GIproject/qupath_points.py at master · goldenberg-lab/GIproject · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
"""
SCRIPT TO PARSE THE QU-PATH POINTS
"""

import sys
import os
from zipfile import ZipFile
import shutil
import numpy as np
import pandas as pd

import seaborn as sns
from matplotlib import pyplot as plt
from PIL import Image

dir_base = os.getcwd()
dir_data = os.path.join(dir_base,'..','data')
dir_cropped = os.path.join(dir_data,'cropped')
dir_points = os.path.join(dir_data, 'points')

def stopifnot(cond):
    if not cond:
        sys.exit('error!')

valid_cells = ['eosinophil','neutrophil','plasma','enterocyte','other','lymphocyte']

# Function to parse the zipped file
def zip_points_parse(fn):
    valid_files = ['Points '+str(k+1)+'.txt' for k in range(6)]
    with ZipFile(file=fn,mode='r') as zf:
        names = pd.Series(zf.namelist())
        stopifnot(names.isin(valid_files).all())
        zf.extractall('tmp')
    # Loop through and parse files
    holder = []
    for pp in names:
        s_pp = pd.read_csv(os.path.join(dir_base,'tmp',pp),sep='\t',header=None)
        stopifnot(s_pp.loc[0,0] == 'Name')
        cell_pp = s_pp.loc[0,1].lower()
        stopifnot(cell_pp in valid_cells)
        df_pp = pd.DataFrame(s_pp.loc[3:].values.astype(float),columns=['x','y'])
        stopifnot(df_pp.shape[0] == int(s_pp.loc[2,1])) # number of coords lines up
        df_pp.insert(0,'cell',cell_pp)
        holder.append(df_pp)
    df = pd.concat(holder).reset_index(drop=True)
    shutil.rmtree('tmp', ignore_errors=True) # Get rid of temporary folder
    return(df)


##################################
## --- (1) LOAD IN THE DATA --- ##

# Get the points
fn_points = os.listdir(dir_points)
raw_points = pd.Series(fn_points).str.split('_',expand=True).iloc[:,0:3].apply(lambda z: '_'.join(z), axis=1)

# Load the labels
df_data = pd.read_csv(os.path.join(dir_data,'df_lbls_anon.csv'),usecols=['type','ID','file'])
df_data['raw'] = df_data.file.str.replace('.png','')
stopifnot(all(raw_points.isin(df_data.raw)))


fig, axes = plt.subplots(nrows=len(fn_points),ncols=1,figsize=(10,20))

for ii, fn in enumerate(fn_points):
    print('file %s (%i of %i)' % (fn, ii+1, len(fn_points)))
    path = os.path.join(dir_points, fn)
    df_ii = zip_points_parse(path)
    df_ii.cell = pd.Categorical(df_ii.cell,categories=valid_cells)
    fold = '_'.join(fn.split('_')[0:3])
    tissue = fold.split('_')[2]
    tt = df_data[df_data.raw == fold]['type'].values[0]
    file = df_data[df_data.raw == fold]['file'].values[0]
    ID = df_data[df_data.raw == fold]['ID'].values[0]
    nam = fn.replace('-points.zip','')
    path_img = os.path.join(dir_cropped, tt, ID, tissue, nam)
    img_vals =  np.array(Image.open(path_img))
    axes[ii].imshow(img_vals)
    #axes.imshow(img_vals)
    sns.scatterplot(x='x',y='y',hue='cell',data=df_ii,ax=axes[ii])