Add/improve show(...) methods

Author: Frankie Robertson

src/aggregators/Aggregators.jl

@@ -33,6 +33,7 @@ using PsychometricsBazaarBase.Integrators: Integrators,
 using PsychometricsBazaarBase.Optimizers: OneDimOptimOptimizer, Optimizer
 using PsychometricsBazaarBase.ConstDistributions: std_normal, std_mv_normal
 import Distributions: pdf
+import Base: show
 
 import FittedItemBanks
 import PsychometricsBazaarBase.IntegralCoeffs

src/aggregators/ability_estimator.jl

@@ -212,6 +212,15 @@ function ModeAbilityEstimator(bits...)
     ModeAbilityEstimator(dist_est, optimizer)
 end
 
+function show(io::IO, ::MIME"text/plain", ability_estimator::ModeAbilityEstimator)
+    println(io, "Estimate ability using its mode")
+    indent_io = indent(io, 2; skip_first=true)
+    print(indent_io, "Distribution estimator ")
+    show(indent_io, ability_estimator.dist_est)
+    print(indent_io, "Optimizer: ")
+    show(indent_io, ability_estimator.optim)
+end
+
 struct MeanAbilityEstimator{
     DistEst <: DistributionAbilityEstimator,
     IntegratorT <: AbilityIntegrator

src/aggregators/optimizers.jl

@@ -10,6 +10,22 @@ function (optim::FunctionOptimizer)(f::F,
     optim.optim(comp_f)
 end
 
+function show(io::IO, ::MIME"text/plain", optim::FunctionOptimizer)
+    indent_io = indent(io, 2)
+    if optim.optim isa OneDimOptimOptimizer || optim.optim isa MultiDimOptimOptimizer || optim.optim isa NativeOneDimOptimOptimizer
+        inner = optim.optim
+        println(io, "Optimizer:")
+        if optim.optim isa NativeOneDimOptimOptimizer
+            name = typeof(inner.method).name.name
+        else
+            name = typeof(inner.optim).name.name
+        end
+        print(indent_io, "Method: ", name)
+        print(indent_io, "Lo: ", inner.lo)
+        print(indent_io, "Hi: ", inner.hi)
+    end
+end
+
 #=
 """
 Argmax + max over the ability likihood given a set of responses with a given

src/next_item_rules/combinators/expectation.jl

@@ -39,6 +39,14 @@ function Aggregators.response_expectation(
         item_idx)
 end
 
+function show(io::IO, ::MIME"text/plain", point_response_expectation::PointResponseExpectation)
+    println(io, "Expected response at point ability estimate")
+    indent_io = indent(io, 2)
+    print(indent_io, "Ability estimator: ")
+    summary(indent_io, point_response_expectation.ability_estimator)
+    println(io)
+end
+
 struct DistributionResponseExpectation{
     DistributionAbilityEstimatorT <: DistributionAbilityEstimator,
     AbilityIntegratorT <: AbilityIntegrator
@@ -124,3 +132,12 @@ function compute_criterion(
     end
     res
 end
+
+function show(io::IO, ::MIME"text/plain", item_criterion::ExpectationBasedItemCriterion)
+    println(io, "Optimize an state/item/item-category criterion based on an expected response")
+    indent_io = indent(io, 2)
+    print(indent_io, "Expected response obtained by: ")
+    show(indent_io, MIME"text/plain"(), item_criterion.response_expectation)
+    print(indent_io, "Criterion: ")
+    show(indent_io, MIME"text/plain"(), item_criterion.criterion)
+end

src/next_item_rules/criteria/pointwise/information.jl

@@ -22,6 +22,10 @@ function compute_criterion_vec(
     -actual
 end
 
+function show(io::IO, ::MIME"text/plain", ::ObservedInformationPointwiseItemCategoryCriterion)
+    println(io, "Observed pointwise item-category information")
+end
+
 """
 See EmpiricalInformationPointwiseItemCategoryCriterion for more details.
 """
@@ -46,6 +50,11 @@ function compute_criterion_vec(
     -actual
 end
 
+
+function show(io::IO, ::MIME"text/plain", ::RawEmpiricalInformationPointwiseItemCategoryCriterion)
+    println(io, "Raw empirical pointwise item-category information")
+end
+
 """
 In equation 10 of [1] we see that we can compute information using 2nd derivatives of log likelihood or 1st derivative squared.
 For single categories, we need to an extra term which disappears when we calculate the total see [2].
@@ -95,6 +104,10 @@ function compute_criterion_vec(
     -actual
 end
 
