make default params available to serde (#167)

* add seed param to search params

* make default params available to serde

* lints

* create defaults for enums

* lint

Author: montanalow

src/algorithm/neighbour/mod.rs

@@ -59,6 +59,12 @@ pub enum KNNAlgorithmName {
     CoverTree,
 }
 
+impl Default for KNNAlgorithmName {
+    fn default() -> Self {
+        KNNAlgorithmName::CoverTree
+    }
+}
+
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug)]
 pub(crate) enum KNNAlgorithm<T: RealNumber, D: Distance<Vec<T>, T>> {

src/cluster/dbscan.rs

@@ -65,17 +65,22 @@ pub struct DBSCAN<T: RealNumber, D: Distance<Vec<T>, T>> {
     eps: T,
 }
 
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 /// DBSCAN clustering algorithm parameters
 pub struct DBSCANParameters<T: RealNumber, D: Distance<Vec<T>, T>> {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// a function that defines a distance between each pair of point in training data.
     /// This function should extend [`Distance`](../../math/distance/trait.Distance.html) trait.
     /// See [`Distances`](../../math/distance/struct.Distances.html) for a list of available functions.
     pub distance: D,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The number of samples (or total weight) in a neighborhood for a point to be considered as a core point.
     pub min_samples: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The maximum distance between two samples for one to be considered as in the neighborhood of the other.
     pub eps: T,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// KNN algorithm to use.
     pub algorithm: KNNAlgorithmName,
 }
@@ -113,14 +118,18 @@ impl<T: RealNumber, D: Distance<Vec<T>, T>> DBSCANParameters<T, D> {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct DBSCANSearchParameters<T: RealNumber, D: Distance<Vec<T>, T>> {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// a function that defines a distance between each pair of point in training data.
     /// This function should extend [`Distance`](../../math/distance/trait.Distance.html) trait.
     /// See [`Distances`](../../math/distance/struct.Distances.html) for a list of available functions.
     pub distance: Vec<D>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The number of samples (or total weight) in a neighborhood for a point to be considered as a core point.
     pub min_samples: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The maximum distance between two samples for one to be considered as in the neighborhood of the other.
     pub eps: Vec<T>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// KNN algorithm to use.
     pub algorithm: Vec<KNNAlgorithmName>,
 }
@@ -221,7 +230,7 @@ impl<T: RealNumber> Default for DBSCANParameters<T, Euclidian> {
             distance: Distances::euclidian(),
             min_samples: 5,
             eps: T::half(),
-            algorithm: KNNAlgorithmName::CoverTree,
+            algorithm: KNNAlgorithmName::default(),
         }
     }
 }

src/cluster/kmeans.rs

@@ -102,13 +102,17 @@ impl<T: RealNumber> PartialEq for KMeans<T> {
     }
 }
 
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 /// K-Means clustering algorithm parameters
 pub struct KMeansParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of clusters.
     pub k: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Maximum number of iterations of the k-means algorithm for a single run.
     pub max_iter: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Determines random number generation for centroid initialization.
     /// Use an int to make the randomness deterministic
     pub seed: Option<u64>,
@@ -141,10 +145,13 @@ impl Default for KMeansParameters {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct KMeansSearchParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of clusters.
     pub k: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Maximum number of iterations of the k-means algorithm for a single run.
     pub max_iter: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Determines random number generation for centroid initialization.
     /// Use an int to make the randomness deterministic
     pub seed: Vec<Option<u64>>,

src/decomposition/pca.rs

@@ -83,11 +83,14 @@ impl<T: RealNumber, M: Matrix<T>> PartialEq for PCA<T, M> {
     }
 }
 
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 /// PCA parameters
 pub struct PCAParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of components to keep.
     pub n_components: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// By default, covariance matrix is used to compute principal components.
     /// Enable this flag if you want to use correlation matrix instead.
     pub use_correlation_matrix: bool,
@@ -120,8 +123,10 @@ impl Default for PCAParameters {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct PCASearchParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of components to keep.
     pub n_components: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// By default, covariance matrix is used to compute principal components.
     /// Enable this flag if you want to use correlation matrix instead.
     pub use_correlation_matrix: Vec<bool>,

src/decomposition/svd.rs

@@ -69,9 +69,11 @@ impl<T: RealNumber, M: Matrix<T>> PartialEq for SVD<T, M> {
     }
 }
 
