mirror of
https://github.com/Xevion/schedule-extract.git
synced 2025-12-08 04:08:26 -06:00
Add date-fns for date parsing, use inspect for deep objects, class date range, general details
This commit is contained in:
56
index.ts
56
index.ts
@@ -1,24 +1,30 @@
|
||||
import {Cheerio, Element, load} from "cheerio";
|
||||
import * as fs from "fs";
|
||||
import {readFileSync} from "fs";
|
||||
import {parse} from "date-fns";
|
||||
import {inspect} from "util";
|
||||
|
||||
const $ = load(fs.readFileSync('list.html'));
|
||||
const $ = load(readFileSync('list.html'));
|
||||
const classes = $('#scheduleListView').children('.listViewWrapper');
|
||||
|
||||
function getOffset(period: string) {
|
||||
switch (period) {
|
||||
case "AM":
|
||||
return 0;
|
||||
case "PM":
|
||||
return 12;
|
||||
default:
|
||||
throw new Error(`Unknown period: ${period}`);
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
function extractDetails(source: Cheerio<Element>) {
|
||||
source.find('div.list-view-crn-info-div > span.bold').toArray().forEach((descriptor) => {
|
||||
const details: [string, string][] = source.find('div.list-view-crn-info-div > span.bold').toArray().map((descriptor) => {
|
||||
const descriptor_element = $(descriptor);
|
||||
|
||||
// console.log([descriptor_element.text(), descriptor_element.next().text()])
|
||||
return [descriptor_element.text(), descriptor_element.next().text()]
|
||||
})
|
||||
|
||||
// console.log(
|
||||
// source.find('div.listViewMeetingInformation').children('span').last().text()
|
||||
// )
|
||||
|
||||
// For some reason this div divides the information into two parts.
|
||||
const div_information_divier = source.find('div.list-view-pillbox.ui-pillbox');
|
||||
const timing_information = div_information_divier.next();
|
||||
const timing_information = source.find('div.list-view-pillbox.ui-pillbox').next();
|
||||
|
||||
const raw_time = timing_information.text();
|
||||
const match = raw_time.match(/(\d{2})\s*:\s*(\d{2})\s*(AM|PM)\D*(\d{2})\s*:\s*(\d{2})\s*(AM|PM)/)
|
||||
@@ -29,22 +35,24 @@ function extractDetails(source: Cheerio<Element>) {
|
||||
const [start_hour, start_minute, end_hour, end_minute] = [1, 2, 4, 5].map((index) => parseInt(match[index]));
|
||||
const [start_period, end_period] = [match[3], match[6]];
|
||||
|
||||
const raw_date = source.find('span.meetingTimes').text();
|
||||
const [start_date, end_date] = raw_date.split("--").map((date) => parse(date.trim(), "MM/dd/yyyy", new Date()));
|
||||
|
||||
return {
|
||||
start: {minute: start_minute, hour: start_hour, period: start_period},
|
||||
end: {minute: end_minute, hour: end_hour, period: end_period},
|
||||
date: {
|
||||
start: start_date,
|
||||
end: end_date,
|
||||
},
|
||||
time: {
|
||||
start: {minute: start_minute, hour: start_hour + getOffset(start_period)},
|
||||
end: {minute: end_minute, hour: end_hour + getOffset(end_period)},
|
||||
},
|
||||
days: source.find('div.ui-pillbox-summary.screen-reader').text().split(",").map((day) => day.trim()),
|
||||
name: source.find('span.list-view-course-title > a.section-details-link').text().trim()
