Filter binary covariates on the server-side via SQL

[anthonysena]

Building on the work done in #174, we'd like to use the minCharacterizationMean to filter out covariates that are smaller than this threshold. The current implementation (v3.5.1) does this in R after downloading all of the data via SQL. A more performant option would be to filter this data directly when retrieving the data via SQL.

To outline a plan for this: the minCharacterizationMean parameter would need to get passed into this function:

https://github.com/OHDSI/FeatureExtraction/blob/a93992be1a2f7d3710aa8249287af937872d50f3/java/org/ohdsi/featureExtraction/FeatureExtraction.java#L703-L746

The resulting SQL could be augmented to then obtain the average_value where it is >= minCharacterizationMean.