# frozen_string_literal: true

require 'rumale/values'
require 'rumale/utils'
require 'rumale/base/base_estimator'
require 'rumale/base/classifier'
require 'rumale/tree/decision_tree_classifier'

module Rumale
  module Ensemble
    # AdaBoostClassifier is a class that implements AdaBoost (SAMME.R) for classification.
    # This class uses decision tree for a weak learner.
    #
    # @example
    #   estimator =
    #     Rumale::Ensemble::AdaBoostClassifier.new(
    #       n_estimators: 10, criterion: 'gini', max_depth: 3, max_leaf_nodes: 10, min_samples_leaf: 5, random_seed: 1)
    #   estimator.fit(training_samples, traininig_labels)
    #   results = estimator.predict(testing_samples)
    #
    # *Reference*
    # - J. Zhu, S. Rosset, H. Zou, and T.Hashie, "Multi-class AdaBoost," Technical Report No. 430, Department of Statistics, University of Michigan, 2005.
    class AdaBoostClassifier
      include Base::BaseEstimator
      include Base::Classifier

      # Return the set of estimators.
      # @return [Array<DecisionTreeClassifier>]
      attr_reader :estimators

      # Return the class labels.
      # @return [Numo::Int32] (size: n_classes)
      attr_reader :classes

      # Return the importance for each feature.
      # @return [Numo::DFloat] (size: n_features)
      attr_reader :feature_importances

      # Return the random generator for random selection of feature index.
      # @return [Random]
      attr_reader :rng

      # Create a new classifier with AdaBoost.
      #
      # @param n_estimators [Integer] The numeber of decision trees for contructing AdaBoost classifier.
      # @param criterion [String] The function to evalue spliting point. Supported criteria are 'gini' and 'entropy'.
      # @param max_depth [Integer] The maximum depth of the tree.
      #   If nil is given, decision tree grows without concern for depth.
      # @param max_leaf_nodes [Integer] The maximum number of leaves on decision tree.
      #   If nil is given, number of leaves is not limited.
      # @param min_samples_leaf [Integer] The minimum number of samples at a leaf node.
      # @param max_features [Integer] The number of features to consider when searching optimal split point.
      #   If nil is given, split process considers all features.
      # @param random_seed [Integer] The seed value using to initialize the random generator.
      #   It is used to randomly determine the order of features when deciding spliting point.
      def initialize(n_estimators: 50,
                     criterion: 'gini', max_depth: nil, max_leaf_nodes: nil, min_samples_leaf: 1,
                     max_features: nil, random_seed: nil)
        check_params_type_or_nil(Integer, max_depth: max_depth, max_leaf_nodes: max_leaf_nodes,
                                          max_features: max_features, random_seed: random_seed)
        check_params_integer(n_estimators: n_estimators, min_samples_leaf: min_samples_leaf)
        check_params_string(criterion: criterion)
        check_params_positive(n_estimators: n_estimators, max_depth: max_depth,
                              max_leaf_nodes: max_leaf_nodes, min_samples_leaf: min_samples_leaf,
                              max_features: max_features)
        @params = {}
        @params[:n_estimators] = n_estimators
        @params[:criterion] = criterion
        @params[:max_depth] = max_depth
        @params[:max_leaf_nodes] = max_leaf_nodes
        @params[:min_samples_leaf] = min_samples_leaf
        @params[:max_features] = max_features
        @params[:random_seed] = random_seed
        @params[:random_seed] ||= srand
        @estimators = nil
        @classes = nil
        @feature_importances = nil
        @rng = Random.new(@params[:random_seed])
      end

