p2ch12_explore_data.py

# ---
# jupyter:
#   jupytext:
#     text_representation:
#       extension: .py
#       format_name: light
#       format_version: '1.5'
#       jupytext_version: 1.16.3
#   kernelspec:
#     display_name: Python 3
#     language: python
#     name: python3
# ---

# %matplotlib inline
import copy
import numpy as np
import matplotlib.pyplot as plt

# +
import torch
from p2ch12.dsets import getCandidateInfoList, getCt, LunaDataset
from util.util import xyz2irc


candidateInfo_list = getCandidateInfoList(requireOnDisk_bool=False)
candidateInfo_list[0]
# -

from p2ch12.vis import findPositiveSamples, showCandidate
positiveSample_list = findPositiveSamples()

# +
augmentation_dict = {}
augmentation_list = [
    ('None', {}),
    ('flip', {'flip': True}),
    ('offset', {'offset': 0.1}),
    ('scale', {'scale': 0.2}),
    ('rotate', {'rotate': True}),
    ('noise', {'noise': 25.0}),    
]
ds_list = [
    LunaDataset(sortby_str='label_and_size', augmentation_dict=augmentation_dict) 
    for title_str, augmentation_dict in augmentation_list
]

all_dict = {}
for title_str, augmentation_dict in augmentation_list:
    all_dict.update(augmentation_dict)
all_ds = LunaDataset(sortby_str='label_and_size', augmentation_dict=all_dict)

augmentation_list.extend([('All', augmentation_dict)] * 3)
ds_list.extend([all_ds] * 3)



# +
sample_ndx = 100
sample_ndx = 154
sample_ndx = 155


sample_tup = all_ds[sample_ndx]
print(sample_tup[0].shape, sample_tup[1:])

fig = plt.figure(figsize=(30, 30))

clim=(-1000.0, 300)

for i, ((title_str, _), ds) in enumerate(zip(augmentation_list, ds_list)):
    sample_tup = ds[sample_ndx]
    subplot = fig.add_subplot(3, 3, i+1)
    subplot.set_title(title_str, fontsize=30)
    for label in (subplot.get_xticklabels() + subplot.get_yticklabels()):
        label.set_fontsize(20)
    plt.imshow(sample_tup[0][0][16], clim=clim, cmap='gray')


# -

series_uid = positiveSample_list[1][2]
showCandidate(series_uid)