Support include_metadata Query Parameter

[SteadBytes]

Issue Summary

Keen provides an include_metadata=true query parameter to return execution metadata for ad-hoc queries (docs) which is not currently supported.

This would be useful to have as an optional kwarg for the ad-hoc query methods on KeenClient.

Example API's

Return a dictionary with result and execution_metadata keys:

>>> keen.count("purchases", timeframe="this_14_days", include_metadata=True)
{
    "result": 100,
    "execution_metadata": {
        "total_processing_time": 0.09057211875915527,
        "events_scanned": 1000,
        "properties_per_event": 4,
        "total_properties_scanned": 4000,
    },
}

Return a (result, execution_metadata) tuple:

>>> keen.count("purchases", timeframe="this_14_days", include_metadata=True)
(
    100,
    {
        "total_processing_time": 0.09057211875915527,
        "events_scanned": 1000,
        "properties_per_event": 4,
        "total_properties_scanned": 4000,
    },
)

Very happy to implement and PR for it after discussion on whether to support the feature and on the desired API 👍

[masojus]

Providing a means for retrieving execution metadata is certainly on the radar. I haven't thought about how to include it yet, but your examples aren't likely too far off from what we'd want. The optional parameter list for those queries is already fairly large so there might be some consideration for how to future-proof ourselves against having to add parameters like this to 15 call site. I'll think about it and come back with a plan.

[SteadBytes]

@masojus Yes that's a good point, currently we would need to add the kwarg to all of those methods individually. I'll have a think as well 🤔

Would you want to hold off adding include_metadata until then? (We would benefit great from it at work that's all)