rename module

readme.md

@@ -14,7 +14,7 @@
 
 ### Use:
 
-`python -m nd2shrink path_to_nd2`
+`python -m nd2tif path_to_nd2`
 
 # Bin 16x stitched nd2:
 

segment/.ipynb_checkpoints/segment-checkpoint.py

+import cv2
+import numpy as np
+from skimage.measure import regionprops, label
+from scipy.ndimage import gaussian_filter, binary_erosion
+import matplotlib.pyplot as plt
+import matplotlib
+import logging
+
+matplotlib.logging.basicConfig(level=logging.INFO)
+logging.basicConfig(level=logging.DEBUG)
+
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.INFO)
+
+
+def findSpheroid(
+    imCropped: np.ndarray,
+    sigma: float = 5,
+    erode: int = 3,
+    threshold: float = 0.5,
+    max_ecc: float = 0.95,
+    lim_major_axis_length: tuple = (100, 400),
+    plot: bool = False,
+):
+
+    """
+
+    We find the spheroid by thresholding the intensity
+    and area filling. Sph. must have a dark border around
+    it.
+
+    returns mask
+    """
+
+    result1, result2 = np.gradient(imCropped)
+    grad = np.sqrt(result1 ** 2 + result2 ** 2)
+    if plot:
+        f, (ax1, ax2, ax3, ax4) = plt.subplots(
+            1, 4, sharey=True, sharex=True, figsize=(10, 3)
+        )
+        ax1.imshow(imCropped, cmap="gray")
+        ax1.set_title("raw")
+        ax2.imshow(grad)
+        ax2.set_title("grad")
+
+    toThresh = gaussian_filter(grad, sigma=sigma)
+
+    imThresh = toThresh > np.max(toThresh) * threshold
+
+    cnts, h = cv2.findContours(
+        imThresh.astype("uint8"), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE
+    )
+    mask = cv2.drawContours(
+        imThresh.astype("uint8"),
+        cnts,
+        -1,
+        (255, 255, 255),
+        thickness=cv2.FILLED,
+    )
+    eroded_mask = binary_erosion(
+        mask, structure=np.ones((erode, erode))
+    ).astype(mask.dtype)
+
+    temp = eroded_mask
+
+    imLabel = label(temp)
+    logger.debug(f"found {imLabel.max()} regions")
+    if plot:
+        ax3.imshow(imLabel)
+        ax3.set_title("all regions")
+
+    for i, region in enumerate(regionprops(imLabel)):
+        logger.debug(f"filtering region {i}")
+
+        if region.eccentricity > max_ecc:
+            # check it is inside or outside
+            logger.debug(f"eccentricity {region.eccentricity} > {max_ecc}")
+
+            temp[imLabel == region.label] = 0
+            # region given same value as sph. border
+
+        if (l := region.major_axis_length) < (lim := lim_major_axis_length)[
+            0
+        ] or l > lim[1]:
+            logger.debug(f"major_axis_length {l} outside limits {lim}")
+            temp[imLabel == region.label] = 0
+
+    if plot:
+        ax4.imshow(temp)
+        ax4.set_title("Selected regions")
+        return temp, f
+    return temp
+
+
+def get_props(mask: np.ndarray, **kwargs):
+    lab = label(mask)
+    # assert lab.max() == 1, f'number of segments is {lab.max()}'
+    reg = regionprops(lab)
+    return [
+        {
+            "area": r.area,
+            "eccentricity": r.eccentricity,
+            "major_axis_length": r.major_axis_length,
+            **kwargs,
+        }
+        for r in reg
+    ]

tests/test_transform.py

@@ -2,7 +2,7 @@
 
 import numpy as np
 from numpy.testing import assert_array_equal
-from nd2shrink import transform, read
+from nd2tif import transform, read
 from scipy.ndimage import binary_erosion
 import logging
 

tests/test_write.py

-from nd2shrink import save
+from nd2tif import save
 import numpy
-from nd2shrink import read
+from nd2tif import read
 
 import os, sys