xevion/schedule-extract
1
0
Fork 0
mirror of https://github.com/Xevion/schedule-extract.git synced 2025-12-06 11:16:19 -06:00
Code Issues Packages Projects Releases Wiki Activity

Work on parsing/traversing HTML tree

Browse Source
This commit is contained in:
Xevion
2023-08-11 03:23:53 -05:00
parent d643aa9a6a
commit dfcd0a0ef2
1 changed files with 44 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
index.ts
View File

@@ -1,13 +1,50 @@
import * as cheerio from "cheerio"; import {Cheerio, Element, load} from "cheerio";
import * as fs from "fs"; import * as fs from "fs";
const $ = cheerio.load(fs.readFileSync('list.html')); const $ = load(fs.readFileSync('list.html'));
const classes = $('#scheduleListView').children('.listViewWrapper').toArray(); const classes = $('#scheduleListView').children('.listViewWrapper');
const getClassName = (html) => {
console.log(html); function extractDetails(source: Cheerio<Element>) {
return html.find('list-view-course-title').text(); source.find('div.list-view-crn-info-div > span.bold').toArray().forEach((descriptor) => {
const descriptor_element = $(descriptor);
// console.log([descriptor_element.text(), descriptor_element.next().text()])
})
// console.log(
// source.find('div.listViewMeetingInformation').children('span').last().text()
// )
// For some reason this div divides the information into two parts.
const div_information_divier = source.find('div.list-view-pillbox.ui-pillbox');
const timing_information = div_information_divier.next();
const raw_time = timing_information.text();
const match = raw_time.match(/(\d{2})\s*:\s*(\d{2})\s*(AM|PM)\D*(\d{2})\s*:\s*(\d{2})\s*(AM|PM)/)
if (match === null) {
throw new Error(`Could not parse time: ${raw_time}`);
}
const [start_hour, start_minute, end_hour, end_minute] = [1, 2, 4, 5].map((index) => parseInt(match[index]));
const [start_period, end_period] = [match[3], match[6]];
return {
start: {minute: start_minute, hour: start_hour, period: start_period},
end: {minute: end_minute, hour: end_hour, period: end_period},
days: source.find('div.ui-pillbox-summary.screen-reader').text().split(",").map((day) => day.trim()),
name: source.find('span.list-view-course-title > a.section-details-link').text().trim()
};
} }
console.log(`${classes.length} classes identified.`); console.log(`${classes.length} classes identified.`);
console.log(`First class: ${getClassName(classes[0])}`);
classes.toArray().forEach((element) => {
console.log(extractDetails(
$(element)
))
});
// console.log(classes.toArray().map((element) => element.attribs.class))
// console.log(`First class: ${extractDetails(classes)}`);