Spam filter on the backend side

[umputun]

On new comment post and edit, we may run spam-check and reject spammy-looking messages. Not sure what the simplest way to do it, maybe some third-part library or service (akismet?), or maybe implement it from scratch.

somewhat related to #344

[paskal]

If we'll pick third-party service, it would make sense to support at least two from the start as there are at least a dozen of them and most are paid. From a quick search, I have no idea which two to pick, aside from the idea that ones with free tier would be more useful for our users.

Also, we should support reviewing and approving \ deleting such messages for the site administrators which sounds like way more complex problem than just marking a message as spam on a backend. I think it would make sense to start implementing this not earlier than we'll have some idea about what to do on the frontend and how to notify the site administrator about these comments.