Add a data stream setting for ignoring duplicate data

This came up in chat with Don Blair, around the question of redundancy in sensor data:

Don:

i wonder if redundancy could be handled at the farmOS level -- you send directly to farmOS, and via geoduck, and via etc ... and farmOS could have some logic that de-duplicates?

Mike:

Hmm yeah could put some logic into farmos that would ignore duplicate data (as an optional setting perhaps)

The implementation would be:

• Add a boolean checkbox setting to listener sensors: "Ignore duplicate data"
• Each time a data point comes in, query the database for an existing data point with the same timestamp, name, and value. If one is found, do not put the new point in.

The reason this should be optional (and disabled by default) is that it adds an additional database query that must run with EVERY data point. So if you're pushing a lot of data, that could have an impact on farmOS performance. Probably not much, since the query is very simple, but still worth making it optional and forcing users to consider it by manually checking the box, in my opinion.