testing without objects

Took 1 hour 19 minutes

sklarpy/tests/multivariate/conftest.py

@@ -1,8 +1,8 @@
 # Contains pytest fixtures for testing SklarPy multivariate code
 import pandas as pd
+import numpy as np
 import pytest
 import scipy.stats
-import numpy as np
 
 from sklarpy.multivariate import distributions_map
 
@@ -41,3 +41,164 @@ def pd_mvt_discrete_data():
 @pytest.fixture(scope="session", autouse=True)
 def mv_dists_to_test():
     return distributions_map['all']
+
+
+@pytest.fixture(scope="session", autouse=True)
+def params_2d():
+    return {
+        'multivariate_normal': (
+            np.array([[-0.09739442], [0.06748115]]),
+            np.array([[1.22604687e+00, 9.60076079e-18],
+                      [9.60076079e-18, 7.30245004e-01]])),
+
+        'multivariate_student_t': (
+            100.0,
+            np.array([[-0.09739442], [0.06748115]]),
+            np.array([[1.20152593, 0.19624622],
+                      [0.19624622, 0.7156401]])),
+
+        'multivariate_gen_hyperbolic': (
+            -1.1788364839820724, 10.0, 10.0,
+            np.array([[0.84415713], [-0.45785498]]),
+            np.array([[1.2253664, 0.26909617],
+                      [0.26909617, 0.75701993]]),
+            np.array([[-1.00453046], [0.56047505]])),
+
+        'multivariate_marginal_hyperbolic': (
+            6.430016040119274, 10.0,
+            np.array([[0.69052415], [-0.39804144]]),
+            np.array([[1.22869356, 0.2616279],
+                      [0.2616279, 0.75174415]]),
+            np.array([[-0.82414258], [0.48692467]])),
+
+        'multivariate_hyperbolic': (
+            5.584106488470228, 10.0,
+            np.array([[0.65995643], [-0.3874908]]),
+            np.array([[1.2296575, 0.26089854],
+                      [0.26089854, 0.75084492]]),
+            np.array([[-0.78967095], [0.47438798]])),
+
+        'multivariate_nig': (
+            8.986208938433768, 10.0,
+            np.array([[0.78169119], [-0.43129769]]),
+            np.array([[1.22628047, 0.26411374],
+                      [0.26411374, 0.7542892]]),
+            np.array([[-0.92734838], [0.52616235]])),
+
+        'multivariate_skewed_t': (
+            10.0,
+            np.array([[0.15711164], [-0.10459758]]),
+            np.array([[1.22892318, 0.21921408],
+                      [0.21921408, 0.7350084]]),
+            np.array([[-0.20360485], [0.13766299]])),
+
+        'multivariate_sym_gen_hyperbolic': (
+            -0.4268651272656109, 10.0, 10.0,
+            np.array([[-0.09739442], [0.06748115]]),
+            np.array([[1.22604687, 0.20025125],
+                      [0.20025125, 0.730245]])),
+
+        'multivariate_sym_marginal_hyperbolic': (
+            7.523270258556909, 10.0,
+            np.array([[-0.09739442], [0.06748115]]),
+            np.array([[1.22604687, 0.20025125],
+                      [0.20025125, 0.730245]])),
+
+        'multivariate_sym_hyperbolic': (
+            6.610028122530992, 10.0,
+            np.array([[-0.09739442], [0.06748115]]),
+            np.array([[1.22604687, 0.20025125],
+                      [0.20025125, 0.730245]])),
+
+        'multivariate_sym_nig': (
+            10.0, 10.0,
+            np.array([[-0.09739442], [0.06748115]]),
+            np.array([[1.22604687, 0.20025125],
+                      [0.20025125, 0.730245]]))
+    }
+
+
+@pytest.fixture(scope="session", autouse=True)
+def params_3d():
+    return {
+        'multivariate_normal': (
+            np.array([[-0.06068104], [-0.10572418], [0.03489784]]),
+            np.array([[1.21898949e+00, 9.34403992e-17, 8.06538495e-17],
+                      [3.91669234e-18, 9.03438954e-01, 7.39590407e-19],
+                      [5.18678009e-18, 3.39803556e-17, 1.05766110e+00]])),
