Desktop: Fixes #3153: Make GotoAnyting work with East Asian charactors (

#3180)

CliClient/tests/services_SearchEngine.js

@@ -257,22 +257,29 @@ describe('services_SearchEngine', function() {
 
 	it('should support queries with Chinese characters', asyncTest(async () => {
 		let rows;
-		const n1 = await Note.save({ title: '我是法国人' });
+		const n1 = await Note.save({ title: '我是法国人', body: '中文测试' });
 
 		await engine.syncTables();
 
 		expect((await engine.search('我')).length).toBe(1);
 		expect((await engine.search('法国人')).length).toBe(1);
+		expect((await engine.search('法国人*'))[0].fields.sort()).toEqual(['body', 'title']); // usually assume that keyword was matched in body
+		expect((await engine.search('测试')).length).toBe(1);
+		expect((await engine.search('测试'))[0].fields).toEqual(['body']);
+		expect((await engine.search('测试*'))[0].fields).toEqual(['body']);
 	}));
 
 	it('should support queries with Japanese characters', asyncTest(async () => {
 		let rows;
-		const n1 = await Note.save({ title: '私は日本語を話すことができません' });
+		const n1 = await Note.save({ title: '私は日本語を話すことができません', body: 'テスト' });
 
 		await engine.syncTables();
 
 		expect((await engine.search('日本')).length).toBe(1);
 		expect((await engine.search('できません')).length).toBe(1);
+		expect((await engine.search('できません*'))[0].fields.sort()).toEqual(['body', 'title']); // usually assume that keyword was matched in body
+		expect((await engine.search('テスト'))[0].fields.sort()).toEqual(['body']);
+
 	}));
 
 	it('should support queries with Korean characters', asyncTest(async () => {
@@ -302,10 +309,15 @@ describe('services_SearchEngine', function() {
 		await engine.syncTables();
 
 		expect((await engine.search('title:你好*')).length).toBe(1);
+		expect((await engine.search('title:你好*'))[0].fields).toEqual(['title']);
+		expect((await engine.search('body:法国人')).length).toBe(1);
+		expect((await engine.search('body:法国人'))[0].fields).toEqual(['body']);
 		expect((await engine.search('body:你好')).length).toBe(0);
 		expect((await engine.search('title:你好 body:法国人')).length).toBe(1);
+		expect((await engine.search('title:你好 body:法国人'))[0].fields.sort()).toEqual(['body', 'title']);
 		expect((await engine.search('title:你好 body:bla')).length).toBe(0);
 		expect((await engine.search('title:你好 我是')).length).toBe(1);
+		expect((await engine.search('title:你好 我是'))[0].fields.sort()).toEqual(['body', 'title']);
 		expect((await engine.search('title:bla 我是')).length).toBe(0);
 
 		// For non-alpha char, only the first field is looked at, the following ones are ignored

ReactNativeClient/lib/services/SearchEngine.js

@@ -235,14 +235,37 @@ class SearchEngine {
 		return occurenceCount / spread;
 	}
 
-	processResults_(rows, parsedQuery) {
+	processBasicSearchResults_(rows, parsedQuery) {
+		const valueRegexs = parsedQuery.keys.includes('_') ? parsedQuery.terms['_'].map(term => term.valueRegex || term.value) : [];
+		const isTitleSearch = parsedQuery.keys.includes('title');
+		const isOnlyTitle = parsedQuery.keys.length === 1 && isTitleSearch;
+
 		for (let i = 0; i < rows.length; i++) {
 			const row = rows[i];
-			const offsets = row.offsets.split(' ').map(o => Number(o));
-			row.weight = this.calculateWeight_(offsets, parsedQuery.termCount);
-			row.fields = this.fieldNamesFromOffsets_(offsets);
+			const testTitle = regex => new RegExp(regex, 'ig').test(row.title);
+			const matchedFields = {
+				title: isTitleSearch || valueRegexs.some(testTitle),
+				body: !isOnlyTitle,
+			};
+
+			row.fields = Object.keys(matchedFields).filter(key => matchedFields[key]);
+			row.weight = 0;
+		}
+	}
+
+	processResults_(rows, parsedQuery, isBasicSearchResults = false) {
+		if (isBasicSearchResults) {
+			this.processBasicSearchResults_(rows, parsedQuery);
+		} else {
+			for (let i = 0; i < rows.length; i++) {
+				const row = rows[i];
+				const offsets = row.offsets.split(' ').map(o => Number(o));
+				row.weight = this.calculateWeight_(offsets, parsedQuery.termCount);
+				row.fields = this.fieldNamesFromOffsets_(offsets);
+			}
 		}
 
+
 		rows.sort((a, b) => {
 			if (a.fields.includes('title') && !b.fields.includes('title')) return -1;
 			if (!a.fields.includes('title') && b.fields.includes('title')) return +1;
@@ -383,6 +406,8 @@ class SearchEngine {
 		const searchOptions = {};
 
 		for (const key of parsedQuery.keys) {
+			if (parsedQuery.terms[key].length === 0) continue;
+
 			const term = parsedQuery.terms[key][0].value;
 			if (key === '_') searchOptions.anywherePattern = `*${term}*`;
 			if (key === 'title') searchOptions.titlePattern = `*${term}*`;
@@ -415,18 +440,20 @@ class SearchEngine {
 		query = this.normalizeText_(query);
 
 		const searchType = this.determineSearchType_(query, options.searchType);
+		const parsedQuery = this.parseQuery(query);
 
 		if (searchType === SearchEngine.SEARCH_TYPE_BASIC) {
 			// Non-alphabetical languages aren't support by SQLite FTS (except with extensions which are not available in all platforms)
-			return this.basicSearch(query);
+			const rows = await this.basicSearch(query);
+			this.processResults_(rows, parsedQuery, true);
+			return rows;
 		} else { // SEARCH_TYPE_FTS
 			// FTS will ignore all special characters, like "-" in the index. So if
 			// we search for "this-phrase" it won't find it because it will only
 			// see "this phrase" in the index. Because of this, we remove the dashes
 			// when searching.
 			// https://github.com/laurent22/joplin/issues/1075#issuecomment-459258856
 			query = query.replace(/-/g, ' ');
-			const parsedQuery = this.parseQuery(query);
 			const sql = `
 				SELECT
 					notes_fts.id,