Refactored and added method for parsing attachments

package.json

@@ -1,6 +1,6 @@
 {
   "name": "code-with-mosh-downloader",
-  "version": "0.0.2",
+  "version": "0.0.4",
   "description": "Download courses from codewithmosh.com",
   "main": ".webpack/main",
   "scripts": {
@@ -88,6 +88,7 @@
     "node-html-parser": "^4.1.0",
     "node-loader": "^2.0.0",
     "react-is": "^17.0.2",
+    "sanitize-filename": "^1.6.3",
     "spectron": "^15.0.0",
     "style-loader": "^3.2.1",
     "styled-components": "^5.3.0",

src/main/pageParser.ts

@@ -1,7 +1,7 @@
 import {DownloadQueueItemType} from 'Types/types';
 import {getString} from 'Main/client';
 import {HTMLElement} from 'node-html-parser';
-import {ITypeParser, ParsedItem} from 'MainTypes/types';
+import {ITypeParser, ParsedAttachment, ParsedItem} from 'MainTypes/types';
 
 export async function parsePageFromUrl(url: string, type: DownloadQueueItemType) {
   if (!type) {
@@ -94,6 +94,38 @@ export class CourseParser implements ITypeParser {
 }
 
 export class VideoParser implements ITypeParser {
+  private static getVideo(html: HTMLElement) {
+	const wistiaIdElement = html.querySelector('.attachment-wistia-player');
+
+	return wistiaIdElement ? wistiaIdElement.getAttribute('data-wistia-id') : null;
+  }
+
+  private static getAttachments(html: HTMLElement) {
+    const attachmentElements = html.querySelectorAll('.lecture-attachment:not(.lecture-attachment-type-video)');
+
+    const attachments: ParsedAttachment[] = [];
+
+    attachmentElements.forEach(x => {
+      if (x.hasAttribute('lecture-attachment-type-text')) {
+        const textContainer = x.querySelector('.lecture-text-container');
+
+        attachments.push({
+		  type: 'text',
+		  data: textContainer.innerHTML
+		});
+	  } else if (x.hasAttribute('.lecture-attachment-type-file')) {
+        const downloadLink = x.querySelector('a');
+
+        attachments.push({
+		  type: 'download',
+		  data: downloadLink.getAttribute('href')
+		});
+	  }
+	});
+
+    return attachments;
+  }
+
   async parse(html: HTMLElement) {
 	const lectureId = html.querySelector('#lecture_heading').getAttribute('data-lecture-id');
 	const videoTitle = html.querySelector('#lecture_heading').textContent.trim().replace(/(\d+)(-)(.+)/gmi, '$1 $2$3');
@@ -103,7 +135,8 @@ export class VideoParser implements ITypeParser {
 	const initialCourseSectionHeading = courseSection.querySelector('.section-title').textContent.trim();
 	const courseSectionHeading = /(.+)\s\(\d+m\)/gmi.exec(initialCourseSectionHeading)[1];
 
-	const wistiaId = html.querySelector('.attachment-wistia-player').getAttribute('data-wistia-id');
+	const wistiaId = VideoParser.getVideo(html);
+	const attachments = VideoParser.getAttachments(html);
 
 	const courseTitle = html.querySelector('.course-sidebar-head > h2').textContent;
 
@@ -113,7 +146,8 @@ export class VideoParser implements ITypeParser {
 		extraData: {
 		  courseTitle,
 		  courseSectionHeading,
-		  videoTitle
+		  videoTitle,
+		  attachments
 		}
 	  }
 	] as ParsedItem[];

src/main/types/types.d.ts

@@ -5,6 +5,11 @@ export interface ITypeParser {
   parse(html: HTMLElement): Promise<ParsedItem[]>
 }
 
+export type ParsedAttachment = {
+  type: 'text' | 'download',
+  data: unknown
+}
+
 export type ParsedItem = {
   nextUrl: string,
   nextType?: DownloadQueueItemType,

yarn.lock

@@ -6135,6 +6135,13 @@ safe-buffer@>=5.1.0, safe-buffer@^5.0.1, safe-buffer@^5.1.0, safe-buffer@^5.1.2,
   resolved "https://registry.yarnpkg.com/safer-buffer/-/safer-buffer-2.1.2.tgz#44fa161b0187b9549dd84bb91802f9bd8385cd6a"
   integrity sha512-YZo3K82SD7Riyi0E1EQPojLz7kpepnSQI9IyPbHHg1XXXevb5dJI7tpyN2ADxGcQbHG7vcyRHk0cbwqcQriUtg==
 
+sanitize-filename@^1.6.3:
+  version "1.6.3"
+  resolved "https://registry.yarnpkg.com/sanitize-filename/-/sanitize-filename-1.6.3.tgz#755ebd752045931977e30b2025d340d7c9090378"
+  integrity sha512-y/52Mcy7aw3gRm7IrcGDFx/bCk4AhRh2eI9luHOQM86nZsqwiRkkq2GekHXBBD+SmPidc8i2PqtYZl+pWJ8Oeg==
+  dependencies:
+    truncate-utf8-bytes "^1.0.0"
+
 sax@^1.2.4:
   version "1.2.4"
   resolved "https://registry.yarnpkg.com/sax/-/sax-1.2.4.tgz#2816234e2378bddc4e5354fab5caa895df7100d9"
@@ -6886,6 +6893,13 @@ trim-repeated@^1.0.0:
   dependencies:
     escape-string-regexp "^1.0.2"
 
+truncate-utf8-bytes@^1.0.0:
+  version "1.0.2"
+  resolved "https://registry.yarnpkg.com/truncate-utf8-bytes/-/truncate-utf8-bytes-1.0.2.tgz#405923909592d56f78a5818434b0b78489ca5f2b"
+  integrity sha1-QFkjkJWS1W94pYGENLC3hInKXys=
+  dependencies:
+    utf8-byte-length "^1.0.1"
+
 ts-loader@^9.2.2:
   version "9.2.3"
   resolved "https://registry.yarnpkg.com/ts-loader/-/ts-loader-9.2.3.tgz#dc3b6362a4d4382493cd4f138d345f419656de68"
@@ -7078,6 +7092,11 @@ username@^5.1.0:
     execa "^1.0.0"
     mem "^4.3.0"
 
+utf8-byte-length@^1.0.1:
+  version "1.0.4"
+  resolved "https://registry.yarnpkg.com/utf8-byte-length/-/utf8-byte-length-1.0.4.tgz#f45f150c4c66eee968186505ab93fcbb8ad6bf61"
+  integrity sha1-9F8VDExm7uloGGUFq5P8u4rWv2E=
+
 util-deprecate@^1.0.1, util-deprecate@~1.0.1:
   version "1.0.2"
   resolved "https://registry.yarnpkg.com/util-deprecate/-/util-deprecate-1.0.2.tgz#450d4dc9fa70de732762fbd2d4a28981419a0ccf"