+function show(io::IO, ::MIME"text/plain", ::EmpiricalInformationPointwiseItemCategoryCriterion)
+    println(io, "Empirical pointwise item-category information")
+end
+
 #=
 """
 This implements Fisher information as a pointwise item criterion.
@@ -116,4 +129,17 @@ function compute_criterion(
     ability
 )
     sum(compute_criterion_vec(tii.pcic, ir, ability))
+end
+
+function show(io::IO, ::MIME"text/plain", rule::TotalItemInformation)
+    if rule.pcic isa ObservedInformationPointwiseItemCategoryCriterion
+        println(io, "Observed pointwise item information")
+    elseif rule.pcic isa RawEmpiricalInformationPointwiseItemCategoryCriterion
+        println(io, "Raw empirical pointwise item information")
+    elseif rule.pcic isa EmpiricalInformationPointwiseItemCategoryCriterion
+        println(io, "Empirical pointwise item information")
+    else
+        print(io, "Total ")
+        show(io, MIME("text/plain"), rule.pcic)
+    end
 end

src/next_item_rules/criteria/state/ability_variance.jl

@@ -68,6 +68,15 @@ function compute_criterion(
         denom)
 end
 
+function show(io::IO, ::MIME"text/plain", criterion::AbilityVarianceStateCriterion)
+    println(io, "Minimise variance of ability estimate")
+    indent_io = indent(io, 2)
+    print(indent_io, "Distribution estimator: ")
+    show(indent_io, MIME"text/plain"(), criterion.dist_est)
+    print(indent_io, "Integrator: ")
+    show(indent_io, MIME"text/plain"(), criterion.integrator)
+end
+
 struct AbilityCovarianceStateMultiCriterion{
     DistEstT <: DistributionAbilityEstimator,
     IntegratorT <: AbilityIntegrator

src/next_item_rules/prelude/next_item_rule.jl

@@ -53,11 +53,13 @@ function best_item(rule::NextItemRule, tracked_responses::TrackedResponses)
     best_item(rule, tracked_responses, tracked_responses.item_bank)
 end
 
-function Base.show(io::IO, ::MIME"text/plain", next_item_rule::ItemStrategyNextItemRule)
-    println(io, "Strategy:")
-    show(indent_io, MIME"text/plain"(), rules.strategy)
-    println(io, "Item criterion:")
-    show(indent_io, MIME"text/plain"(), rules.criterion)
+function Base.show(io::IO, ::MIME"text/plain", rule::ItemStrategyNextItemRule)
+    println(io, "Pick optimal item criterion according to strategy")
+    indent_io = indent(io, 2)
+    print(indent_io, "Strategy: ")
+    show(indent_io, MIME"text/plain"(), rule.strategy)
+    print(indent_io, "Item criterion: ")
+    show(indent_io, MIME"text/plain"(), rule.criterion)
 end
 
 # Default implementation

src/next_item_rules/strategies/balance.jl

@@ -41,9 +41,8 @@ end
 
 function show(io::IO, ::MIME"text/plain", rule::GreedyForcedContentBalancer)
     indent_io = indent(io, 2)
-    println(io, "Greedy + forced content balancer")
+    println(io, "Greedy + forced content balancing")
     println(indent_io, "Target ratio: " * join(rule.targets, ", "))
-    print(indent_io, "Using rule: ")
     show(indent_io, MIME("text/plain"), rule.inner_rule)
 end
 

src/next_item_rules/strategies/pointwise.jl

@@ -16,6 +16,16 @@ function best_item(rule::PointwiseNextItemRule, responses::TrackedResponses, ite
     return idx
 end
 
+function show(io::IO, ::MIME"text/plain", rule::PointwiseNextItemRule)
+    println(io, "Optimize a pointwise criterion at specified points")
+    indent_io = indent(io, 2)
+    points_desc = join(rule.points, ", ")
+    println(indent_io, "Points: $points_desc")
+    print(indent_io, "Criterion: ")
+    show(indent_io, MIME("text/plain"), rule.criterion)
+end
+
+
 function PointwiseFirstNextItemRule(criterion, points, rule)
     FixedRuleSequencer((length(points),), (PointwiseNextItemRule(criterion, points), rule))
 end

src/next_item_rules/strategies/randomesque.jl

@@ -51,4 +51,8 @@ function best_item(
     items
 ) where {ItemCriterionT <: ItemCriterion}
     randomesque(rule.strategy.rng, rule.criterion, responses, items, rule.strategy.k)[1]
+end
+
+function show(io::IO, ::MIME"text/plain", rule::RandomesqueStrategy)
+    println(io, "Randomesque strategy with k = $(rule.k)")
 end