Add retry to coursebook scraper

[justinschwerdtfeger]

This PR was moved from a fork to a branch

This PR improves coursebook scraper's ability to retry and recover. Here's the main changes

• log.Fatals have been replaced with simply returning an error or log.panic where the error shouldn't be recovered. log.Fatal runs Os.Exit, which is not recoverable. We should consider this for other tools as well.
• Improve error handling and behavior during network timeouts or non-200 server response codes adding more exponential retry.
• Retry flag will fully restart coursebook scraper upon error or panic. With the other changes, this is unlikely to fix any issues that went wrong during scraping, but it could. It acts as mostly a last resort, so the flag is optional
• Wrong parameters end scraping early with log.Fatal and a clear error.