Batch improvements and fix total timeout

[Alextopher]

In #45 I shared how I think the total timeout condition is not being applied correctly. This PR fixes this by adding another helper function.

While I was making this change I noticed some improvements that can be made to the logic of batch lookups.

• I removed a few temporary arrays. Instead I preform the related calculations immediately. This is a free memory and runtime improvement
• I now remove duplicate ips on the client, instead of relying on the server.
• Batches can be made with .chunks()
• I filtered bogon ips from reaching the cache. This should improve cache performance in those use case.

This should be a completely transparent (semver patch) change.

[Alextopher]

There is still required discussion on how to handle errors for batched lookups. I believe this PR doesn't effect the status-quo in that regard and therefor could be merged without considering changes to the error type.

[Alextopher]

The tests seem to be failing because of missing secrets on your end.