Massively optimize inactive-users script by caching API safe-states

[jagdish-15]

The scripts/analyze-inactivity.js script currently polls the Leetcode API for every single user in our database on a daily basis to see if they've crossed a 90-day inactivity threshold.

If a user solved a problem yesterday, they are mathematically guaranteed to not become inactive for another 89 days. Polling their API during that 89-day padding window wastes massive amounts of compute time (and hits rate limits).

We need to implement a state cache tracking safeUntil timestamps for active users to completely skip unnecessary backend polling.

Implementation Requirements

1. The script should read and maintain a new local JSON file (e.g., data/activity-state.json).
2. Inside the concurrency loop, before fetching a user's baseUrl + username data, check if activity-state.json has a safeUntil date in the future for this user.
3. If Date.now() < safeUntil, we immediately know they are active. Skip the fetchData call completely and just log them as active by continuing the iteration.
4. If they don't have a safeUntil date, or the date has expired, fetch the actual profile via fetchData.
5. Look at their submissionCalendar to find the latest timestamp.
6. Calculate their new safeUntil boundary: latestTimestamp + (90 days). Save this back into the activity-state cache.
7. Save the updated activity-state.json file at the bottom of the script alongside inactive-users.json.

Note: You must thoughtfully construct the cache structure so it doesn't lose old user tracking on subsequent runs!

[Yashaswini-K-P]

I would like to work on this