Merge 9348dbc into 05310cd

.travis.yml

@@ -1,5 +1,5 @@
 language: dart
 dart:
-  - "2.4.1"
+  - "2.7.2"
 script: pub run grinder start
 after_success: pub run grinder finish

build.yaml

@@ -0,0 +1,14 @@
+targets:
+  $default:
+    builders:
+      json_serializable:
+        options:
+          any_map: false
+          checked: true
+          create_to_json: true
+          disallow_unrecognized_keys: true
+          explicit_to_json: false
+          field_rename: none
+          ignore_unannotated: false
+          include_if_null: true
+          nullable: true

lib/src/classifier/classifier.dart

@@ -4,8 +4,6 @@ import 'package:ml_dataframe/ml_dataframe.dart';
 /// An interface for any classifier (linear, non-linear, parametric,
 /// non-parametric, etc.)
 abstract class Classifier extends Predictor {
-  List<String> get classNames;
-
   /// Returns predicted distribution of probabilities for each observation in
   /// the passed [testFeatures]
   DataFrame predictProbabilities(DataFrame testFeatures);

lib/src/classifier/decision_tree_classifier/_helpers/create_decision_tree_classifier.dart → lib/src/classifier/decision_tree_classifier/_helper/create_decision_tree_classifier.dart

@@ -4,8 +4,8 @@ import 'package:ml_algo/src/di/dependencies.dart';
 import 'package:ml_algo/src/helpers/validate_train_data.dart';
 import 'package:ml_algo/src/helpers/validate_tree_solver_max_depth.dart';
 import 'package:ml_algo/src/helpers/validate_tree_solver_min_samples_count.dart';
-import 'package:ml_algo/src/helpers/validate_tree_solver_minimal_error.dart';
-import 'package:ml_algo/src/tree_solver/_helpers/create_decision_tree_solver.dart';
+import 'package:ml_algo/src/helpers/validate_tree_solver_min_error.dart';
+import 'package:ml_algo/src/tree_trainer/_helpers/create_decision_tree_trainer.dart';
 import 'package:ml_dataframe/ml_dataframe.dart';
 import 'package:ml_linalg/dtype.dart';
 
@@ -18,20 +18,15 @@ DecisionTreeClassifier createDecisionTreeClassifier(
   DType dtype,
 ) {
   validateTrainData(trainData, [targetName]);
-  validateTreeSolverMinimalError(minError);
+  validateTreeSolverMinError(minError);
   validateTreeSolversMinSamplesCount(minSamplesCount);
   validateTreeSolverMaxDepth(maxDepth);
 
-  final solver = createDecisionTreeSolver(
-    trainData,
-    targetName,
-    minError,
-    minSamplesCount,
-    maxDepth,
-    dtype,
-  );
+  final trainer = createDecisionTreeTrainer(trainData, targetName, minError,
+    minSamplesCount, maxDepth);
+  final treeRootNode = trainer.train(trainData.toMatrix(dtype));
 
   return dependencies
       .getDependency<DecisionTreeClassifierFactory>()
-      .create(solver, targetName, dtype);
+      .create(treeRootNode, targetName, dtype);
 }

lib/src/classifier/decision_tree_classifier/_helper/create_decision_tree_classifier_from_json.dart

@@ -0,0 +1,15 @@
+import 'dart:convert';
+
+import 'package:ml_algo/src/classifier/decision_tree_classifier/decision_tree_classifier.dart';
+import 'package:ml_algo/src/classifier/decision_tree_classifier/decision_tree_classifier_impl.dart';
+
+DecisionTreeClassifier createDecisionTreeClassifierFromJson(String json) {
+  if (json.isEmpty) {
+    throw Exception('Provided JSON object is empty, please provide a proper '
+        'JSON object');
+  }
+
+  final decodedJson = jsonDecode(json) as Map<String, dynamic>;
+
+  return DecisionTreeClassifierImpl.fromJson(decodedJson);
+}

