You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm working on some integration with the Java implementation and found that it could be helpful to have more verbose construction such that I can recreate a t-digest given the component parts (compression / centroids (and min/max?)). This is basically already done in internally in the fromBytes code but isn't accessible via a public method.
It could also aid in re-construction/ergonomics to have other serialization options such as JSON to help with cross-language support, in the case I create my t-digest via a Spark UDAF but want to consume it in my python application without re-implementing the bytes encoding.
The text was updated successfully, but these errors were encountered:
I'm working on some integration with the Java implementation and found that it could be helpful to have more verbose construction such that I can recreate a t-digest given the component parts (compression / centroids (and min/max?)). This is basically already done in internally in the fromBytes code but isn't accessible via a public method.
It could also aid in re-construction/ergonomics to have other serialization options such as JSON to help with cross-language support, in the case I create my t-digest via a Spark UDAF but want to consume it in my python application without re-implementing the bytes encoding.
The text was updated successfully, but these errors were encountered: