fix(scraper): add request timeout and HTTP error handling to prevent workflow hangs

[glatinone]

While looking at scrape_url in scraper.py, I noticed the HTTP request to GitHub Trending doesn’t set a timeout and relies on an assert to validate the status code. Given this repository runs the scraper daily via schedule.yml, a slow or transient network condition could block the job or abort it early.

Context and impact:
: The GET call currently has no timeout, so a stalled connection can hang the job indefinitely.
: assert r.status_code == 200 raises an AssertionError on any non-200 response, exiting the run instead of skipping a problematic page.
: The notifier script (check_stars.py) already uses explicit timeouts for outbound requests, so adding the same resilience to the scraper aligns behavior across jobs.

Changes introduced:
: Add a REQUEST_TIMEOUT environment toggle with a default (15s) and pass it to requests.get.
: Replace assert with r.raise_for_status().
: Wrap the request and parsing in try/except to catch Timeout and RequestException, logging the error and returning a safe fallback structure so the loop continues for other languages.

Stability benefits:
: Prevents the scheduled workflow from hanging on slow responses, improving determinism for the daily update.
: Avoids abrupt termination on non-200 responses and gracefully skips transient failures, reducing noise and failed runs.
: Keeps behavior consistent with the existing star checker’s timeout usage, consolidating network hygiene across scripts.