created fallback implementation for legal_action_space_mask (#644)

* created fallback implementation for legal_action_space_mask

* fixed doc typo

Author: baedan

src/ReinforcementLearningBase/src/interface.jl

@@ -450,9 +450,14 @@ legal_action_space(::MinimalActionSet, env, player) = action_space(env)
 """
     legal_action_space_mask(env, player=current_player(env)) -> AbstractArray{Bool}
 
-Required for environments of [`FULL_ACTION_SET`](@ref).
+Required for environments of [`FULL_ACTION_SET`](@ref). As a default implementation,
+     [`legal_action_space_mask`](@ref) creates a mask of [`action_space`](@ref) with
+     the subset [`legal_action_space`](@ref).
 """
-@multi_agent_env_api legal_action_space_mask(env::AbstractEnv, player = current_player(env))
+@multi_agent_env_api legal_action_space_mask(env::AbstractEnv, player = current_player(env)) = 
+    map(action_space(env, player)) do action
+        action in legal_action_space(env, player)
+    end
 
 """
     state(env, style=[DefaultStateStyle(env)], player=[current_player(env)])

src/ReinforcementLearningCore/src/policies/q_based_policies/explorers/epsilon_greedy_explorer.jl

@@ -24,7 +24,7 @@ Two kinds of epsilon-decreasing strategy are implemented here (`linear` and `exp
 - `ϵ_stable::Float64`: the epsilon after `warmup_steps + decay_steps`.
 - `is_break_tie=false`: randomly select an action of the same maximum values if set to `true`.
 - `rng=Random.GLOBAL_RNG`: set the internal RNG.
-- `is_training=true`, in training mode, `step` will not be updated. And the `ϵ` will be set to 0.
+- `is_training=true`: when not in training mode, `step` will not be updated. And the `ϵ` will be set to 0.
 
 # Example