fix(api): update logic to check schedule week date

sws2apps · Dec 19, 2023 · 2ad9482 · 2ad9482
1 parent 0175463
commit 2ad9482
Show file tree

Hide file tree

Showing 3 changed files with 22 additions and 15 deletions.
diff --git a/src/common/html_validation.ts b/src/common/html_validation.ts
@@ -4,7 +4,7 @@ import { HTMLElement, parse } from 'node-html-parser';
 export const isValidHTML = (name: string): boolean => {
 	let valid = false;
 
-	if (name.startsWith('OEBPS') && name.endsWith('.xhtml')) {
+	if (name.startsWith('OEBPS') && name.endsWith('.xhtml') && name.indexOf('-extracted') === -1) {
 		const fileName = name.split('/')[1].split('.')[0];
 		if (!isNaN(parseFloat(fileName))) {
 			valid = true;

diff --git a/src/common/parser.ts b/src/common/parser.ts
@@ -260,12 +260,14 @@ export const parseWSchedule = (article: HTMLElement, content: HTMLElement, wLang
 
 	const studyDate = getWStudyDate(article);
 
-	if (isEnhancedParsing) {
-		const wStudyEnhanced = getWTStudyDateEnhanced(studyDate, wLang);
-		weekItem.w_study_date = wStudyEnhanced;
-		weekItem.w_study_date_locale = studyDate;
-	} else {
-		weekItem.w_study_date = studyDate;
+	if (studyDate.length > 0) {
+		if (isEnhancedParsing) {
+			const wStudyEnhanced = getWTStudyDateEnhanced(studyDate, wLang);
+			weekItem.w_study_date = wStudyEnhanced;
+			weekItem.w_study_date_locale = studyDate;
+		} else {
+			weekItem.w_study_date = studyDate;
+		}
 	}
 
 	const studyTitle = getWStudyTitle(article);

diff --git a/src/common/parsing_rules.ts b/src/common/parsing_rules.ts
@@ -5,17 +5,22 @@ export const extractMonthName = (src: string, lang: string) => {
 	let varDay;
 	let monthIndex;
 
+	const text = src.toLowerCase();
+	const split = text.split(/[–-—]/);
 	const monthNames = getMonthNames(lang);
 
-	for (const month of monthNames) {
-		const monthLang = month.name;
-		const regex = new RegExp(`(${monthLang.toLowerCase()})`);
-		const array = regex.exec(src.toLowerCase());
+	outerLoop: for (const splitted of split) {
+		for (const month of monthNames) {
+			const monthLang = month.name.toLowerCase();
+			const regex = new RegExp(`(${monthLang})`);
 
-		if (Array.isArray(array)) {
-			varDay = +src.match(/(\d+)/)![0];
-			monthIndex = month.index;
-			break;
+			const array = regex.exec(splitted);
+
+			if (Array.isArray(array)) {
+				varDay = +text.match(/(\d+)/)![0];
+				monthIndex = month.index;
+				break outerLoop;
+			}
 		}
 	}