+
+        'multivariate_student_t': (
+            100.0,
+            np.array([[-0.06068104], [-0.10572418], [0.03489784]]),
+            np.array([[1.1946097, 0.09775865, -0.00776843],
+                      [0.09775865, 0.88537018, 0.08621716],
+                      [-0.00776843,  0.08621716, 1.03650788]])),
+
+        'multivariate_gen_hyperbolic': (
+            -1.782382671254414, 10.0, 10.0,
+            np.array([[-0.62581127], [-1.87660553], [0.13564066]]),
+            np.array([[ 1.38090727e+00, -2.01241655e-04, -2.60836067e-03],
+                      [-2.01241655e-04,  6.86684932e-01,  1.23061470e-01],
+                      [-2.60836067e-03, 1.23061470e-01, 1.22915669e+00]]),
+            np.array([[0.63853468], [2.00090016], [-0.11382825]])),
+
+        'multivariate_marginal_hyperbolic': (
+            6.205679755772674, 10.0,
+            np.array([[-0.4760037], [-1.50649459], [0.0978438 ]]),
+            np.array([[1.33664701, 0.03084231, -0.0052302],
+                      [0.03084231, 0.73684715, 0.1103901],
+                      [-0.0052302, 0.1103901, 1.17990614]]),
+            np.array([[0.44089418], [1.4870162], [-0.06682156]])),
+
+        'multivariate_hyperbolic': (
+            4.669931146010268, 10.0,
+            np.array([[-0.43035465], [-1.40305972], [0.08801923]]),
+            np.array([[1.32964083, 0.03793152, -0.00580953],
+                      [0.03793152, 0.74609703, 0.10788308],
+                      [-0.00580953, 0.10788308, 1.17118297]]),
+            np.array([[0.38622221], [1.35541132], [-0.05549939]])),
+
+        'multivariate_nig': (
+            8.517549891625043, 10.0,
+            np.array([[-0.54136409], [-1.66194387], [0.11216012]]),
+            np.array([[1.34739657, 0.02110339, -0.00440445],
+                      [0.02110339, 0.72340899, 0.11398558],
+                      [-0.00440445, 0.11398558, 1.19285811]]),
+            np.array([[0.52083656], [1.68621739], [-0.08371633]])),
+
+        'multivariate_skewed_t': (
+            10.0,
+            np.array([[-0.1552794], [-0.58786565], [0.03910667]]),
+            np.array([[ 1.25348961, 0.0871567, -0.00803451],
+                      [0.0871567, 0.8514119, 0.09138582],
+                      [-0.00803451, 0.09138582, 1.09025713]]),
+            np.array([[0.07567869], [0.38571318], [-0.00336706]])),
+
+        'multivariate_sym_gen_hyperbolic': (
+            -0.33835936636262204, 10.0, 10.0,
+            np.array([[-0.06068104], [-0.10572418], [0.03489784]]),
+            np.array([[1.21898949, 0.09975372, -0.00792697],
+                      [0.09975372, 0.90343895, 0.08797669],
+                      [-0.00792697, 0.08797669, 1.0576611]])),
+
+        'multivariate_sym_marginal_hyperbolic': (
+            7.799728823974279, 10.0,
+            np.array([[-0.06068104], [-0.10572418], [0.03489784]]),
+            np.array([[1.21898949, 0.09975372, -0.00792697],
+                      [0.09975372, 0.90343895, 0.08797669],
+                      [-0.00792697, 0.08797669, 1.0576611]])),
+
+        'multivariate_sym_hyperbolic': (
+            5.992259894031568, 10.0,
+            np.array([[-0.06068104], [-0.10572418], [0.03489784]]),
+            np.array([[1.21898949, 0.09975372, -0.00792697],
+                      [0.09975372, 0.90343895, 0.08797669],
+                      [-0.00792697, 0.08797669, 1.0576611]])),
+
+        'multivariate_sym_nig': (
+            10.0, 9.96376429525435,
+            np.array([[-0.06068104], [-0.10572418], [0.03489784]]),
+            np.array([[1.21898949, 0.09975372, -0.00792697],
+                      [0.09975372, 0.90343895, 0.08797669],
+                      [-0.00792697, 0.08797669, 1.0576611]]))
+    }

