alphapairs and alphavectors

docs/src/alpha_vector.md

@@ -1,12 +1,13 @@
 # Alpha Vector Policy
 
-Represents a policy with a set of alpha vectors (See `AlphaVectorPolicy` constructor docstring). In addition to finding the optimal action with `action`, pairs mapping each alpha vector to an action can be iterated through with [`alpha_actions`](@ref).
+Represents a policy with a set of alpha vectors (See `AlphaVectorPolicy` constructor docstring). In addition to finding the optimal action with `action`, the alpha vectors can be accessed with [`alphavectors`](@ref) or [`alphapairs`](@ref).
 
 Determining the estimated value and optimal action depends on calculating the dot product between alpha vectors and a belief vector. [`POMDPPolicies.beliefvec(pomdp, b)`](@ref) is used to create this vector and can be overridden for new belief types for efficiency.
 
 
 ```@docs
 AlphaVectorPolicy
-alpha_actions
+alphavectors
+alphapairs
 POMDPPolicies.beliefvec
 ``` 

src/POMDPPolicies.jl

@@ -25,7 +25,8 @@ export
 
 export
     AlphaVectorPolicy,
-    alpha_actions
+    alphavectors,
+    alphapairs
 
 include("alpha_vector.jl")
 

src/alpha_vector.jl

@@ -11,7 +11,7 @@ Construct a policy from alpha vectors.
 
 Represents a policy with a set of alpha vectors.
 
-Use `action` to get the best action for a belief, an `alpha_actions` to iterate through alpha-vector => action pairs.
+Use `action` to get the best action for a belief, and `alphavectors` and `alphapairs` to 
 
 # Fields
 - `pomdp::P` the POMDP problem 
@@ -43,9 +43,14 @@ end
 updater(p::AlphaVectorPolicy) = DiscreteUpdater(p.pomdp)
 
 """
-Return an iterator of alpha-vector => action pairs in the policy.
+Return an iterator of alpha vector-action pairs in the policy.
 """
-alpha_actions(p::AlphaVectorPolicy) = (p.alphas[i]=>p.action_map[i] for i in 1:length(p.alphas))
+alphapairs(p::AlphaVectorPolicy) = (p.alphas[i]=>p.action_map[i] for i in 1:length(p.alphas))
+
+"""
+Return the alpha vectors.
+"""
+alphavectors(p::AlphaVectorPolicy) = p.alphas
 
 # The three functions below rely on beliefvec being implemented for the belief type 
 # Implementations of beliefvec are below

test/test_alpha_policy.jl

@@ -9,7 +9,8 @@ let
     alphas = [ -16.0629 -19.4557; -36.5093 -29.4557]
     policy = AlphaVectorPolicy(pomdp, alphas)
 
-    @test collect(alpha_actions(policy)) == [[-16.0629, -36.5093]=>false, [-19.4557, -29.4557]=>true]
+    @test Set(alphapairs(policy)) == Set([[-16.0629, -36.5093]=>false, [-19.4557, -29.4557]=>true])
+    @test Set(alphavectors(policy)) == Set([[-16.0629, -36.5093], [-19.4557, -29.4557]])
 
     # initial belief is 100% confidence in baby not being hungry
     @test isapprox(value(policy, b0), -16.0629)