|
||||
};
|
||||
}
|
||||
|
||||
console.log(`${classes.length} classes identified.`);
|
||||
|
||||
classes.toArray().forEach((element) => {
|
||||
console.log(extractDetails(
|
||||
$(element)
|
||||
))
|
||||
});
|
||||
|
||||
// console.log(classes.toArray().map((element) => element.attribs.class))
|
||||
|
||||
// console.log(`First class: ${extractDetails(classes)}`);
|
||||
console.log(
|
||||
inspect(classes.toArray().map((element) => extractDetails($(element))), {colors: true, depth: null})
|
||||
);
|
||||
|
||||
@@ -15,6 +15,7 @@
|
||||
"dependencies": {
|
||||
"@types/node": "^20.4.9",
|
||||
"cheerio": "^1.0.0-rc.12",
|
||||
"date-fns": "^2.30.0",
|
||||
"ics": "^3.2.0",
|
||||
"ts-node": "^10.9.1",
|
||||
"typescript": "^5.1.6"
|
||||
|
||||
19
yarn.lock
19
yarn.lock
@@ -2,6 +2,13 @@
|
||||
# yarn lockfile v1
|
||||
|
||||
|
||||
"@babel/runtime@^7.21.0":
|
||||
version "7.22.10"
|
||||
resolved "https://registry.yarnpkg.com/@babel/runtime/-/runtime-7.22.10.tgz#ae3e9631fd947cb7e3610d3e9d8fef5f76696682"
|
||||
integrity sha512-21t/fkKLMZI4pqP2wlmsQAWnYW1PDyKyyUV4vCi+B25ydmdaYTKXPwCj0BzSUnZf4seIiYvSA3jcZ3gdsMFkLQ==
|
||||
dependencies:
|
||||
regenerator-runtime "^0.14.0"
|
||||
|
||||
"@cspotcode/source-map-support@^0.8.0":
|
||||
version "0.8.1"
|
||||
resolved "https://registry.yarnpkg.com/@cspotcode/source-map-support/-/source-map-support-0.8.1.tgz#00629c35a688e05a88b1cda684fb9d5e73f000a1"
|
||||
@@ -118,6 +125,13 @@ css-what@^6.1.0:
|
||||
resolved "https://registry.yarnpkg.com/css-what/-/css-what-6.1.0.tgz#fb5effcf76f1ddea2c81bdfaa4de44e79bac70f4"
|
||||
integrity sha512-HTUrgRJ7r4dsZKU6GjmpfRK1O76h97Z8MfS1G0FozR+oF2kG6Vfe8JE6zwrkbxigziPHinCJ+gCPjA9EaBDtRw==
|
||||
|
||||
date-fns@^2.30.0:
|
||||
version "2.30.0"
|
||||
resolved "https://registry.yarnpkg.com/date-fns/-/date-fns-2.30.0.tgz#f367e644839ff57894ec6ac480de40cae4b0f4d0"
|
||||
integrity sha512-fnULvOpxnC5/Vg3NCiWelDsLiUc9bRwAPs/+LfTLNvetFCtCTN+yQz15C/fs4AwX1R9K5GLtLfn8QW+dWisaAw==
|
||||
dependencies:
|
||||
"@babel/runtime" "^7.21.0"
|
||||
|
||||
diff@^4.0.1:
|
||||
version "4.0.2"
|
||||
resolved "https://registry.yarnpkg.com/diff/-/diff-4.0.2.tgz#60f3aecb89d5fae520c11aa19efc2bb982aade7d"
|
||||
@@ -213,6 +227,11 @@ property-expr@^2.0.5:
|
||||
resolved "https://registry.yarnpkg.com/property-expr/-/property-expr-2.0.5.tgz#278bdb15308ae16af3e3b9640024524f4dc02cb4"
|
||||
integrity sha512-IJUkICM5dP5znhCckHSv30Q4b5/JA5enCtkRHYaOVOAocnH/1BQEYTC5NMfT3AVl/iXKdr3aqQbQn9DxyWknwA==
|
||||
|
||||
regenerator-runtime@^0.14.0:
|
||||
version "0.14.0"
|
||||
resolved "https://registry.yarnpkg.com/regenerator-runtime/-/regenerator-runtime-0.14.0.tgz#5e19d68eb12d486f797e15a3c6a918f7cec5eb45"
|
||||
integrity sha512-srw17NI0TUWHuGa5CFGGmhfNIeja30WMBfbslPNhf6JrqQlLN5gcrvig1oqPxiVaXb0oW0XRKtH6Nngs5lKCIA==
|
||||
|
||||
tiny-case@^1.0.3:
|
||||
version "1.0.3"
|
||||
resolved "https://registry.yarnpkg.com/tiny-case/-/tiny-case-1.0.3.tgz#d980d66bc72b5d5a9ca86fb7c9ffdb9c898ddd03"
|
||||
|
||||
Reference in New Issue
Block a user