secondary_utils.py

import numpy as np
import imageio
import matplotlib.pyplot as plt
import scipy.misc


def add_white_noise(arr, mu, sigma, factor, size):
    """ sigma = std 
    var = sigma^2
    """
    noisy_arr = arr + factor * np.random.normal(loc = mu, scale = sigma, size = size) 
    return noisy_arr


def imsave(image, path, type=None):
  
  """
    imageio will save values in its orginal form even if its float
    if type='orginal' is specified
    else scipy save will save the image in (0 - 255 values)
    scipy new update has removed imsave from scipy.misc due
    to reported errors ... so just use imwrite from imageio 
    by declaring orginal and changing the data types accordingly
  """
  if type is "orginal":
    return(imageio.imwrite(path, image))
  else:
    return scipy.misc.imsave(path, image)


def normalize_data_ab(a, b, data):
    min_x = min(data.ravel())
    max_x = max(data.ravel())  
    range_x = max_x - min_x 
    return((b-a)*((data-min_x)/range_x)+a)


def imageio_imread(path):
  """
   imageio based imread reads image in its orginal form even if its in
   - ve floats
  """
  return(imageio.imread(path))


def plot2dlayers(arr, xlabel=None, ylabel=None, title=None, cmap=None, colorbar=True):
    """
	'brg' is the best colormap for reb-green-blue image
	'brg_r': in 'brg' colormap green color area will have
		high values whereas in 'brg_r' blue area will have
		the highest values
    """
    if xlabel is None:
        xlabel=''
    if ylabel is None:
        ylabel=''
    if title is None:
        title=''
    if cmap is None:
        cmap='Greys_r'
    plt.imshow(arr, cmap=cmap)
    cb = plt.colorbar()
    if colorbar is False:
      cb.remove()
    plt.xlabel(xlabel)
    plt.ylabel(ylabel)
    plt.title(title)
    plt.show()
    return


def multi2dplots(nrows, ncols, fig_arr, axis, passed_fig_att=None):
    default_att= {"suptitle": '',
            "split_title": np.asanyarray(['']*(nrows*ncols)),
            "supfontsize": 12,
            "xaxis_vis"  : False,
            "yaxis_vis"  : False,
            "out_path"   : '',
            "figsize"    : [8, 8],
            "cmap"       : 'Greys_r',
            "plt_tight"  : True,
            "colorbar"   : True
                 }
    if passed_fig_att is None:
        fig_att = default_att
    else:
        fig_att = default_att
        for key, val in passed_fig_att.items():
            fig_att[key]=val
    
    f, axarr = plt.subplots(nrows, ncols, figsize = fig_att["figsize"])
    img_ind  = 0
    f.suptitle(fig_att["suptitle"], fontsize = fig_att["supfontsize"])
    for i in range(nrows):
        for j in range(ncols):                
            if (axis==0):
                each_img = fig_arr[img_ind, :, :]
            if (axis==1):
                each_img = fig_arr[:, img_ind, :]
            if (axis==2):
                each_img = fig_arr[:, :, img_ind]
                
            if(nrows==1):
                ax = axarr[j]
            elif(ncols ==1):
                ax =axarr[i]
            else:
                ax = axarr[i,j]
            im = ax.imshow(each_img, cmap = fig_att["cmap"])
            if fig_att["colorbar"] is True:  f.colorbar(im, ax=ax)
            ax.set_title(fig_att["split_title"][img_ind])
            ax.get_xaxis().set_visible(fig_att["xaxis_vis"])
            ax.get_yaxis().set_visible(fig_att["yaxis_vis"])
            img_ind = img_ind + 1
            if fig_att["plt_tight"] is True: plt.tight_layout()
            
    if (len(fig_att["out_path"])==0):
        plt.show()
    else:
        plt.savefig(fig_att["out_path"])