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 /// SVD parameters
 pub struct SVDParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of components to keep.
     pub n_components: usize,
 }
@@ -94,6 +96,7 @@ impl SVDParameters {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct SVDSearchParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Maximum number of iterations of the k-means algorithm for a single run.
     pub n_components: Vec<usize>,
 }

src/ensemble/random_forest_classifier.rs

@@ -67,20 +67,28 @@ use crate::tree::decision_tree_classifier::{
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct RandomForestClassifierParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Split criteria to use when building a tree. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub criterion: SplitCriterion,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Tree max depth. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub max_depth: Option<u16>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The minimum number of samples required to be at a leaf node. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub min_samples_leaf: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The minimum number of samples required to split an internal node. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub min_samples_split: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The number of trees in the forest.
     pub n_trees: u16,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of random sample of predictors to use as split candidates.
     pub m: Option<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Whether to keep samples used for tree generation. This is required for OOB prediction.
     pub keep_samples: bool,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Seed used for bootstrap sampling and feature selection for each tree.
     pub seed: u64,
 }
@@ -198,20 +206,28 @@ impl<T: RealNumber, M: Matrix<T>> Predictor<M, M::RowVector> for RandomForestCla
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct RandomForestClassifierSearchParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Split criteria to use when building a tree. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub criterion: Vec<SplitCriterion>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Tree max depth. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub max_depth: Vec<Option<u16>>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The minimum number of samples required to be at a leaf node. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub min_samples_leaf: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The minimum number of samples required to split an internal node. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub min_samples_split: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The number of trees in the forest.
     pub n_trees: Vec<u16>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of random sample of predictors to use as split candidates.
     pub m: Vec<Option<usize>>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Whether to keep samples used for tree generation. This is required for OOB prediction.
     pub keep_samples: Vec<bool>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Seed used for bootstrap sampling and feature selection for each tree.
     pub seed: Vec<u64>,
 }

src/ensemble/random_forest_regressor.rs

@@ -65,18 +65,25 @@ use crate::tree::decision_tree_regressor::{
 /// Parameters of the Random Forest Regressor
 /// Some parameters here are passed directly into base estimator.
 pub struct RandomForestRegressorParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Tree max depth. See [Decision Tree Regressor](../../tree/decision_tree_regressor/index.html)
     pub max_depth: Option<u16>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The minimum number of samples required to be at a leaf node. See [Decision Tree Regressor](../../tree/decision_tree_regressor/index.html)
     pub min_samples_leaf: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The minimum number of samples required to split an internal node. See [Decision Tree Regressor](../../tree/decision_tree_regressor/index.html)
     pub min_samples_split: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The number of trees in the forest.
     pub n_trees: usize,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of random sample of predictors to use as split candidates.
     pub m: Option<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Whether to keep samples used for tree generation. This is required for OOB prediction.
     pub keep_samples: bool,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Seed used for bootstrap sampling and feature selection for each tree.
     pub seed: u64,
 }
@@ -181,18 +188,25 @@ impl<T: RealNumber, M: Matrix<T>> Predictor<M, M::RowVector> for RandomForestReg
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct RandomForestRegressorSearchParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Tree max depth. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub max_depth: Vec<Option<u16>>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The minimum number of samples required to be at a leaf node. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub min_samples_leaf: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The minimum number of samples required to split an internal node. See [Decision Tree Classifier](../../tree/decision_tree_classifier/index.html)
     pub min_samples_split: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The number of trees in the forest.
     pub n_trees: Vec<usize>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Number of random sample of predictors to use as split candidates.
     pub m: Vec<Option<usize>>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Whether to keep samples used for tree generation. This is required for OOB prediction.
     pub keep_samples: Vec<bool>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Seed used for bootstrap sampling and feature selection for each tree.
     pub seed: Vec<u64>,
 }

