch ? tMatch[0] : ''}"` :
+				tSecondPart = tDetails.freeFormText !== '' ? `"${tDetails.freeFormText.trim()}"` :
 					tDetails.punctuation !== '' ? tDetails.punctuation :
 					tDetails.wordList && tDetails.wordList.datasetName !== '' ? tDetails.wordList.datasetName :
 					tDetails.what === 'any number' ? 'anyNumber' : ''

src/stores/target_store.ts

@@ -208,7 +208,30 @@ export class TargetStore {
 			const option = (iNewFeature.info.details as SearchDetails).where;
 			const tBegins = option === featureDescriptors.containsOptions[2] ? '^' : '';
 			const tEnds = option === featureDescriptors.containsOptions[3] ? '$' : '';
-			const tParamString = `${tTargetAttr},"${tBegins}\\\\\\\\b${(iNewFeature.info.details as SearchDetails).freeFormText}\\\\\\\\b${tEnds}"`;
+			const text = (iNewFeature.info.details as SearchDetails).freeFormText.trim();
+			// note: the multiple slash escaping is due to all the layers between this code and the CODAP formula evaluator
+			const escapedText = text
+				.replace(/[.*+?^${}()|[\]\\]/g, '\\\\\\\\$&') // escape regex modifiers
+				.replace(/\s+/g, '\\\\\\\\s+') // allow multiple spaces between words
+				.replace(/['"“”‘’]/g, (match) => { // allow both regular and smart quotes to match each other
+					switch (match) {
+						case '"':
+						case '“':
+						case '”':
+							return `["“”]`;
+						case "'":
+						case '‘':
+						case '’':
+							return `['‘’]`;
+						default:
+							return match;
+					}
+				});
+			// don't add word boundaries when the user input starts/ends with non-word characters, like ! or , as that would fail matching
+			const wordBoundary = `\\\\\\\\b`;
+			const maybeStartingWordBoundary = /^\w/.test(text) ? wordBoundary : '';
+			const maybeEndingWordBoundary = /\w$/.test(text) ? wordBoundary : '';
+			const tParamString = `${tTargetAttr},"${tBegins}${maybeStartingWordBoundary}${escapedText}${maybeEndingWordBoundary}${tEnds}"`;
 			let tResult = '';
 			switch (option) {//['contain', 'not contain', 'start with', 'end with']
 				case featureDescriptors.containsOptions[0]:	// contain

src/utilities/utilities.ts

@@ -1,4 +1,4 @@
-import {wordTokenizer} from "../lib/one_hot";
+// import {wordTokenizer} from "../lib/one_hot";
 
 /**
  *
@@ -7,7 +7,7 @@ import {wordTokenizer} from "../lib/one_hot";
  * @param iSpecialFeatures - an array of features that don't appear in the text but may appear in iSelectedWords
  */
 export function textToObject( iText:string, iSelectedWords:any, iSpecialFeatures:string[]):any {
-	let segment = '';
+	// let segment = '';
 	let tResultArray:any = [];
 	iSpecialFeatures.forEach( iFeature => {
 		if( iSelectedWords.indexOf( iFeature) >= 0)
@@ -16,6 +16,12 @@ export function textToObject( iText:string, iSelectedWords:any, iSpecialFeatures
 			});
 	});
 
+	// do not highlight text for now
+	tResultArray.push({text: iText});
+
+	/*
+	NOTE: this code is broken and doesn't match phrases or match lists like personal pronoun lists
+
 	let words:string[] = wordTokenizer(iText, false, false);
 	words.forEach((iWord) => {
 		let tRawWord = iWord.toLowerCase();
@@ -39,6 +45,8 @@ export function textToObject( iText:string, iSelectedWords:any, iSpecialFeatures
 	});
 	if( segment !== '')
 		tResultArray.push({ text: segment });
+	*/
+
 	return tResultArray;
 }