[BUG] - Clean up browser and page

[leegunwoo98]

We are currently trying to use this as a long running microservice with exposed api to accept request for getting tiktok data. However, we noticed that pages and contexts that have been detected as a bot or failed for some reason, is not being cleaned. This is evident when using non headless mode and using a global TiktokAPI instance, where we keep a certain number of active sessions and run create session when the number of active session drops below a certain threshold. This would incur increased memory consumption if the unclosed browsers and pages piles up.

[leegunwoo98]

#1217

This PR just wraps it in a try and tries closing the page and context when it reaches an exception