src/linear/elastic_net.rs

@@ -71,16 +71,21 @@ use crate::linear::lasso_optimizer::InteriorPointOptimizer;
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct ElasticNetParameters<T: RealNumber> {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Regularization parameter.
     pub alpha: T,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The elastic net mixing parameter, with 0 <= l1_ratio <= 1.
     /// For l1_ratio = 0 the penalty is an L2 penalty.
     /// For l1_ratio = 1 it is an L1 penalty. For 0 < l1_ratio < 1, the penalty is a combination of L1 and L2.
     pub l1_ratio: T,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// If True, the regressors X will be normalized before regression by subtracting the mean and dividing by the standard deviation.
     pub normalize: bool,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The tolerance for the optimization
     pub tol: T,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The maximum number of iterations
     pub max_iter: usize,
 }
@@ -139,16 +144,21 @@ impl<T: RealNumber> Default for ElasticNetParameters<T> {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct ElasticNetSearchParameters<T: RealNumber> {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Regularization parameter.
     pub alpha: Vec<T>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The elastic net mixing parameter, with 0 <= l1_ratio <= 1.
     /// For l1_ratio = 0 the penalty is an L2 penalty.
     /// For l1_ratio = 1 it is an L1 penalty. For 0 < l1_ratio < 1, the penalty is a combination of L1 and L2.
     pub l1_ratio: Vec<T>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// If True, the regressors X will be normalized before regression by subtracting the mean and dividing by the standard deviation.
     pub normalize: Vec<bool>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The tolerance for the optimization
     pub tol: Vec<T>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The maximum number of iterations
     pub max_iter: Vec<usize>,
 }

src/linear/lasso.rs

@@ -38,13 +38,17 @@ use crate::math::num::RealNumber;
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct LassoParameters<T: RealNumber> {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Controls the strength of the penalty to the loss function.
     pub alpha: T,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// If true the regressors X will be normalized before regression
     /// by subtracting the mean and dividing by the standard deviation.
     pub normalize: bool,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The tolerance for the optimization
     pub tol: T,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The maximum number of iterations
     pub max_iter: usize,
 }
@@ -116,13 +120,17 @@ impl<T: RealNumber, M: Matrix<T>> Predictor<M, M::RowVector> for Lasso<T, M> {
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct LassoSearchParameters<T: RealNumber> {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Controls the strength of the penalty to the loss function.
     pub alpha: Vec<T>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// If true the regressors X will be normalized before regression
     /// by subtracting the mean and dividing by the standard deviation.
     pub normalize: Vec<bool>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The tolerance for the optimization
     pub tol: Vec<T>,
+    #[cfg_attr(feature = "serde", serde(default))]
     /// The maximum number of iterations
     pub max_iter: Vec<usize>,
 }

src/linear/linear_regression.rs

@@ -71,19 +71,21 @@ use crate::linalg::Matrix;
 use crate::math::num::RealNumber;
 
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
-#[derive(Debug, Clone, Eq, PartialEq)]
+#[derive(Debug, Default, Clone, Eq, PartialEq)]
 /// Approach to use for estimation of regression coefficients. QR is more efficient but SVD is more stable.
 pub enum LinearRegressionSolverName {
     /// QR decomposition, see [QR](../../linalg/qr/index.html)
     QR,
+    #[default]
     /// SVD decomposition, see [SVD](../../linalg/svd/index.html)
     SVD,
 }
 
 /// Linear Regression parameters
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
-#[derive(Debug, Clone)]
+#[derive(Debug, Default, Clone)]
 pub struct LinearRegressionParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Solver to use for estimation of regression coefficients.
     pub solver: LinearRegressionSolverName,
 }
@@ -105,18 +107,11 @@ impl LinearRegressionParameters {
     }
 }
 
-impl Default for LinearRegressionParameters {
-    fn default() -> Self {
-        LinearRegressionParameters {
-            solver: LinearRegressionSolverName::SVD,
-        }
-    }
-}
-
 /// Linear Regression grid search parameters
 #[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
 #[derive(Debug, Clone)]
 pub struct LinearRegressionSearchParameters {
+    #[cfg_attr(feature = "serde", serde(default))]
     /// Solver to use for estimation of regression coefficients.
     pub solver: Vec<LinearRegressionSolverName>,
 }
@@ -353,5 +348,9 @@ mod tests {
             serde_json::from_str(&serde_json::to_string(&lr).unwrap()).unwrap();
 
         assert_eq!(lr, deserialized_lr);
+
+        let default = LinearRegressionParameters::default();
+        let parameters: LinearRegressionParameters = serde_json::from_str("{}").unwrap();
+        assert_eq!(parameters.solver, default.solver);
     }
 }