Tig 209 tiger path egr course sliding window#519
Conversation
… only), and add owner (owners only). Meant for owner > admin hierarchy
The OIT MobileApp API (error 900908 unauthorized) has been replaced
with the Princeton registrar front-end API, which requires no OIT
subscription and is publicly accessible via Bearer token.
- scrape_registrar.py: use cloudscraper to bypass Cloudflare and
extract apiToken and apiBaseUrl from registrar.princeton.edu;
add get_course_ids_for_term() via /classes/{term} endpoint;
add get_full_course_details() returning all course fields in one call
- scrape_parse.py: remove MobileApp dependency entirely; derive
semester dates from term code; use registrar API in parallel
- scrape_all.py: warn and skip instead of crashing on 0-course terms
(historical terms 1214-1234 are not in the registrar API)
- requirements.txt: add cloudscraper==1.2.71
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…GR-CourseSlidingWindow
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9bc8988e16
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| .then((data) => { | ||
| setDetails(data); | ||
| setLoading(false); |
There was a problem hiding this comment.
Each course selection triggers a new fetch, but prior requests are never cancelled or ignored; if a user clicks courses quickly, an older response can resolve last and overwrite details for the currently selected course. In that case the panel can display ratings/comments from course A while the header still shows course B, which is a correctness issue in normal interactive use.
Useful? React with 👍 / 👎.
| if not evaluations: | ||
| return JsonResponse(empty) |
There was a problem hiding this comment.
This early return exits before the DB description lookup and course metadata enrichment, so courses with no evaluation rows always get an empty payload even when a description exists locally. The UI explicitly supports rendering description without eval data, so this branch drops useful details for exactly that case.
Useful? React with 👍 / 👎.
| evaluations.sort(key=lambda e: e.get("evalTerm", ""), reverse=True) | ||
| raw_rating = evaluations[0].get("rating") | ||
| rating = round(float(raw_rating), 2) if raw_rating is not None else None | ||
| return course_id, rating |
There was a problem hiding this comment.
The helper checks course_details_* cache but never stores successful fetches back into that key. As a result, callers like get_schedule repeatedly hit the external evaluations API for the same courses (especially when quality_rating is missing or null), which adds avoidable latency and external load instead of converging to cache hits.
Useful? React with 👍 / 👎.
2 participants
No description provided.