# -*- coding: utf-8 -*-

# <nbformat>3.0</nbformat>

# <codecell>

import os

import numpy as np

from scipy.ndimage import maximum_filter, minimum_filter, binary_fill_holes

import skimage as ski

from skimage import io, filter, color, exposure, morphology, feature, draw, measure, transform

figsize(16, 10)

# <codecell>

l = os.listdir("../data")

print l

qstain = np.array([[.26451728, .5205347, .81183386], [.9199094, .29797825, .25489032], [.28947765, .80015373, .5253158]])

# <codecell>

# Original image

A = transform.rescale(io.imread("../data/" + l[0]), 0.25)

io.imshow(A)

# <codecell>

# Colour Deconvolution

deconv = ski.img_as_float(color.separate_stains(A, np.linalg.inv(qstain)))

"""

io.imshow(exposure.equalize_adapthist(

          exposure.adjust_sigmoid(

            filter.gaussian_filter(

                exposure.rescale_intensity(

                    1-A[:, :, 2], 

                    out_range=(0, 1)), 

                4), 

            gain=7, cutoff=0.7)))

"""

#io.imshow(exposure.adjust_sigmoid(filter.gaussian_filter(exposure.equalize_adapthist(exposure.rescale_intensity(1 - A[:, :, 2], out_range=(0, 1)), ntiles_y=1), 3), cutoff=0.55))

# <codecell>

# Blood

"""

blood = \

morphology.remove_small_objects(

    morphology.binary_closing(

        morphology.remove_small_objects(

            filter.threshold_adaptive(

                exposure.adjust_sigmoid(

                    filter.gaussian_filter(

                    exposure.equalize_adapthist(

                        exposure.rescale_intensity(

                            1. - A[:, :, 2], 

                            out_range=(0, 1)), 

                        ntiles_y = 1),

                        3), 

                    cutoff=0.55),

                100, offset=-0.2),

            50),

        morphology.disk(20)).astype(bool),

    100)

io.imshow(blood)

print type(blood)

"""

# <codecell>

# Veins

veins = \

morphology.remove_small_objects(

    binary_fill_holes(

        morphology.binary_closing(

            morphology.remove_small_objects(      

                filter.threshold_adaptive(

                    filter.gaussian_filter(      

                        maximum_filter(

                            deconv[:, :, 2] / deconv[:, :, 0],

                            3),

                        7),

                    170, offset=-0.13),

                60),

            morphology.disk(29))

        ),

    250)

io.imshow(veins)

# <codecell>

io.imshow( \

          filter.gaussian_filter(      

                            maximum_filter(

                                exposure.rescale_intensity(

                                deconv[:, :, 2] / deconv[:, :, 0], out_range=(0,1)),

                                2),

                            11))

# <codecell>

# Inflammation

inflammation = \

morphology.binary_closing(

    morphology.remove_small_objects(

        filter.threshold_adaptive(

            exposure.adjust_sigmoid(

                filter.gaussian_filter(

                    exposure.equalize_adapthist(

                        exposure.rescale_intensity(

                            deconv[:, :, 1],

                            out_range = (0, 1)),

                        ntiles_y = 3),

                    5),

                cutoff = 0.6),

            50, offset = -0.1),

        150),

    morphology.disk(19))

"""

morphology.remove_small_objects(

    filter.threshold_adaptive(

        exposure.adjust_sigmoid(

            filter.gaussian_filter(

                exposure.rescale_intensity(

                    deconv[:, :, 1],

                    out_range = (0, 1)),

                25),

            gain = 12, cutoff = 0.32),

        501, offset = -0.1),

    4000)

"""

io.imshow(inflammation)

#dec = exposure.adjust_sigmoid(filter.gaussian_filter(exposure.equalize_adapthist(exposure.rescale_intensity(deconv[:, :, 1], out_range=(0, 1)), ntiles_y=3), 5), cutoff=0.6)

#io.imshow(filter.threshold_adaptive(dec, 50, offset=-0.1))

#io.imshow(filter.threshold_adaptive(exposure.adjust_sigmoid(filter.gaussian_filter(exposure.rescale_intensity(deconv[:, :, 1], out_range=(0, 1)), 19), gain=15, cutoff=0.4), 401))

# <codecell>

# Labelled

total = np.zeros_like(A)

#total[:, :, 0] = blood

total[:, :, 1] = veins

total[:, :, 2] = inflammation

io.imshow(total)

# <codecell>

# All regions of interest

#%timeit

"""

all = \

minimum_filter(

    maximum_filter(

        np.logical_or(np.logical_or(blood, veins), inflammation).astype(float),

        size = 101),

    size = 51)

"""

"""

morphology.binary_erosion(

    morphology.binary_dilation(

        np.logical_or(np.logical_or(blood, veins), inflammation),

        morphology.disk(31)),

    morphology.disk(21))

"""

#full = (np.logical_or(np.logical_or(blood, veins), inflammation))

#io.imshow(binary_fill_holes(minimum_filter(maximum_filter(full, size=51, footprint=morphology.diamond(31)), size=51)))

# <codecell>

#io.imshow(ski.img_as_float(A)*0.5 + full[:, :, newaxis]*0.5)

#io.imshow(morphology.binary_closing(full, morphology.disk(15)))

# <codecell>

io.imsave("__1.gif", A)

io.imsave("__2.gif", total)

#io.imsave("__2.gif", (morphology.binary_closing(full, morphology.disk(15))).astype(float))

# <codecell>

os.system("gifsicle --delay=80 --loop __*.gif > test.gif")

# <codecell>

print B.shape

print transform.rescale(B, 0.5).shape

# <codecell>