[doc/fix]add a Jupyternotebook for bezierfit and fix some small bugs in bezierfit

Author: ZhenHuangLab

build/lib/memxterminator/GUI/__init__.py

@@ -1,2 +0,0 @@
-import sys
-sys.path.insert(0, '/data3/Zhen/MemXTerminator/src/')

build/lib/memxterminator/GUI/particle_membrane_subtraction_bezierfit.py

@@ -107,7 +107,7 @@ def retranslateUi(self, ParticleMembraneSubtraction_bezierfit):
         self.controlpoints_label.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Control points file"))
         self.template_browse_pushButton.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Browse..."))
         self.particle.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Particle .cs file"))
-        self.points_step_lineEdit.setText(_translate("ParticleMembraneSubtraction_bezierfit", "0.005"))
+        self.points_step_lineEdit.setText(_translate("ParticleMembraneSubtraction_bezierfit", "0.001"))
         self.points_step_label.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Points_step"))
         self.physical_membrane_dist_label.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Physical membrane distance(Å)"))
         self.physical_membrane_dist_lineEdit.setText(_translate("ParticleMembraneSubtraction_bezierfit", "35"))

build/lib/memxterminator/__init__.py

@@ -1,3 +1,3 @@
 __author__ = 'Zhen Huang'
 __email__ = 'zhen.victor.huang@gmail.com'
-__version__ = '1.2.1'
+__version__ = '1.2.2'

build/lib/memxterminator/bezierfit/bin/mem_subtract_main.py

@@ -80,7 +80,6 @@ def membrane_subtract(particle_filename):
     shifts = shift_list[mask]
     classes = class_list[mask]
     for particle_idx, psi, pixel_size, shift, class_ in zip(particle_idxes, psis, pixel_sizes, shifts, classes):
-        # class_得根据control_points.json找到control_points
         # if str(class_) in control_points_dict:
         control_points = np.array(control_points_dict[str(class_)])
         # print(control_points)

build/lib/memxterminator/bezierfit/lib/bezierfit.py

@@ -14,25 +14,13 @@
 import json
 from scipy.ndimage import zoom
 
-
-# def bezier_curve(control_points, t):
-#     B = np.outer((1 - t) ** 3, control_points[0]) + \
-#         np.outer(3 * (1 - t) ** 2 * t, control_points[1]) + \
-#         np.outer(3 * (1 - t) * t ** 2, control_points[2]) + \
-#         np.outer(t ** 3, control_points[3])
-#     return B.squeeze()
 def bezier_curve(control_points, t):
     n = len(control_points) - 1
     B = np.zeros_like(control_points[0], dtype=float)
     for i, point in enumerate(control_points):
         B += comb(n, i) * (1 - t) ** (n - i) * t ** i * point
     return B
-# def bezier_curve_derivative(control_points, t):
-#     control_points = np.array(control_points)
-#     B_prime = 3 * (1 - t) ** 2 * (control_points[1] - control_points[0]) + \
-#               6 * (1 - t) * t * (control_points[2] - control_points[1]) + \
-#               3 * t ** 2 * (control_points[3] - control_points[2])
-#     return B_prime
+
 def bezier_curve_derivative(control_points, t):
     n = len(control_points) - 1
     B_prime = np.zeros(2)
@@ -41,25 +29,6 @@ def bezier_curve_derivative(control_points, t):
         B_prime += coef * (control_points[i+1] - control_points[i])
     return B_prime
 
-# def bezier_curvature(control_points, t):
-#     dB0 = -3 * (1 - t) ** 2
-#     dB1 = 3 * (1 - t) ** 2 - 6 * t * (1 - t)
-#     dB2 = 6 * t * (1 - t) - 3 * t ** 2
-#     dB3 = 3 * t ** 2
-
-#     ddB0 = 6 * (1 - t)
-#     ddB1 = 6 - 18 * t
-#     ddB2 = 18 * t - 6
-#     ddB3 = 6 * t
-
-#     p = control_points
-#     dx = sum([p[i, 0] * [dB0, dB1, dB2, dB3][i] for i in range(4)])
-#     dy = sum([p[i, 1] * [dB0, dB1, dB2, dB3][i] for i in range(4)])
-#     ddx = sum([p[i, 0] * [ddB0, ddB1, ddB2, ddB3][i] for i in range(4)])
-#     ddy = sum([p[i, 1] * [ddB0, ddB1, ddB2, ddB3][i] for i in range(4)])
-
-#     curvature = abs(dx * ddy - dy * ddx) / (dx * dx + dy * dy) ** 1.5
-#     return curvature
 def bezier_curvature(control_points, t, threshold=1e-6, high_curvature_value=1e6):
     n = len(control_points) - 1
 
