|
1
|
|
|
import pydicom |
|
2
|
|
|
import sys |
|
3
|
|
|
import os |
|
4
|
|
|
import argparse |
|
5
|
|
|
import getpass |
|
6
|
|
|
import logging |
|
7
|
|
|
from pydicom.data import get_testdata_files |
|
8
|
|
|
|
|
9
|
|
|
logging.basicConfig(stream=sys.stdout, level=logging.INFO) |
|
10
|
|
|
|
|
11
|
|
|
def DICOM_decompressor(path): |
|
12
|
|
|
""" |
|
13
|
|
|
Decompress the dicoms in case they are compressed. |
|
14
|
|
|
:param path: |
|
15
|
|
|
:return: |
|
16
|
|
|
""" |
|
17
|
|
|
|
|
18
|
|
|
ds = pydicom.dcmread(path) |
|
19
|
|
|
|
|
20
|
|
|
ds.decompress() |
|
21
|
|
|
ds.save_as("Test.dcm") |
|
22
|
|
|
|
|
23
|
|
|
|
|
24
|
|
|
def DICOM_anonymizer(path, PSCID): |
|
25
|
|
|
""" |
|
26
|
|
|
Anonymize the DICOMS to remove any identifiable information |
|
27
|
|
|
:param path: |
|
28
|
|
|
:param PSCID: |
|
29
|
|
|
:return: |
|
30
|
|
|
""" |
|
31
|
|
|
|
|
32
|
|
|
return DICOM_anonymizer_save_as(path, PSCID, path) |
|
33
|
|
|
|
|
34
|
|
|
|
|
35
|
|
|
def DICOM_anonymizer_save_as(path, PSCID, out_path): |
|
36
|
|
|
""" |
|
37
|
|
|
Anonymize the DICOMS to remove any identifiable information |
|
38
|
|
|
:param path: |
|
39
|
|
|
:param PSCID: |
|
40
|
|
|
:return: |
|
41
|
|
|
""" |
|
42
|
|
|
|
|
43
|
|
|
success, dataset = DICOM_validator(path) |
|
44
|
|
|
if not success: |
|
45
|
|
|
return False |
|
46
|
|
|
|
|
47
|
|
|
dataset.PatientID = PSCID |
|
48
|
|
|
dataset.PatientName = PSCID |
|
49
|
|
|
dataset.save_as(out_path) |
|
50
|
|
|
|
|
51
|
|
|
return True |
|
52
|
|
|
|
|
53
|
|
|
|
|
54
|
|
|
def DICOM_recursive_loader(path, action): |
|
55
|
|
|
""" |
|
56
|
|
|
load all the files, validate and then pass to decompress or anonimize. |
|
57
|
|
|
:param path: |
|
58
|
|
|
:return: |
|
59
|
|
|
""" |
|
60
|
|
|
|
|
61
|
|
|
|
|
62
|
|
|
def DICOM_validator(file_path): |
|
63
|
|
|
""" |
|
64
|
|
|
validate to check if the DICOM file is an actual DICOM file. |
|
65
|
|
|
:param file_path: |
|
66
|
|
|
:return: |
|
67
|
|
|
""" |
|
68
|
|
|
logger = logging.getLogger(__name__) |
|
69
|
|
|
|
|
70
|
|
|
global dicom |
|
71
|
|
|
dicom = None |
|
72
|
|
|
|
|
73
|
|
|
from pydicom.filereader import InvalidDicomError |
|
74
|
|
|
from pydicom.filereader import read_file |
|
75
|
|
|
try: |
|
76
|
|
|
dicom = read_file(file_path) |
|
77
|
|
|
except InvalidDicomError: |
|
78
|
|
|
logger.info(file_path + " is not a DICOM file. Skipping") |
|
79
|
|
|
return False, None |
|
80
|
|
|
return True, dicom |
|
81
|
|
|
|
|
82
|
|
|
|
|
83
|
|
|
def DICOM_batchValidator(dir_path): |
|
84
|
|
|
""" |
|
85
|
|
|
Some basic information of the participants must be consistent across the files, such as the SCAN DATE (assuming they are not scanning across MIDNIGHT POINT) |
|
86
|
|
|
Birthday date, subject name, etc MUST BE CONSISTENT across a SINGLE subject's folder, RIGHT! |
|
87
|
|
|
|
|
88
|
|
|
:param dir_path: |