      # Fit the model with given training data.
      #
      # @param x [Numo::DFloat] (shape: [n_samples, n_features]) The training data to be used for fitting the model.
      # @param y [Numo::Int32] (shape: [n_samples]) The labels to be used for fitting the model.
      # @return [AdaBoostClassifier] The learned classifier itself.
      def fit(x, y) # rubocop:disable Metrics/AbcSize
        check_sample_array(x)
        check_label_array(y)
        check_sample_label_size(x, y)
        ## Initialize some variables.
        n_samples, n_features = x.shape
        @estimators = []
        @feature_importances = Numo::DFloat.zeros(n_features)
        @params[:max_features] = n_features unless @params[:max_features].is_a?(Integer)
        @params[:max_features] = [[1, @params[:max_features]].max, n_features].min
        @classes = Numo::Int32.asarray(y.to_a.uniq.sort)
        n_classes = @classes.shape[0]
        ## Boosting.
        classes_arr = @classes.to_a
        y_codes = Numo::DFloat.zeros(n_samples, n_classes) - 1.fdiv(n_classes - 1)
        n_samples.times { |n| y_codes[n, classes_arr.index(y[n])] = 1.0 }
        observation_weights = Numo::DFloat.zeros(n_samples) + 1.fdiv(n_samples)
        @params[:n_estimators].times do |_t|
          # Fit classfier.
          ids = Rumale::Utils.choice_ids(n_samples, observation_weights, @rng)
          break if y[ids].to_a.uniq.size != n_classes
          tree = Tree::DecisionTreeClassifier.new(
            criterion: @params[:criterion], max_depth: @params[:max_depth],
            max_leaf_nodes: @params[:max_leaf_nodes], min_samples_leaf: @params[:min_samples_leaf],
            max_features: @params[:max_features], random_seed: @rng.rand(Rumale::Values.int_max)
          )
          tree.fit(x[ids, true], y[ids])
          # Calculate estimator error.
          proba = tree.predict_proba(x).clip(1.0e-15, nil)
          p = Numo::Int32.asarray(Array.new(n_samples) { |n| @classes[proba[n, true].max_index] })
          inds = p.ne(y)
          error = (observation_weights * inds).sum / observation_weights.sum
          # Store model.
          @estimators.push(tree)
          @feature_importances += tree.feature_importances
          break if error.zero?
          # Update observation weights.
          log_proba = Numo::NMath.log(proba)
          observation_weights *= Numo::NMath.exp(-1.0 * (n_classes - 1).fdiv(n_classes) * (y_codes * log_proba).sum(1))
          observation_weights = observation_weights.clip(1.0e-15, nil)
          sum_observation_weights = observation_weights.sum
          break if sum_observation_weights.zero?
          observation_weights /= sum_observation_weights
        end
        @feature_importances /= @feature_importances.sum
        self
      end

      # Calculate confidence scores for samples.
      #
      # @param x [Numo::DFloat] (shape: [n_samples, n_features]) The samples to compute the scores.
      # @return [Numo::DFloat] (shape: [n_samples, n_classes]) Confidence score per sample.
      def decision_function(x)
        check_sample_array(x)
        n_samples, = x.shape
        n_classes = @classes.size
        sum_probs = Numo::DFloat.zeros(n_samples, n_classes)
        @estimators.each do |tree|
          log_proba = Numo::NMath.log(tree.predict_proba(x).clip(1.0e-15, nil))
          sum_probs += (n_classes - 1) * (log_proba - 1.fdiv(n_classes) * Numo::DFloat[log_proba.sum(1)].transpose)
        end
        sum_probs /= @estimators.size
      end

      # Predict class labels for samples.
      #
      # @param x [Numo::DFloat] (shape: [n_samples, n_features]) The samples to predict the labels.
      # @return [Numo::Int32] (shape: [n_samples]) Predicted class label per sample.
      def predict(x)
        check_sample_array(x)
        n_samples, = x.shape
        probs = decision_function(x)
        Numo::Int32.asarray(Array.new(n_samples) { |n| @classes[probs[n, true].max_index] })
      end

      # Predict probability for samples.
      #
      # @param x [Numo::DFloat] (shape: [n_samples, n_features]) The samples to predict the probailities.
      # @return [Numo::DFloat] (shape: [n_samples, n_classes]) Predicted probability of each class per sample.
      def predict_proba(x)
        check_sample_array(x)
        n_classes = @classes.size
        probs = Numo::NMath.exp(1.fdiv(n_classes - 1) * decision_function(x))
        sum_probs = probs.sum(1)
        probs /= Numo::DFloat[sum_probs].transpose
        probs
      end

      # Dump marshal data.
      # @return [Hash] The marshal data about AdaBoostClassifier.
      def marshal_dump
        { params: @params,
          estimators: @estimators,
          classes: @classes,
          feature_importances: @feature_importances,
          rng: @rng }
      end

      # Load marshal data.
      # @return [nil]
      def marshal_load(obj)
        @params = obj[:params]
        @estimators = obj[:estimators]
        @classes = obj[:classes]
        @feature_importances = obj[:feature_importances]
        @rng = obj[:rng]
        nil
      end
    end
  end
end