lib/src/classifier/decision_tree_classifier/decision_tree_classifier.dart

@@ -1,5 +1,7 @@
 import 'package:ml_algo/src/classifier/classifier.dart';
-import 'package:ml_algo/src/classifier/decision_tree_classifier/_helpers/create_decision_tree_classifier.dart';
+import 'package:ml_algo/src/classifier/decision_tree_classifier/_helper/create_decision_tree_classifier.dart';
+import 'package:ml_algo/src/classifier/decision_tree_classifier/_helper/create_decision_tree_classifier_from_json.dart';
+import 'package:ml_algo/src/common/serializable/serializable.dart';
 import 'package:ml_algo/src/model_selection/assessable.dart';
 import 'package:ml_dataframe/ml_dataframe.dart';
 import 'package:ml_linalg/dtype.dart';
@@ -10,21 +12,24 @@ import 'package:ml_linalg/dtype.dart';
 /// subsets until the subsets conforming certain stop criteria are found.
 ///
 /// Process of forming such a recursive subsets structure is called
-/// decision tree learning. Once a decision tree learned, it may use to classify
-/// new samples with the same features, as were used to learn the tree.
-abstract class DecisionTreeClassifier implements Classifier, Assessable {
+/// decision tree learning. Once a decision tree learned, it may be used to
+/// classify new samples with the same features that were used to learn the
+/// tree.
+abstract class DecisionTreeClassifier implements
+    Classifier, Assessable, Serializable {
   /// Parameters:
   ///
-  /// [trainData] A [DataFrame] with observations, that will be used by the
+  /// [trainData] A [DataFrame] with observations that will be used by the
   /// classifier to learn a decision tree. Must contain [targetName] column.
   ///
   /// [targetName] A name of a column in [trainData] that contains class
   /// labels
   ///
-  /// [minError] A value from range 0..1 (both inclusive). The value denotes a
-  /// minimal error on a single decision tree node and is used as a stop
-  /// criteria to avoid farther decision's tree node splitting: if the node is
-  /// good enough, there is no need to split it and thus it will become a leaf.
+  /// [minError] A value within the range 0..1 (both inclusive). The value
+  /// denotes a minimal error on a single decision tree node and is used as a
+  /// stop criteria to avoid farther decision's tree node splitting: if the
+  /// node is good enough, there is no need to split it and thus it will become
+  /// a leaf.
   ///
   /// [minSamplesCount] A minimal number of samples (observations) on the
   /// decision's tree node. The value is used as a stop criteria to avoid
@@ -47,4 +52,7 @@ abstract class DecisionTreeClassifier implements Classifier, Assessable {
     maxDepth,
     dtype,
   );
+
+  factory DecisionTreeClassifier.fromJson(String json) =>
+      createDecisionTreeClassifierFromJson(json);
 }

lib/src/classifier/decision_tree_classifier/decision_tree_classifier_factory.dart

@@ -1,10 +1,10 @@
 import 'package:ml_algo/ml_algo.dart';
-import 'package:ml_algo/src/tree_solver/tree_solver.dart';
+import 'package:ml_algo/src/tree_trainer/tree_node/tree_node.dart';
 import 'package:ml_linalg/dtype.dart';
 
 abstract class DecisionTreeClassifierFactory {
   DecisionTreeClassifier create(
-      TreeSolver solver,
+      TreeNode root,
       String targetName,
       DType dtype,
   );

lib/src/classifier/decision_tree_classifier/decision_tree_classifier_factory_impl.dart

@@ -1,18 +1,21 @@
 import 'package:ml_algo/src/classifier/decision_tree_classifier/decision_tree_classifier.dart';
 import 'package:ml_algo/src/classifier/decision_tree_classifier/decision_tree_classifier_factory.dart';
 import 'package:ml_algo/src/classifier/decision_tree_classifier/decision_tree_classifier_impl.dart';
-import 'package:ml_algo/src/tree_solver/tree_solver.dart';
+import 'package:ml_algo/src/tree_trainer/tree_node/tree_node.dart';
 import 'package:ml_linalg/dtype.dart';
 
 class DecisionTreeClassifierFactoryImpl implements
     DecisionTreeClassifierFactory {
-
   const DecisionTreeClassifierFactoryImpl();
 
   @override
   DecisionTreeClassifier create(
-      TreeSolver solver,
+      TreeNode root,
       String targetName,
       DType dtype,
-  ) => DecisionTreeClassifierImpl(solver, targetName, dtype);
+  ) => DecisionTreeClassifierImpl(
+    root,
+    targetName,
+    dtype,
+  );
 }

lib/src/classifier/decision_tree_classifier/decision_tree_classifier_impl.dart

@@ -1,31 +1,65 @@
+import 'package:json_annotation/json_annotation.dart';
 import 'package:ml_algo/src/classifier/decision_tree_classifier/decision_tree_classifier.dart';
+import 'package:ml_algo/src/classifier/decision_tree_classifier/decision_tree_json_keys.dart';
+import 'package:ml_algo/src/common/dtype_serializer/dtype_to_json.dart';
+import 'package:ml_algo/src/common/dtype_serializer/from_dtype_json.dart';
+import 'package:ml_algo/src/common/serializable/serializable_mixin.dart';
 import 'package:ml_algo/src/predictor/assessable_predictor_mixin.dart';
-import 'package:ml_algo/src/tree_solver/tree_solver.dart';
+import 'package:ml_algo/src/tree_trainer/leaf_label/leaf_label.dart';
+import 'package:ml_algo/src/tree_trainer/tree_node/_helper/from_tree_node_json.dart';
+import 'package:ml_algo/src/tree_trainer/tree_node/_helper/tree_node_to_json.dart';
+import 'package:ml_algo/src/tree_trainer/tree_node/tree_node.dart';
 import 'package:ml_dataframe/ml_dataframe.dart';
 import 'package:ml_linalg/dtype.dart';
 import 'package:ml_linalg/matrix.dart';
 import 'package:ml_linalg/vector.dart';
 
-class DecisionTreeClassifierImpl with AssessablePredictorMixin
-    implements DecisionTreeClassifier {
+part 'decision_tree_classifier_impl.g.dart';
 
