vizabi / core

Data fetching and transformation module with API based on Grammar of Graphics
1 stars 1 forks source link

Transformation results should be memoized on encoding instead of on marker #27

Open jheeffer opened 4 years ago

jheeffer commented 4 years ago

This makes encodings less coupled to the marker. Encodings would have a specific place to "pick up", an observable, where to get input data from. This can be set to a marker, which saves a reference to the previous transformation's result, or a local dataframe in case of a stand-alone encoding.

This transformInputs place can also be useful as it may be used by other helper methods in the encoding, like limits in trail.