@@ -79,7 +48,6 @@ def bezier_curvature(control_points, t, threshold=1e-6, high_curvature_value=1e6
 
     magnitude_squared = dx * dx + dy * dy
 
-    # 规避除数接近零的问题
     if magnitude_squared < threshold:
         return high_curvature_value
 
@@ -331,22 +299,3 @@ def generate_curve_within_boundaries(control_points, image_shape, step):
             break
     fitted_curve_points = np.array([bezier_curve(control_points, t_val) for t_val in t_values])
     return np.array(fitted_curve_points), np.array(t_values)
-
-if __name__ == '__main__':
-    multiprocessing.set_start_method('spawn', force=True)
-    with mrcfile.open('/data3/kzhang/cryosparc/CS-vsv/J354/templates_selected.mrc') as f:
-        image = f.data[2]
-    image = zoom(image, 2)
-    coarsefit = Coarsefit(image, 600, 3, 300, 20)
-    initial_control_points = coarsefit()
-    ga_refine = GA_Refine(image, 1.068, 0.05, 50, 700, 18)
-    refined_control_points = ga_refine(initial_control_points, image)
-    refined_control_points = np.array(refined_control_points)
-    fitted_curve_points, t_values = generate_curve_within_boundaries(refined_control_points, image.shape, 0.01)
-    # save the control points in JSON format
-    with open('control_points.json', 'w') as f:
-        json.dump(refined_control_points.tolist(), f)
-    plt.imshow(image, cmap='gray')
-    plt.plot(fitted_curve_points[:, 0], fitted_curve_points[:, 1], 'r-')
-    plt.plot(refined_control_points[:, 0], refined_control_points[:, 1], 'g.')
-    plt.show()

build/lib/memxterminator/bezierfit/lib/subtraction.py

@@ -153,9 +153,9 @@ def generate_2d_mask(self, image, fitted_points, membrane_distance):
         membrane_mask[mask_outside_distance & ~mask_small_gray_value] = gray_value[mask_outside_distance & ~mask_small_gray_value]
         return membrane_mask
 
-    def average_1d(self, image_gpu, fitted_points, normals, extra_mem_dist):
+    def average_1d(self, image_gpu, fitted_points, normals, mem_dist):
         average_1d_lst = []
-        for membrane_dist in range(-extra_mem_dist, extra_mem_dist+1):
+        for membrane_dist in range(-mem_dist, mem_dist+1):
             normals_points = fitted_points + membrane_dist * normals
             # Ensure the points are within the image boundaries
             mask = (normals_points[:, 0] >= 0) & (normals_points[:, 0] < image_gpu.shape[1]) & \
@@ -227,11 +227,11 @@ def average_1d(self, image_gpu, fitted_points, normals, extra_mem_dist):
     #         new_image[cp.isnan(new_image)] = 0
     #     return new_image.astype(image.dtype)
 
-    def average_2d(self, image_gpu, fitted_points, normals, average_1d_lst, extra_mem_dist):
+    def average_2d(self, image_gpu, fitted_points, normals, average_1d_lst, mem_dist):
         image = image_gpu.get()
         new_image = np.zeros_like(image)
         count_image = np.zeros_like(image)
-        for membrane_dist, average_1d in zip(range(-extra_mem_dist, extra_mem_dist+1), average_1d_lst):
+        for membrane_dist, average_1d in zip(range(-mem_dist, mem_dist+1), average_1d_lst):
             # start_time = time.time()
             normals_points = fitted_points + membrane_dist * normals
             mask = (normals_points[:, 0] >= 0) & (normals_points[:, 0] < image.shape[1]) & \
@@ -247,11 +247,11 @@ def average_2d(self, image_gpu, fitted_points, normals, average_1d_lst, extra_me
             new_image[np.isnan(new_image)] = 0
         return new_image.astype(image.dtype)
 