|
89
|
|
|
:return: |
|
90
|
|
|
""" |
|
91
|
|
|
|
|
92
|
|
|
|
|
93
|
|
|
def DICOM_TransferSyntax(file_path): |
|
94
|
|
|
""" |
|
95
|
|
|
Used to find if a file is compressed |
|
96
|
|
|
:param file_path: |
|
97
|
|
|
:return: |
|
98
|
|
|
""" |
|
99
|
|
|
from pydicom.filereader import read_file_meta_info |
|
100
|
|
|
|
|
101
|
|
|
# Validity check: |
|
102
|
|
|
success, DICOM = DICOM_validator(file_path) |
|
103
|
|
|
if not success: |
|
104
|
|
|
raise IOError |
|
105
|
|
|
|
|
106
|
|
|
# Now read the meta information. |
|
107
|
|
|
dicom_file = read_file_meta_info(file_path) |
|
108
|
|
|
transfer_syntax = dicom_file.TransferSyntaxUID |
|
109
|
|
|
|
|
110
|
|
|
return transfer_syntax |
|
111
|
|
|
|
|
112
|
|
|
|
|
113
|
|
|
def DICOM_computeScanAge(file_path): |
|
114
|
|
|
""" |
|
115
|
|
|
Read the PatientID field which normally used as MRN number. |
|
116
|
|
|
:param file_path: |
|
117
|
|
|
:return: Age as a relative delta time object. |
|
118
|
|
|
""" |
|
119
|
|
|
|
|
120
|
|
|
from dateutil.relativedelta import relativedelta |
|
121
|
|
|
|
|
122
|
|
|
success, DICOM = DICOM_validator(file_path) |
|
123
|
|
|
if not success: |
|
124
|
|
|
return False, None |
|
125
|
|
|
from datetime import datetime |
|
126
|
|
|
scan_date = datetime.strptime(DICOM.SeriesDate, "%Y%m%d") |
|
127
|
|
|
birthday = datetime.strptime(DICOM.PatientBirthDate, "%Y%m%d") |
|
128
|
|
|
age = relativedelta(scan_date, birthday) |
|
129
|
|
|
return True, age |
|
130
|
|
|
|
|
131
|
|
|
|
|
132
|
|
|
def DICOM_retrieveMRN(file_path): |
|
133
|
|
|
""" |
|
134
|
|
|
Read the PatientID field which normally used as MRN number. |
|
135
|
|
|
:param file_path: |
|
136
|
|
|
:return: MRN number, as a STRING |
|
137
|
|
|
""" |
|
138
|
|
|
success, DICOM = DICOM_validator(file_path) |
|
139
|
|
|
if not success: |
|
140
|
|
|
return False, None |
|
141
|
|
|
|
|
142
|
|
|
MRN_number = DICOM.PatientID |
|
143
|
|
|
return True, MRN_number |
|
144
|
|
|
|
|
145
|
|
|
|
|
146
|
|
|
def DICOM_TransferSyntaxCheck(transfer_syntax): |
|
147
|
|
|
""" |
|
148
|
|
|
Determine if the transfer syntax symbolize LEE or JPEG compressed! |
|
149
|
|
|
:param transfer_syntax: |
|
150
|
|
|
:return: whether the DICOM files are compressed. |
|
151
|
|
|
""" |
|
152
|
|
|
|
|
153
|
|
|
if not ("1.2.840.10008.1.2" in transfer_syntax): |
|
154
|
|
|
raise ValueError |
|
155
|
|
|
elif transfer_syntax == "1.2.840.10008.1.2" or transfer_syntax[18] == '1' or transfer_syntax[18] == '2': |
|
156
|
|
|
return True |
|
157
|
|
|
elif transfer_syntax[18] == '4' or transfer_syntax[18] == '5' or transfer_syntax[18] == '6': |
|
158
|
|
|
return False |
|
159
|
|
|
else: |
|
160
|
|
|
raise ValueError |
|
161
|
|
|
|
|
162
|
|
|
|
|
163
|
|
|
if __name__ == '__main__': |
|
164
|
|
|
DICOM_decompressor("0000000A") |
|
165
|
|
|
|
|
166
|
|
|
|
|
167
|
|
|
|
CNBP /
DICOMTransit
