add to docs: fit/predict/trans must support subsampled data

Author: ablaom

docs/src/common_implementation_patterns.md

@@ -31,8 +31,8 @@ implementations fall into one (or more) of the following informally understood p
 
 - Missing Value Imputation
 
-- [Transformers](@ref): Other transformers, such as standardizers, and categorical
-  encoders.
+- [Transformers](@ref transformers): Other transformers, such as standardizers, and
+  categorical encoders.
 
 - [Static Algorithms](@ref): Algorithms that do not learn, in the sense they must be
   re-executed for each new data set (do not generalize), but which have hyperparameters

docs/src/patterns/transformers.md

@@ -1,4 +1,4 @@
-# Transformers
+# [Transformers](@id transformers)
 
 Check out the following examples:
 

src/obs.jl

@@ -57,7 +57,13 @@ fit(learner, observations)` is equivalent to `model = fit(learner, data)`, whene
 `predict(model, ..., observations)` and `transform(model, observations)` must be supported
 alternatives with the same output, whenever `observations = obs(model, data)`.
 
-Implicit in the above requirements is that `obs(learner, _)` and `obs(model, _)` are
+If `LearnAPI.data_interface(learner) == RandomAccess()` (the default), then `fit`,
+`predict` and `transform` must additionally accept `obs` output that has been *subsampled*
+using `MLUtils.getobs`, with the obvious interpretation applying to the outcomes of such
+calls (e.g., if *all* observations are subsampled, then outcomes should be the same as if
+using the original data).
+
+Implicit in preceding requirements is that `obs(learner, _)` and `obs(model, _)` are
 involutive, meaning both the following hold:
 
 ```julia