-    def average_2d_gpu(self, image_gpu, fitted_points, normals, average_1d_lst, extra_mem_dist):
+    def average_2d_gpu(self, image_gpu, fitted_points, normals, average_1d_lst, mem_dist):
         new_image = cp.zeros_like(image_gpu)
         count_image = cp.zeros_like(image_gpu)
         fitted_points = cp.asarray(fitted_points)
-        membrane_dists = cp.arange(-extra_mem_dist, extra_mem_dist + 1)
+        membrane_dists = cp.arange(-mem_dist, mem_dist + 1)
         # Expand dimensions for broadcasting
         membrane_dists = membrane_dists[:, cp.newaxis, cp.newaxis]
         # Calculate all normals_points at once
@@ -275,14 +275,11 @@ def average_2d_gpu(self, image_gpu, fitted_points, normals, average_1d_lst, extr
     def mem_subtract(self):
         control_points = self.control_points_trasf(self.control_points, self.psi, self.origin_x, self.origin_y)
         fitted_curve_points, t_values = generate_curve_within_boundaries(control_points, self.image.shape, self.points_step)
-        # plt.imshow(self.image, cmap='gray')
-        # plt.plot(fitted_curve_points[:, 0], fitted_curve_points[:, 1], 'r-')
-        # plt.plot(control_points[:, 0], control_points[:, 1], 'g.')
-        # plt.show()
+        extra_mem_dist = 10
         mem_mask = self.generate_2d_mask(self.image_gpu, fitted_curve_points, self.mem_dist)
-        raw_image_average_1d_lst = self.average_1d(self.image_gpu, fitted_curve_points, points_along_normal(control_points, t_values).get(), self.mem_dist)
-        raw_image_average_2d = self.average_2d(self.image_gpu, fitted_curve_points, points_along_normal(control_points, t_values).get(), raw_image_average_1d_lst, self.mem_dist)
-        raw_image_average_2d = cp.asarray(raw_image_average_2d)
+        raw_image_average_1d_lst = self.average_1d(self.image_gpu, fitted_curve_points, points_along_normal(control_points, t_values).get(), self.mem_dist+extra_mem_dist)
+        raw_image_average_2d = self.average_2d(self.image_gpu, fitted_curve_points, points_along_normal(control_points, t_values).get(), raw_image_average_1d_lst, self.mem_dist+extra_mem_dist)
+        raw_image_average_2d = cp.asarray(raw_image_average_2d) * mem_mask
         kernel = gaussian_kernel(5, 1)
         image_conv = convolve2d(self.image_gpu, kernel, mode = 'same')
         raw_image_average_2d_conv = convolve2d(raw_image_average_2d, kernel, mode = 'same')

build/lib/memxterminator/cli/__init__.py

@@ -1,2 +0,0 @@
-import sys
-sys.path.insert(0, '/data3/Zhen/MemXTerminator/src/')

build/lib/memxterminator/cli/bezierfit_cli.py

@@ -105,12 +105,11 @@ def kill_process(self):
                 os.remove(self.PID_FILE)
             self.timer.stop()
     def update_log(self):
-        # 读取日志文件内容
         try:
             with open('run.out', 'r') as f:
-                f.seek(self.last_read_position)  # 跳转到上次读取的位置
-                new_content = f.read()  # 读取新内容
-                self.last_read_position = f.tell()  # 更新读取的位置
+                f.seek(self.last_read_position)
+                new_content = f.read()
+                self.last_read_position = f.tell()
             if new_content:
                 self.LOG_textBrowser.append(new_content)
         except FileNotFoundError:
@@ -198,12 +197,11 @@ def kill_process(self):
                 os.remove(self.PID_FILE)
             self.timer.stop()
     def update_log(self):
-        # 读取日志文件内容
         try:
             with open('run.out', 'r') as f:
-                f.seek(self.last_read_position)  # 跳转到上次读取的位置
-                new_content = f.read()  # 读取新内容
-                self.last_read_position = f.tell()  # 更新读取的位置
+                f.seek(self.last_read_position)
+                new_content = f.read()
+                self.last_read_position = f.tell()
             if new_content:
                 self.LOG_textBrowser.append(new_content)
         except FileNotFoundError:

build/lib/memxterminator/cli/radonfit_cli.py

@@ -222,12 +222,11 @@ def kill_process(self):
                 os.remove(self.PID_FILE)
             self.timer.stop()
     def update_log(self):
-        # 读取日志文件内容
         try:
             with open('run.out', 'r') as f:
-                f.seek(self.last_read_position)  # 跳转到上次读取的位置
-                new_content = f.read()  # 读取新内容
-                self.last_read_position = f.tell()  # 更新读取的位置
+                f.seek(self.last_read_position)
+                new_content = f.read()
+                self.last_read_position = f.tell()
             if new_content:
                 self.textBrowser_log.append(new_content)
         except FileNotFoundError:
@@ -274,7 +273,7 @@ def __init__(self, parent=None):
         self.last_read_position = 0
         self.timer = QtCore.QTimer(self)
         self.timer.timeout.connect(self.update_log)
-        self.timer.start(1000)  # 每秒更新一次
+        self.timer.start(1000)
 
 
     def browse_mem_analysis_starfile(self):

dist/MemXTerminator-1.2.2-py3-none-any.whl

@@ -124,5 +124,6 @@ nav:
   - Frequently Asked Questions: ./tutorials/faq.md
   - Reference Pages:
     - "Visualize Radonfit (.ipynb)": ./tutorials/reference/radonfit-mem-analysis-visualizer.ipynb
+    - "Visualize Bezierfit (.ipynb)": ./tutorials/reference/bezierfit-mem-analysis-visualizer.ipynb
     - "Conventions": ./tutorials/reference/conventions.md
     - "API Reference": ./tutorials/reference/api.md

dist/MemXTerminator-1.2.2.tar.gz

@@ -2,7 +2,7 @@
 
 setup(
     name='MemXTerminator',
-    version='1.2.1',
+    version='1.2.2',
     packages=find_packages(where='src'),
     package_dir={'': 'src'},
     author='Zhen Huang',

mkdocs.yml

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: MemXTerminator
-Version: 1.2.1
+Version: 1.2.2
 Summary: A software for membrane analysis and subtraction in cryo-EM
 Home-page: https://github.com/ZhenHuangLab/MemXTerminator
 Author: Zhen Huang

setup.py

@@ -1,2 +0,0 @@
-import sys
-sys.path.insert(0, '/data3/Zhen/MemXTerminator/src/')

src/MemXTerminator.egg-info/PKG-INFO

@@ -107,7 +107,7 @@ def retranslateUi(self, ParticleMembraneSubtraction_bezierfit):
         self.controlpoints_label.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Control points file"))
         self.template_browse_pushButton.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Browse..."))
         self.particle.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Particle .cs file"))
-        self.points_step_lineEdit.setText(_translate("ParticleMembraneSubtraction_bezierfit", "0.005"))
+        self.points_step_lineEdit.setText(_translate("ParticleMembraneSubtraction_bezierfit", "0.001"))
         self.points_step_label.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Points_step"))
         self.physical_membrane_dist_label.setText(_translate("ParticleMembraneSubtraction_bezierfit", "Physical membrane distance(Å)"))
         self.physical_membrane_dist_lineEdit.setText(_translate("ParticleMembraneSubtraction_bezierfit", "35"))

src/memxterminator/GUI/__init__.py

@@ -1,3 +1,3 @@
 __author__ = 'Zhen Huang'
 __email__ = 'zhen.victor.huang@gmail.com'
-__version__ = '1.2.1'
+__version__ = '1.2.2'

src/memxterminator/GUI/particle_membrane_subtraction_bezierfit.py

@@ -80,7 +80,6 @@ def membrane_subtract(particle_filename):
     shifts = shift_list[mask]
     classes = class_list[mask]
     for particle_idx, psi, pixel_size, shift, class_ in zip(particle_idxes, psis, pixel_sizes, shifts, classes):
-        # class_得根据control_points.json找到control_points
         # if str(class_) in control_points_dict:
         control_points = np.array(control_points_dict[str(class_)])
         # print(control_points)

src/memxterminator/__init__.py

@@ -14,25 +14,13 @@
 import json
 from scipy.ndimage import zoom
 
-
-# def bezier_curve(control_points, t):
-#     B = np.outer((1 - t) ** 3, control_points[0]) + \
-#         np.outer(3 * (1 - t) ** 2 * t, control_points[1]) + \
-#         np.outer(3 * (1 - t) * t ** 2, control_points[2]) + \
-#         np.outer(t ** 3, control_points[3])
-#     return B.squeeze()
 def bezier_curve(control_points, t):
     n = len(control_points) - 1
     B = np.zeros_like(control_points[0], dtype=float)
     for i, point in enumerate(control_points):
         B += comb(n, i) * (1 - t) ** (n - i) * t ** i * point
     return B
-# def bezier_curve_derivative(control_points, t):
-#     control_points = np.array(control_points)
-#     B_prime = 3 * (1 - t) ** 2 * (control_points[1] - control_points[0]) + \
-#               6 * (1 - t) * t * (control_points[2] - control_points[1]) + \
-#               3 * t ** 2 * (control_points[3] - control_points[2])
-#     return B_prime
+
 def bezier_curve_derivative(control_points, t):
     n = len(control_points) - 1
     B_prime = np.zeros(2)
@@ -41,25 +29,6 @@ def bezier_curve_derivative(control_points, t):
         B_prime += coef * (control_points[i+1] - control_points[i])
     return B_prime
 
-# def bezier_curvature(control_points, t):
-#     dB0 = -3 * (1 - t) ** 2
-#     dB1 = 3 * (1 - t) ** 2 - 6 * t * (1 - t)
-#     dB2 = 6 * t * (1 - t) - 3 * t ** 2
-#     dB3 = 3 * t ** 2
-
-#     ddB0 = 6 * (1 - t)
-#     ddB1 = 6 - 18 * t
-#     ddB2 = 18 * t - 6
-#     ddB3 = 6 * t
-
-#     p = control_points
-#     dx = sum([p[i, 0] * [dB0, dB1, dB2, dB3][i] for i in range(4)])
-#     dy = sum([p[i, 1] * [dB0, dB1, dB2, dB3][i] for i in range(4)])
-#     ddx = sum([p[i, 0] * [ddB0, ddB1, ddB2, ddB3][i] for i in range(4)])
-#     ddy = sum([p[i, 1] * [ddB0, ddB1, ddB2, ddB3][i] for i in range(4)])
-
-#     curvature = abs(dx * ddy - dy * ddx) / (dx * dx + dy * dy) ** 1.5
-#     return curvature
 def bezier_curvature(control_points, t, threshold=1e-6, high_curvature_value=1e6):
     n = len(control_points) - 1
 
@@ -79,7 +48,6 @@ def bezier_curvature(control_points, t, threshold=1e-6, high_curvature_value=1e6
 
     magnitude_squared = dx * dx + dy * dy
 
-    # 规避除数接近零的问题
     if magnitude_squared < threshold:
         return high_curvature_value
 
@@ -331,22 +299,3 @@ def generate_curve_within_boundaries(control_points, image_shape, step):
             break
     fitted_curve_points = np.array([bezier_curve(control_points, t_val) for t_val in t_values])
     return np.array(fitted_curve_points), np.array(t_values)
-
-if __name__ == '__main__':
-    multiprocessing.set_start_method('spawn', force=True)
-    with mrcfile.open('/data3/kzhang/cryosparc/CS-vsv/J354/templates_selected.mrc') as f:
-        image = f.data[2]
-    image = zoom(image, 2)
-    coarsefit = Coarsefit(image, 600, 3, 300, 20)
-    initial_control_points = coarsefit()
-    ga_refine = GA_Refine(image, 1.068, 0.05, 50, 700, 18)
-    refined_control_points = ga_refine(initial_control_points, image)
-    refined_control_points = np.array(refined_control_points)
-    fitted_curve_points, t_values = generate_curve_within_boundaries(refined_control_points, image.shape, 0.01)
-    # save the control points in JSON format
-    with open('control_points.json', 'w') as f:
-        json.dump(refined_control_points.tolist(), f)
-    plt.imshow(image, cmap='gray')
-    plt.plot(fitted_curve_points[:, 0], fitted_curve_points[:, 1], 'r-')
-    plt.plot(refined_control_points[:, 0], refined_control_points[:, 1], 'g.')
-    plt.show()