fix: skip experiments pylint

experiments/seathru/dbscan_seathru.py

@@ -1,15 +1,21 @@
-#%%
+# pylint: skip-file
+
+
+import itertools as it
+
+# %%
 from pathlib import Path
-import matplotlib.pyplot as plt
+from typing import Tuple
+
 import cv2
+import matplotlib.pyplot as plt
 import numpy as np
+import skimage
 from PIL import Image
-from sklearn.cluster import KMeans, DBSCAN
 from scipy.optimize import minimize
-from typing import Tuple
-import itertools as it
 from skimage.util import img_as_ubyte
-import skimage
+from sklearn.cluster import DBSCAN, KMeans
+
 # %%
 # name = "T_S04856"
 name = "T_S04923"
@@ -18,12 +24,17 @@
 tif_file = Path(f"../../data/seathru/D3/D3/depth/depth{name}.tif")
 # %%
 clusters = 10
+
+
 # %%
 def uint8_2_double(array: np.ndarray):
     return array.astype(np.float64) / 255.0
 
+
 def double_2_uint8(array: np.ndarray):
     return (array * 255).astype(np.uint8)
+
+
 # %%
 def imshow(img: np.ndarray, color_channel="rgb"):
     if img.dtype == np.float64 and len(img.shape) == 3:
@@ -34,10 +45,12 @@ def imshow(img: np.ndarray, color_channel="rgb"):
     elif color_channel == "hsv":
         img = cv2.cvtColor(img, cv2.COLOR_HSV2RGB)
 
-    #plt.imshow(img)
+    # plt.imshow(img)
 
-    #if len(img.shape) == 2:
+    # if len(img.shape) == 2:
     #    plt.colorbar()
+
+
 # %%
 png = uint8_2_double(cv2.imread(png_file))
 height, width, _ = png.shape
@@ -48,7 +61,7 @@ def imshow(img: np.ndarray, color_channel="rgb"):
 # %%
 tif_image = Image.open(tif_file)
 tif = np.array(tif_image)
-tif = cv2.resize(tif, (width, height),  interpolation = cv2.INTER_NEAREST)
+tif = cv2.resize(tif, (width, height), interpolation=cv2.INTER_NEAREST)
 
 imshow(tif)
 # %%
@@ -60,10 +73,12 @@ def imshow(img: np.ndarray, color_channel="rgb"):
 
 tif_flat.shape
 # %%
-kmeans = KMeans(n_clusters=clusters, random_state=0, n_init="auto").fit(tif_flat[tif_flat != 0].reshape(-1, 1))
+kmeans = KMeans(n_clusters=clusters, random_state=0, n_init="auto").fit(
+    tif_flat[tif_flat != 0].reshape(-1, 1)
+)
 labels = np.zeros(tif_flat.shape, dtype=np.uint8)
 labels[tif_flat != 0] = kmeans.labels_
-labels[tif_flat == 0] = 255 # Use 255 to represent nan
+labels[tif_flat == 0] = 255  # Use 255 to represent nan
 means = np.array(kmeans.cluster_centers_).flatten()
 means_args = np.argsort(means)
 
@@ -92,7 +107,9 @@ def imshow(img: np.ndarray, color_channel="rgb"):
     g_filtered = g[mask].flatten()
     r_filtered = r[mask].flatten()
 
-    pixels = np.array([[b,g,r] for b,g,r in zip(b_filtered, g_filtered , r_filtered)])
+    pixels = np.array(
+        [[b, g, r] for b, g, r in zip(b_filtered, g_filtered, r_filtered)]
+    )
     idx = np.nonzero(np.all(pixels <= np.percentile(pixels, 1, axis=0), axis=1))
     selected_pixels = pixels[idx]
 
@@ -110,45 +127,81 @@ def imshow(img: np.ndarray, color_channel="rgb"):
 tif.min(), tif.max()
 # %%
 z_values
+
+
 # %%
-def estimate_backscatter(B_inf: float, beta_B: float, J_prime: float, beta_D_prime: float, z):
-    return B_inf * (1 - np.exp(- beta_B * z)) + (J_prime * np.exp(- beta_D_prime * z))
+def estimate_backscatter(
+    B_inf: float, beta_B: float, J_prime: float, beta_D_prime: float, z
+):
+    return B_inf * (1 - np.exp(-beta_B * z)) + (J_prime * np.exp(-beta_D_prime * z))
+
+
 # %%
-def optimize_estimate_backscatter(arguments: Tuple[float, float, float, float], B_hat: np.ndarray, z: np.ndarray):
+def optimize_estimate_backscatter(
+    arguments: Tuple[float, float, float, float], B_hat: np.ndarray, z: np.ndarray
+):
     B_inf, beta_B, J_prime, beta_D_prime = arguments
 
