#!/usr/bin/env python
# coding: utf-8

# # Tissue neighborhood quantification
# 
# In this notebook, we will analyse neighborhood-relationships between cells. 
# We count the number of neighbors for each cell and take a look at this number in a parametric image.
# Afterwards, we average this number locally between neighbors.
# 
# ## Test data
# Let's generate some cells and take a look at the borders between them

# In[1]:


import pyclesperanto_prototype as cle
import numpy as np
from skimage.io import imshow
import matplotlib


# In[2]:


# Generate artificial cells as test data
tissue = cle.artificial_tissue_2d()

cle.imshow(tissue, labels=True)


# In[3]:


membranes = cle.detect_label_edges(tissue)
cle.imshow(membranes)


# # Analysis and visualization of neighbor count

# In[4]:


touch_matrix = cle.generate_touch_matrix(tissue)
neighbor_count = cle.count_touching_neighbors(touch_matrix)

parametric_image = cle.replace_intensities(tissue, neighbor_count)
cle.imshow(parametric_image, min_display_intensity=0, max_display_intensity=10, color_map='jet')


# ## Average the measurement between cells to reduce noise
# Mean of touching neighbors

# In[5]:


local_mean_neighbor_count = cle.mean_of_touching_neighbors(neighbor_count, touch_matrix)

parametric_image = cle.replace_intensities(tissue, local_mean_neighbor_count)
cle.imshow(parametric_image, min_display_intensity=0, max_display_intensity=10, color_map='jet')


# Median of touching neighbors

# In[6]:


local_median_neighbor_count = cle.median_of_touching_neighbors(neighbor_count, touch_matrix)

parametric_image = cle.replace_intensities(tissue, local_median_neighbor_count)
cle.imshow(parametric_image, min_display_intensity=0, max_display_intensity=10, color_map='jet')


# In[7]:


local_minimum_neighbor_count = cle.minimum_of_touching_neighbors(neighbor_count, touch_matrix)

parametric_image = cle.replace_intensities(tissue, local_minimum_neighbor_count)
cle.imshow(parametric_image, min_display_intensity=0, max_display_intensity=10, color_map='jet')


# In[8]:


local_maximum_neighbor_count = cle.maximum_of_touching_neighbors(neighbor_count, touch_matrix)

parametric_image = cle.replace_intensities(tissue, local_maximum_neighbor_count)
cle.imshow(parametric_image, min_display_intensity=0, max_display_intensity=10, color_map='jet')


# In[9]:


local_standard_deviation_neighbor_count = cle.standard_deviation_of_touching_neighbors(neighbor_count, touch_matrix)

parametric_image = cle.replace_intensities(tissue, local_standard_deviation_neighbor_count)
cle.imshow(parametric_image, min_display_intensity=0, max_display_intensity=10, color_map='jet')


# In[10]:


# most popular number of neighbors locally
local_mode_neighbor_count = cle.mode_of_touching_neighbors(neighbor_count, touch_matrix)

parametric_image = cle.replace_intensities(tissue, local_mode_neighbor_count)
cle.imshow(parametric_image, min_display_intensity=0, max_display_intensity=10, color_map='jet')


# In[ ]: