summaryrefslogtreecommitdiff
path: root/scripts/scrape.js
diff options
context:
space:
mode:
authorYuren Hao <97327730+YurenHao0426@users.noreply.github.com>2025-08-06 02:13:31 -0700
committerGitHub <noreply@github.com>2025-08-06 02:13:31 -0700
commit2dda168a062841725cd59102bb194c958b214639 (patch)
tree1dcd26e92fdd2122e784b0a4b1b60e6df8fef638 /scripts/scrape.js
parent2434b68db4d8f86a2d614dc096a2c2cadcb38a91 (diff)
parente5cbd188c6fc654e8ababd37922c018bc8527c44 (diff)
Merge branch 'main' into l7g40n-codex/test-uiuc-api-cors-responsel7g40n-codex/test-uiuc-api-cors-response
Diffstat (limited to 'scripts/scrape.js')
-rw-r--r--scripts/scrape.js6
1 files changed, 6 insertions, 0 deletions
diff --git a/scripts/scrape.js b/scripts/scrape.js
index 93f6a41..d4b2ecb 100644
--- a/scripts/scrape.js
+++ b/scripts/scrape.js
@@ -2,6 +2,7 @@ import { XMLParser } from "fast-xml-parser";
import fs from "fs/promises";
import path from "path";
import dns from "node:dns";
+
import { ProxyAgent } from "undici";
dns.setDefaultResultOrder("ipv4first");
@@ -20,18 +21,23 @@ async function getXML(url) {
const res = await fetch(url, { dispatcher });
if (!res.ok) throw new Error(`Request failed: ${res.status} ${url}`);
return parser.parse(await res.text());
+
}
async function scrapeSchedule(year, term) {
const catalog = {};
const termRoot = await getXML(`${BASE}/schedule/${year}/${term}.xml`);
+
const subjects = termRoot.term?.subjects?.subject;
if (!subjects) throw new Error(`Unexpected XML structure for ${year} ${term}`);
+
const subjHrefs = Array.isArray(subjects) ? subjects.map(s => s['@_href']) : [subjects['@_href']];
for (const subjURL of subjHrefs) {
const subjXML = await getXML(subjURL);
+
const courses = subjXML.subject?.courses?.course || [];
+
const courseList = Array.isArray(courses) ? courses : [courses];
for (const c of courseList) {
const courseURL = c['@_href'];