-  DecisionTreeClassifierImpl(this._solver, String className, this.dtype)
-      : classNames = [className];
+@JsonSerializable()
+class DecisionTreeClassifierImpl
+    with
+        AssessablePredictorMixin,
+        SerializableMixin
+    implements
+        DecisionTreeClassifier {
+
+  DecisionTreeClassifierImpl(
+      this.treeRootNode,
+      this.targetColumnName,
+      this.dtype,
+  );
+
+  factory DecisionTreeClassifierImpl.fromJson(Map<String, dynamic> json) =>
+      _$DecisionTreeClassifierImplFromJson(json);
 
   @override
+  Map<String, dynamic> toJson() => _$DecisionTreeClassifierImplToJson(this);
+
+  @override
+  @JsonKey(
+    name: dTypeJsonKey,
+    toJson: dTypeToJson,
+    fromJson: fromDTypeJson,
+  )
   final DType dtype;
 
-  final TreeSolver _solver;
+  @JsonKey(name: targetColumnNameJsonKey)
+  final String targetColumnName;
 
-  @override
-  final List<String> classNames;
+  @JsonKey(
+    name: treeRootNodeJsonKey,
+    toJson: treeNodeToJson,
+    fromJson: fromTreeNodeJson,
+  )
+  final TreeNode treeRootNode;
 
   @override
   DataFrame predict(DataFrame features) {
     final predictedLabels = features
         .toMatrix(dtype)
         .rows
-        .map(_solver.getLabelForSample);
+        .map((sample) => _getLabelForSample(sample, treeRootNode));
 
     if (predictedLabels.isEmpty) {
       return DataFrame([<num>[]]);
@@ -38,27 +72,51 @@ class DecisionTreeClassifierImpl with AssessablePredictorMixin
 
     return DataFrame.fromMatrix(
       Matrix.fromColumns([outcomeVector], dtype: dtype),
-      header: classNames,
+      header: [
+        targetColumnName,
+      ],
     );
   }
 
   @override
   DataFrame predictProbabilities(DataFrame features) {
-    final probabilities = Matrix.fromColumns([
-      Vector.fromList(
-        features
-            .toMatrix(dtype)
-            .rows
-            .map(_solver.getLabelForSample)
-            .map((label) => label.probability)
-            .toList(growable: false),
-        dtype: dtype,
-      ),
+    final sampleVectors = features
+        .toMatrix(dtype)
+        .rows;
+
+    final probabilities = sampleVectors
+        .map((sample) => _getLabelForSample(sample, treeRootNode))
+        .map((label) => label.probability)
+        .toList(growable: false);
+
+    final probabilitiesVector = Vector.fromList(
+      probabilities,
+      dtype: dtype,
+    );
+
+    final probabilitiesMatrixColumn = Matrix.fromColumns([
+      probabilitiesVector,
     ], dtype: dtype);
 
     return DataFrame.fromMatrix(
-      probabilities,
-      header: classNames,
+      probabilitiesMatrixColumn,
+      header: [
+        targetColumnName,
+      ],
     );
   }
+
+  TreeLeafLabel _getLabelForSample(Vector sample, TreeNode node) {
+    if (node.isLeaf) {
+      return node.label;
+    }
+
+    for (final childNode in node.children) {
+      if (childNode.isSamplePassed(sample)) {
+        return _getLabelForSample(sample, childNode);
+      }
+    }
+
+    throw Exception('Given sample does not conform any splitting condition');
+  }
 }

lib/src/classifier/decision_tree_classifier/decision_tree_json_keys.dart

@@ -0,0 +1,3 @@
+const dTypeJsonKey = 'DT';
+const targetColumnNameJsonKey = 'T';
+const treeRootNodeJsonKey = 'R';

lib/src/classifier/knn_classifier/knn_classifier_impl.dart

@@ -12,17 +12,16 @@ import 'package:ml_linalg/vector.dart';
 
 class KnnClassifierImpl with AssessablePredictorMixin implements KnnClassifier {
   KnnClassifierImpl(
-      String targetName,
+      this._targetColumnName,
       this._classLabels,
       this._kernel,
       this._solver,
       this.dtype,
-  ) : classNames = [targetName] {
+  ) {
     validateClassLabelList(_classLabels);
   }
 
-  @override
-  final List<String> classNames;
+  final String _targetColumnName;
 
   @override
   final DType dtype;
@@ -47,7 +46,7 @@ class KnnClassifierImpl with AssessablePredictorMixin implements KnnClassifier {
 
     return DataFrame.fromMatrix(
       Matrix.fromColumns([outcomesAsVector], dtype: dtype),
-      header: classNames,
+      header: [_targetColumnName],
     );
   }
 

lib/src/classifier/linear_classifier.dart

@@ -4,6 +4,8 @@ import 'package:ml_linalg/matrix.dart';
 
 /// An interface for all types of linear classifiers
 abstract class LinearClassifier implements Classifier {
+  Iterable<String> get classNames;
+
   /// A function that is used for converting learned coefficients into
   /// probabilities
   LinkFunction get linkFunction;
@@ -19,4 +21,4 @@ abstract class LinearClassifier implements Classifier {
   /// A matrix, where each column is a vector of coefficients, associated with
   /// the specific class
   Matrix get coefficientsByClasses;
-}
+}