alexgiving · alexgiving · May 8, 2023 · May 8, 2023 · May 8, 2023 · May 8, 2023
diff --git a/graph/generate_dimension_histogram.py b/graph/generate_dimension_histogram.py
@@ -1,10 +1,9 @@
 from pathlib import Path
 
 import matplotlib.pyplot as plt
-import numpy as np
 from tap import Tap
 
-from lib.experiment import get_covariance_matrix
+from lib.data import get_experiment_data
 from lib.points_generator import generate_mix_distribution
 
 
@@ -14,22 +13,15 @@ class ArgumentParser(Tap):
 
 def main():
     args = ArgumentParser().parse_args()
-    path = args.path
+    args.path.mkdir(exist_ok=True)
 
     dimension = 20
     n_points = 100
 
-    sigma_list = [1, 1]
-    prob = 0.5
-    mu_list = [np.array([x + [0] * (dimension-2)])
-               for x in [[-4, 0], [4, 0]]]
+    n_clusters, prob, mu_list, cov_matrices = get_experiment_data(experiment_id=1, dimension=dimension)
 
-
-    cov_matrices = [get_covariance_matrix(
-        sigma, dimension) for sigma in sigma_list]
-
-    for t in [0, 0.3, 0.7, 0.9]:
-        filename = path / f'2cluster_hist_t_{t}.png'
+    for t in [0.2, 0.4, 0.9]:
+        filename = args.path / f'{n_clusters}_cluster_hist_t_{t}.png'
         clusters, _, _ = generate_mix_distribution(
             probability=prob,
             mu_list=mu_list,

diff --git a/graph/plot_minkowski.py b/graph/plot_minkowski.py
@@ -0,0 +1,73 @@
+from pathlib import Path
+from typing import Union
+
+import matplotlib.pyplot as plt
+import numpy as np
+from tap import Tap
+
+from lib.data import get_experiment_data
+from lib.kmeans import assign_to_cluster
+from lib.minkowski import minkowski_distance, pairwise_minkowski_distance
+from lib.points_generator import generate_mix_distribution
+
+
+class ArgumentParser(Tap):
+    path: Path = Path('images/')
+    '''Path to save image'''
+
+    p: float | int = 2
+    '''Minkowski parameter'''
+
+    t: float = 0.
+    '''Parameter of data distribution'''
+
+    @staticmethod
+    def to_number(string: str) -> Union[float, int]:
+        return float(string) if '.' in string else int(string)
+
+    def configure(self):
+        self.add_argument('-p', type=self.to_number)
+
+
+# pylint: disable=too-many-locals
+def main():
+    args = ArgumentParser().parse_args()
+    args.path.mkdir(exist_ok=True)
+    p = args.p
+
+    dimension = 20
+    n_points = 10
+
+    n_clusters, prob, mu_list, cov_matrices = get_experiment_data(experiment_id=1, dimension=dimension)
+
+    filename = args.path / f'plot_minkowski_function_with_p_{p}.png'
+    samples, _, centroids = generate_mix_distribution(
+        probability=prob,
+        mu_list=mu_list,
+        cov_matrices=cov_matrices,
+        n_samples=n_points,
+        t=0.1
+    )
+
+    dim = 0
+
+    clusters, _ = assign_to_cluster(samples, centroids, n_clusters, p)
+    cluster = np.array(clusters[0])
+    dimension_data = cluster[:,dim]
+
+    points = np.linspace(min(dimension_data), max(dimension_data), 100)
+    minkowski_values = pairwise_minkowski_distance(
+        point_a = dimension_data,
+        points=points,
+        p=p
+    )
+
+    fig, ax = plt.subplots(figsize=(5, 3))
+    ax.scatter(points, minkowski_values)
+    ax.scatter(centroids[0][dim], minkowski_distance(centroids[0][dim], dimension_data, p))
+    fig.savefig(str(filename), dpi=300, bbox_inches='tight')
+    plt.close(fig)
+
+
+if __name__ == '__main__':
+    main()
diff --git a/lib/data.py b/lib/data.py
@@ -0,0 +1,32 @@
+import numpy as np
+
+from lib.experiment import get_covariance_matrix
+
+
+def get_experiment_data(experiment_id: int, dimension: int) -> tuple[int, float, list[np.ndarray], list[np.ndarray]]:
+
+    n_clusters: int = 0
+    prob: float = 0.
+    mu_prefix: list[list[float | int]] = [[]]
+    sigma_list: list[float | int] = []
+
+    if experiment_id == 1:
+        print('Experiment with 2 clusters')
+        n_clusters = 2
+        sigma_list = [1, 1]
+        prob = 0.5
+        mu_prefix = [[-4, 0], [4, 0]]
+
+    elif experiment_id == 2:
+        print('Experiment with 3 clusters')
+        n_clusters = 3
+        sigma_list = [1, 1, 1]
+        prob = 1/3
+        mu_prefix = [[4, 0, 0], [0, 4, 0], [0, 0, 4]]
+
+    else:
+        KeyError(f'Not supported experiment type: {experiment_id}')
+
+    mu_list = [np.array([x + [0] * (dimension - len(x))]) for x in mu_prefix]
+    cov_matrix = [get_covariance_matrix(sigma, dimension) for sigma in sigma_list]
+    return n_clusters, prob, mu_list, cov_matrix
diff --git a/main.py b/main.py
@@ -1,9 +1,9 @@
 from pathlib import Path
 
-import numpy as np
 from tap import Tap
 
-from lib.experiment import get_covariance_matrix, run_experiment
+from lib.data import get_experiment_data
+from lib.experiment import run_experiment
 
 
 class ArgumentParser(Tap):
@@ -19,35 +19,6 @@ class ArgumentParser(Tap):
     '''
 
 
-def _get_experiment_data(experiment_id: int, dimension: int) -> tuple[int, float, list[np.ndarray], list[np.ndarray]]:
-
-    n_clusters: int = 0
-    prob: float = 0.
-    mu_prefix: list[list[float | int]] = [[]]
-    sigma_list: list[float | int] = []
-
-    if experiment_id == 1:
-        print('Experiment with 2 clusters')
-        n_clusters = 2
-        sigma_list = [1, 1]
-        prob = 0.5
-        mu_prefix = [[-4, 0], [4, 0]]
-
-    elif experiment_id == 2:
-        print('Experiment with 3 clusters')
-        n_clusters = 3
-        sigma_list = [1, 1, 1]
-        prob = 1/3
-        mu_prefix = [[4, 0, 0], [0, 4, 0], [0, 0, 4]]
-
-    else:
-        KeyError(f'Not supported experiment type: {experiment_id}')
-
-    mu_list = [np.array([x + [0] * (dimension - len(x))]) for x in mu_prefix]
-    cov_matrix = [get_covariance_matrix(sigma, dimension) for sigma in sigma_list]
-    return n_clusters, prob, mu_list, cov_matrix
-
-
 def main():
     args = ArgumentParser().parse_args()
     experiments_path = args.path
@@ -59,7 +30,7 @@ def main():
 
     dimension = 20
     T_parameter = [0.4, 1]
-    n_clusters, prob, mu_list, cov_matrices = _get_experiment_data(
+    n_clusters, prob, mu_list, cov_matrices = get_experiment_data(
         experiment_id=args.experiment_id, dimension=dimension)
 
     for points in n_points: