Add file download and histogram stats

docs/tutorial/tutorial-minkowski-functionals.ipynb

@@ -16,7 +16,7 @@
     "4. Euler Characteristic\n",
     "\n",
     "In this notebook, we compare Minkowski functional calculations via 2 methods: \n",
-    "- Counts arrangements of $2^3$ neighborhoods of voxels using Quantimpy [[1]](#1) \n",
+    "- Counts arrangements of $2^3$ neighborhoods of voxels using DPM Tools [[1]](#1) \n",
     "- Creates a mesh using Scikit-Image[[2]](#2), Trimesh, and Porespy[[4]](#4) \n",
     "\n",
     "______\n",
@@ -111,28 +111,7 @@
     "sphere[id_a:id_b, id_a:id_b, id_a:id_b] = ball(radius, dtype=np.uint8)\n",
     "\n",
     "# DPM Tools Image object\n",
-    "sphere = Image(scalar=sphere)\n",
-    "\n",
-    "# plotter_obj = pv.Plotter(notebook=True, off_screen=False, lighting='three lights')\n",
-    "\n",
-    "# # Set background colors\n",
-    "# plotter_obj.set_background(color='w')\n",
-    "\n",
-    "# # Set font colors and sizes\n",
-    "# pv.global_theme.font.color = 'black'\n",
-    "# pv.global_theme.font.size = 18\n",
-    "# pv.global_theme.font.label_size = 14\n",
-    "\n",
-    "# pv.set_jupyter_backend('static')\n",
-    "\n",
-    "# plotter_obj = plot_isosurface(sphere, plotter_obj, show_isosurface=[0.5], \n",
-    "#                 mesh_kwargs={\"opacity\":1, \n",
-    "#                              \"color\":(200 / 255, 181 / 255, 152 / 255), \n",
-    "#                              \"diffuse\": 0.75, \n",
-    "#                              \"ambient\": 0.15})\n",
-    "\n",
-    "\n",
-    "# plotter_obj.show(jupyter_backend='static')"
+    "sphere = Image(scalar=sphere)\n"
    ]
   },
   {

dpm_tools/io/__init__.py

@@ -19,12 +19,11 @@
     natural_sort
     combine_slices
     convert_filetype
-
-
+    download_file
 
 """
 
-from ._io_utils import find_files_with_ext, get_tiff_metadata, natural_sort, combine_slices, convert_filetype
+from ._io_utils import find_files_with_ext, get_tiff_metadata, natural_sort, combine_slices, convert_filetype, download_file
 
 from ._read_data import read_image, Image
 

dpm_tools/io/_io_utils.py

@@ -5,6 +5,7 @@
 import pathlib
 from typing import Any, Tuple
 import re
+import wget
 
 from ._read_data import read_image
 from ._write_data import write_image
@@ -184,3 +185,21 @@ def convert_filetype(filepath: pathlib.Path, convert_to: str, **kwargs) -> None:
     new_filepath = filepath.with_suffix(convert_to)
     write_image(save_path=new_filepath.parent, save_name=new_filepath.name,
                 image=original_image, filetype=convert_to)
+
+def download_file(url: str, save_path: str = ".") -> None:
+    """
+    Download a file from the provided URL and save it to the specified save path.
+
+    Args:
+        url (str): URL of the file to download
+        save_path (pathlib.Path, optional): File path where to save the file to. Defaults to pathlib.Path(".").
+    
+    Returns:
+        None
+    """
+
+    # savepath = pathlib.Path(save_path)
+    wget.download(url, out=save_path)
+
+    return
+

dpm_tools/metrics/__init__.py

@@ -23,10 +23,11 @@
     _morph_drain_config
     heterogeneity_curve
     minkowski_map
+    histogram_statistics
 """
 
 from ._maps import slicewise_edt, slicewise_mis, edt, sdt, mis, chords, time_of_flight, constriction_factor, minkowski_map
 
 from ._feature_utils import _morph_drain_config, _set_linear_trend
 
-from ._scalars import minkowski_functionals, morph_drain, heterogeneity_curve
+from ._scalars import minkowski_functionals, morph_drain, heterogeneity_curve, histogram_statistics

dpm_tools/metrics/_scalars.py

@@ -1,5 +1,7 @@
 import numpy as np
-from typing import Tuple
+import matplotlib.pyplot as plt
+from scipy import stats
+from typing import Tuple, List, Dict
 import edt
 from ._minkowski_coeff import contributions_2d, contributions_3d
 from ._feature_utils import _morph_drain_config, _get_heterogeneity_centers_3d
@@ -172,3 +174,72 @@ def heterogeneity_curve(image: np.ndarray, no_radii: int = 50, n_samples_per_rad
         variance[i] = np.var(porosity)
 
     return radii, variance
+
+def _hist_stats(image: np.ndarray, nbins: int = 256) -> Dict:
+    """
+    Calculate histogram statistics of an image.
+
+    Parameters:
+        image: A 2D or 3D image.
+        nbins: Number of histogram bins. Defaults to 256.
+
+    Returns:
+        dict: A dictionary of image metadata and histogram statistics.
+        Includes the shape, dtype, min, max, mean, median, variance, skewness, and kurtosis.
+    """
+
+    stats_dict = {'shape': image.shape, 
+                  'dtype': image.dtype,
+                  'min': np.amin(image),
+                  'max': np.amax(image),
+                  'mean': np.mean(image),
+                  'median': np.median(image),
+                  'variance': np.var(image)}
+
+    hist, bin_edges = np.histogram(image.flatten(), bins=256)
+    bin_centers = 0.5 * (bin_edges[:-1] + bin_edges[1:])
+
+    stats_dict['skewness'] = stats.skew(bin_centers)
+    stats_dict['kurtosis'] = stats.kurtosis(bin_centers, fisher=True)
+
+    return stats_dict
+
+def histogram_statistics(*images: np.ndarray, plot_histogram: bool = False, nbins: int = 256, legend_elem: List = []) -> List[Dict]:
+    """
+    Get the statistics of multiple images.
+
+    Parameters:
+        plot_histograms: Plot the image histograms. Defaults to False.
+        bins: Number of histogram bins. Defaults to 256.
+
+    Returns:
+        dict: A dictionary of image metadata and histogram statistics.
+        Includes the shape, dtype, min, max, mean, median, variance, skewness, and kurtosis.
+    """     
+
+    img_stats = []
+    img_list = []
+
+    for image in images:
+        stats = _hist_stats(image, nbins)
+
+        print('-'*35)
+        print('Image statistics:')
+        print(f'\tShape: {stats["shape"]}')
+        print(f'\tData type: {stats["dtype"]}')
+        print(f'\tMin: {stats["min"]}, Max: {stats["max"]}, Mean: {stats["mean"]}\n')
+
+
+        img_stats.append(_hist_stats(image, nbins))
+        img_list.append(image.flatten())
+
+    if plot_histogram:
+        assert len(legend_elem) == len(images), 'Number of legend elements must match the number of images'
+        fig, ax = plt.subplots()
+        ax.hist(img_list, bins=nbins, density=True, histtype='step', fill=False)
+        ax.set_xlabel('Image Value', fontsize=16)
+        ax.set_ylabel('Probability Density', fontsize=16)
+        ax.grid(True) 
+        ax.legend(legend_elem)
+
+    return img_stats

examples/grayscale_filtering.py

@@ -0,0 +1,50 @@
+from dpm_tools.io import download_file
+from dpm_tools.metrics import histogram_statistics as hist_stats
+import skimage.io
+import numpy as np
+import matplotlib.pyplot as plt
+
+if __name__ == "__main__":
+    file_url1 = "https://www.digitalrocksportal.org/projects/211/images/126114/download/";
+    filename1 = "Project211_sandstone_slice_0801hr.png"
+    download_file(file_url1, filename1)
+
+    # image location, this is a lower resolution image (see project)
+    file_url2 = "https://www.digitalrocksportal.org/projects/211/images/126113/download/"
+    filename2 = "Project211_sandstone_slice_0801lr.png"
+    download_file(file_url2, filename2)
+    # CHANGE MADE - using one of the three channels
+    img_hr = skimage.io.imread(filename1)
+    img_hr = img_hr[:,:,0] #use only one channel of the three
+
+    img_lr = skimage.io.imread(filename2)
+    img_lr = img_lr[:,:,0] #use only one channel of the three
+
+    f = plt.figure(figsize=(24, 24))
+    f.add_subplot(221)
+    plt.imshow(img_hr) # if the image has three channels, default - and fake - color will appear when using imshow. Color scheme can easily be changed.
+    plt.title('higher res',fontsize=20)
+
+    f.add_subplot(222)
+    plt.imshow(img_lr) # if the image has three channels, default - and fake - color will appear when using imshow
+    plt.title('lower res',fontsize=20)
+
+    # location and length of an image subset in x, y directions
+    locx = 470
+    lenx = 90
+    locy = 440
+    leny = 90
+
+    f.add_subplot(223)
+    plt.imshow(img_hr[locx:(locx+lenx),locy:(locy+leny)])
+    plt.title('zoom in - higher res',fontsize=20)
+
+    f.add_subplot(224)
+    # note that // is floor division (so we get integers as a result)
+    plt.imshow(img_lr[locx // 4:(locx+lenx)//4,locy//4:(locy+leny)//4]) 
+    plt.title('zoom in - lower res', fontsize=20)
+
+    hist_stats(img_hr, img_lr, plot_histogram=True, nbins=128, legend_elem=["img_hr", "img_lr"])
+    plt.show()
+
+

examples/segmentation_example.py

@@ -11,7 +11,6 @@
 # print(segmented.dtype)
 # srm_obj.segment()
 segmented = srm(image[0].astype(np.uint16), Q=5.0)
-print(segmented)
 
 # Plot the result
 fig, ax = plt.subplots(1, 2, figsize=(10, 5))

pyproject.toml

@@ -36,6 +36,7 @@ dependencies = [
     "scikit-fmm", 
     "connected-components-3d",
     "pyfftw",
+    "wget",
     "dpm-srm",
     "dpm-srg"]