-    return np.sum((B_hat - estimate_backscatter(B_inf, beta_B, J_prime, beta_D_prime, z)) ** 2)
+    return np.sum(
+        (B_hat - estimate_backscatter(B_inf, beta_B, J_prime, beta_D_prime, z)) ** 2
+    )
+
+
 # %%
-result = minimize(optimize_estimate_backscatter, [1, 1, 1, 1], args=(dark_pixels[:, 0], z_values), bounds=[(0,1), (0,5), (0,1), (0,5)], method="SLSQP")
+result = minimize(
+    optimize_estimate_backscatter,
+    [1, 1, 1, 1],
+    args=(dark_pixels[:, 0], z_values),
+    bounds=[(0, 1), (0, 5), (0, 1), (0, 5)],
+    method="SLSQP",
+)
 b_args = result.x
 
 b_args
 # %%
 optimize_estimate_backscatter(b_args, dark_pixels[:, 0], z_values)
 # %%
-result = minimize(optimize_estimate_backscatter, [1, 1, 1, 1], args=(dark_pixels[:, 1], z_values), bounds=[(0,1), (0,5), (0,1), (0,5)], method="SLSQP")
+result = minimize(
+    optimize_estimate_backscatter,
+    [1, 1, 1, 1],
+    args=(dark_pixels[:, 1], z_values),
+    bounds=[(0, 1), (0, 5), (0, 1), (0, 5)],
+    method="SLSQP",
+)
 g_args = result.x
 
 g_args
 # %%
 optimize_estimate_backscatter(g_args, dark_pixels[:, 1], z_values)
 # %%
-result = minimize(optimize_estimate_backscatter, [1, 1, 1, 1], args=(dark_pixels[:, 2], z_values), bounds=[(0,1), (0,5), (0,1), (0,5)], method="SLSQP")
+result = minimize(
+    optimize_estimate_backscatter,
+    [1, 1, 1, 1],
+    args=(dark_pixels[:, 2], z_values),
+    bounds=[(0, 1), (0, 5), (0, 1), (0, 5)],
+    method="SLSQP",
+)
 r_args = result.x
 
 r_args
 # %%
 optimize_estimate_backscatter(r_args, dark_pixels[:, 2], z_values)
 # %%
-B_b = estimate_backscatter(b_args[0], b_args[1], b_args[2], b_args[3], tif_flat).reshape((height, width))
+B_b = estimate_backscatter(
+    b_args[0], b_args[1], b_args[2], b_args[3], tif_flat
+).reshape((height, width))
 
 imshow(B_b)
 # %%
-B_g = estimate_backscatter(g_args[0], g_args[1], g_args[2], g_args[3], tif_flat).reshape((height, width))
+B_g = estimate_backscatter(
+    g_args[0], g_args[1], g_args[2], g_args[3], tif_flat
+).reshape((height, width))
 
 imshow(B_g)
 # %%
-B_r = estimate_backscatter(r_args[0], r_args[1], r_args[2], r_args[3], tif_flat).reshape((height, width))
+B_r = estimate_backscatter(
+    r_args[0], r_args[1], r_args[2], r_args[3], tif_flat
+).reshape((height, width))
 
 imshow(B_r)
 # %%
@@ -189,7 +242,12 @@ def optimize_estimate_backscatter(arguments: Tuple[float, float, float, float],
 
 epsilon
 # %%
-X = np.array([(int(y), int(x), tif[y, x]) for y, x in it.product(np.arange(height), np.arange(width))])
+X = np.array(
+    [
+        (int(y), int(x), tif[y, x])
+        for y, x in it.product(np.arange(height), np.arange(width))
+    ]
+)
 
 X
 # %%
@@ -200,17 +258,21 @@ def optimize_estimate_backscatter(arguments: Tuple[float, float, float, float],
     tif_rebuilt[int(X[i, 0]), int(X[i, 1])] = X[i, 2]
 
 imshow(tif_rebuilt)
+
+
 # %%
 def metric(a: np.ndarray, b: np.ndarray):
     if np.linalg.norm(a[:2] - b[:2]) > 1:
         return np.inf
-    
+
     return np.linalg.norm(a[2] - b[2])
+
+
 # %%
 print("Coffee Time")
 dbscan = DBSCAN(eps=epsilon, min_samples=2, n_jobs=-1, metric=metric).fit(X)
 
 dbscan.labels_
 # %%
 print("No more coffee")
-np.savez_compressed("dbscan", dbscan.labels_)
+np.savez_compressed("dbscan", dbscan.labels_)