sklarpy/tests/multivariate/helpers.py

@@ -1,28 +1,35 @@
 # Contains helper functions for testing SklarPy multivariate code
 import os
+import numpy as np
 
 from sklarpy.multivariate import *
-from sklarpy import load
 
 __all__ = ['get_dist']
 
 
-def get_dist(name: str, d: int = 3) -> tuple:
+def get_dist(name: str, params_dict: dict, data: np.ndarray = None) -> tuple:
     """Fits a distribution to data.
 
     Parameters
     ----------
     name : str
         The name of the distribution as in distributions_map
-    d: int
-        The dimension of the data.
+    params_dict: dict
+        Dictionary containing the parameters of each distribution.
+    data: np.ndarray
+        Multivariate data to fit the distribution too if its params are not
+        specified in params_dict.
+
     Returns
     -------
     res: tuple
         non-fitted dist, fitted dist, parameters for fitted dist
     """
     dist = eval(name)
-    path: str = f'{os.getcwd()}\\sklarpy\\tests\\multivariate\\{name}_{d}D'
-    params = load(path)
-    fitted = dist.fit(params=params)
+    if name in params_dict:
+        params: tuple = params_dict[name]
+        fitted = dist.fit(params=params)
+    else:
+        fitted = dist.fit(data=data)
+        params: tuple = fitted.params.to_tuple
     return dist, fitted, params

sklarpy/tests/multivariate/test_fitted_dists.py

@@ -4,22 +4,22 @@
 import pytest
 from typing import Callable
 import matplotlib.pyplot as plt
+import scipy.stats
 
-from sklarpy.multivariate import *
 from sklarpy.tests.multivariate.helpers import get_dist
 from sklarpy._utils import Params
 
 
 def test_fitted_logpdf_pdf_cdf_mc_cdf(
         mvt_continuous_data, mvt_discrete_data, pd_mvt_continuous_data,
-        pd_mvt_discrete_data, mv_dists_to_test):
+        pd_mvt_discrete_data, mv_dists_to_test, params_2d):
     """Testing the logpdf, pdf, cdf and mc-cdf functions of fitted multivariate
      distributions"""
     eps: float = 10 ** -5
     num_generate: int = 10
 
     for name in mv_dists_to_test:
-        _, fitted, _ = get_dist(name, 2)
+        _, fitted, _ = get_dist(name, params_2d, mvt_continuous_data)
         for func_str in ('logpdf', 'pdf', 'cdf', 'mc_cdf'):
             func: Callable = eval(f'fitted.{func_str}')
             cdf_num: int = 10
@@ -67,10 +67,10 @@ def test_fitted_logpdf_pdf_cdf_mc_cdf(
                 func(x=new_dataset, num_generate=num_generate)
 
 
-def test_fitted_rvs(mv_dists_to_test):
+def test_fitted_rvs(mv_dists_to_test, params_2d, mvt_continuous_data):
     """Testing the rvs functions of fitted multivariate distributions."""
     for name in mv_dists_to_test:
-        _, fitted, _ = get_dist(name)
+        _, fitted, _ = get_dist(name, params_2d, mvt_continuous_data)
         for size in (1, 2, 5, 101):
             rvs = fitted.rvs(size=size)
             np_rvs: np.ndarray = np.asarray(rvs)
@@ -87,11 +87,11 @@ def test_fitted_rvs(mv_dists_to_test):
 
 def test_fitted_scalars(mvt_continuous_data, mvt_discrete_data,
                         pd_mvt_continuous_data, pd_mvt_discrete_data,
-                        mv_dists_to_test):
+                        mv_dists_to_test, params_2d):
     """Testing the likelihood, loglikelihood, AIC and BIC functions of
     multivariate fitted distributions."""
     for name in mv_dists_to_test:
-        _, fitted, _ = get_dist(name, 2)
+        _, fitted, _ = get_dist(name, params_2d, mvt_continuous_data)
         for func_str in ('likelihood', 'loglikelihood', 'aic', 'bic'):
             func: Callable = eval(f'fitted.{func_str}')
             for data in (mvt_continuous_data, mvt_discrete_data,
@@ -124,12 +124,14 @@ def test_fitted_scalars(mvt_continuous_data, mvt_discrete_data,
                 func(data=new_dataset)
 
 
-def test_fitted_plots(mv_dists_to_test):
+def test_fitted_plots(mv_dists_to_test, params_2d, params_3d, mvt_continuous_data):
     """Testing the marginal_pairplot, pdf_plot, cdf_plot and mc_cdf_plot
     methods of fitted multivariate distributions."""
+    mvt_continuous_data_3d: np.ndarray = scipy.stats.multivariate_normal.rvs(
+        size=(mvt_continuous_data.shape[0], 3))
     for name in mv_dists_to_test:
-        _, fitted_2d, _ = get_dist(name, 2)
-        _, fitted_3d, _ = get_dist(name, 3)
+        _, fitted_2d, _ = get_dist(name, params_2d, mvt_continuous_data)
+        _, fitted_3d, _ = get_dist(name, params_3d, mvt_continuous_data_3d)
         for func_str in ('marginal_pairplot', 'pdf_plot',
                          'cdf_plot', 'mc_cdf_plot'):
             func_2d: Callable = eval(f'fitted_2d.{func_str}')
@@ -153,10 +155,10 @@ def test_fitted_plots(mv_dists_to_test):
             plt.close()
 
 
-def test_fitted_params(mv_dists_to_test):
+def test_fitted_params(mv_dists_to_test, params_2d, mvt_continuous_data):
     """Testing the params attribute of fitted multivariate distributions."""
     for name in mv_dists_to_test:
-        _, fitted, _ = get_dist(name, 2)
+        _, fitted, _ = get_dist(name, params_2d, mvt_continuous_data)
 
         # checking exists
         assert 'params' in dir(fitted), f"params is not an attribute of " \
@@ -167,11 +169,11 @@ def test_fitted_params(mv_dists_to_test):
             f"params of fitted {name} is not a Params object."
 
 
-def test_fitted_integers(mv_dists_to_test):
+def test_fitted_integers(mv_dists_to_test, params_2d, mvt_continuous_data):
     """Testing the num_params, num_scalar_params, num_variables and
     fitted_num_data_points of fitted multivariate distributions."""
     for name in mv_dists_to_test:
-        _, fitted, params = get_dist(name)
+        _, fitted, params = get_dist(name, params_2d, mvt_continuous_data)
 
         for func_str in ('num_params', 'num_scalar_params', 'num_variables',
                          'fitted_num_data_points'):
@@ -188,10 +190,10 @@ def test_fitted_integers(mv_dists_to_test):
             f"object."
 
 
-def test_fitted_converged(mv_dists_to_test):
+def test_fitted_converged(mv_dists_to_test, params_2d, mvt_continuous_data):
     """Testing converged attributes of fitted multivariate distributions."""
     for name in mv_dists_to_test:
-        _, fitted, _ = get_dist(name)
+        _, fitted, _ = get_dist(name, params_2d, mvt_continuous_data)
 
         # testing exists
         assert 'converged' in dir(fitted), \
@@ -202,10 +204,10 @@ def test_fitted_converged(mv_dists_to_test):
             f"converged is not a boolean value for fitted {name}."
 
 
-def test_fitted_summaries(mv_dists_to_test):
+def test_fitted_summaries(mv_dists_to_test, params_2d, mvt_continuous_data):
     """Testing the summaries of fitted multivariate distributions."""
     for name in mv_dists_to_test:
-        _, fitted, _ = get_dist(name)
+        _, fitted, _ = get_dist(name, params_2d, mvt_continuous_data)
 
         # testing exists
         assert 'summary' in dir(fitted), \