sparktk.dicom.dicom module
# vim: set encoding=utf-8
# Copyright (c) 2016 Intel Corporation
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
import logging
logger = logging.getLogger('sparktk')
from sparktk import TkContext
# import constructors for the API's sake (not actually dependencies of the Dicom class)
from sparktk.dicom.constructors.import_dcm import import_dcm
__all__ = ["Dicom",
"import_dcm",
"load"]
class Dicom(object):
"""
sparktk Dicom
Represents a collection of DICOM data objects. Reference: [https://en.wikipedia.org/wiki/DICOM](https://en.wikipedia.org/wiki/DICOM)
The metadata property is a sparktk frame which defines the metadata of the collection of DICOM objects.
Its schema has a column named "id" which holds a unique integer ID for the record and another column which
holds a string of XML comprised of the metadata. Users can run XQuery or invoke canned column extraction/filter
operations on this frame.
The pixeldata property is a sparktk frame which defines the pixeldata of the collection of DICOM objects.
Its schema has a column named "id" which holds a unique integer ID for the record and another column which
holds a matrix(internally it is a numpy.ndarray) comprised of the pixeldata. Users can run numpy supported transformations on it.
dcm4che-3.x dependencies are used to support various operations on dicom images. It is available as java library
Reference: [https://github.com/dcm4che/dcm4che](https://github.com/dcm4che/dcm4che)
Note: Currently sparktk Dicom supports only uncompressed dicom images
Load a set of uncompressed sample .dcm files from path (integration-tests/datasets/dicom_uncompressed)
and create a dicom object. The below examples helps you to understand how to access dicom object properties.
Examples
--------
#Path can be local/hdfs to dcm file(s)
>>> dicom_path = "../datasets/dicom_uncompressed"
#use import_dcm available inside dicom module to create a dicom object from given dicom_path
>>> dicom = tc.dicom.import_dcm(dicom_path)
#Type of dicom object created
>>> type(dicom)
>>> dicom.metadata.count()
3
>>> dicom.pixeldata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0
AAE= 1.2.840.10008.5.1.4.1.1.4 1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685
#Dimesions of the image matrix stored
>>> image_ndarray.shape
(320, 320)
#Use python matplot lib package to verify image visually
>>> import pylab
>>> pylab.imshow(image_ndarray, cmap=pylab.cm.bone)
>>> pylab.show()
#Save method persists the dicom object to disk
>>> dicom.save("sandbox/dicom_data")
#loads the saved dicom object
>>> load_dicom = tc.load("sandbox/dicom_data")
#Re-check whether we loaded back the dicom object or not
>>> type(load_dicom)
#Again access pixeldata and perform same operations as above
>>> load_pixeldata = load_dicom.pixeldata.take(1)
#Order may differ when you load back dicom object
>>> load_pixeldata
[[0L, array([[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 125., 103., ..., 120., 213., 319.],
[ 0., 117., 94., ..., 135., 223., 325.],
...,
[ 0., 62., 21., ..., 896., 886., 854.],
[ 0., 63., 23., ..., 941., 872., 897.],
[ 0., 60., 30., ..., 951., 822., 906.]])]]
>>> load_image_ndarray= load_pixeldata[0][1]
>>> type(load_image_ndarray)
>>> load_image_ndarray.shape
(320, 320)
#Inspect metadata property to see dicom metadata xml content
>>> load_dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 >> import xml.etree.ElementTree as ET
#Performing add_columns operation.
#Add xml tag as column in dicom metadata frame
#Here we add SOPInstanceUID as column to metadaframe
#sample function to apply on row - add_columns
>>> def extractor(tag_name):
... def _extractor(row):
... root = ET.fromstring(row["metadata"])
... for attribute in root.findall('DicomAttribute'):
... keyword = attribute.get('keyword')
... value = None
... if attribute.find('Value') is not None:
... value = attribute.find('Value').text
... if keyword == tag_name:
... return value
... return _extractor
>>> tag_name = "SOPInstanceUID"
>>> dicom.metadata.add_columns(extractor(tag_name), (tag_name, str))
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata SOPInstanceUID
=======================================================================
[0] 0 Functions
def import_dcm(
dicom_dir_path, tc=<class 'sparktk.arguments.implicit'>)
Creates a dicom object with metadataFrame and pixeldataFrame from a dcm file(s)
Parameters:
| dicom_dir_path | (str): | Local/HDFS path of the dcm file(s) |
| Returns | (Dicom): | returns a dicom object with metadata and pixeldata frames |
Examples:
#Path can be local/hdfs to dcm file(s)
>>> dicom_path = "../datasets/dicom_uncompressed"
#use import_dcm available inside dicom module to create a dicom object from given dicom_path
>>> dicom = tc.dicom.import_dcm(dicom_path)
#Type of dicom object created
>>> type(dicom)
<class 'sparktk.dicom.dicom.Dicom'>
#Inspect metadata property to see dicom metadata xml content
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#pixeldata property is sparktk frame
>>> pixeldata = dicom.pixeldata.take(1)
>>> pixeldata
[[0L, array([[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 7., 5., ..., 5., 7., 8.],
[ 0., 7., 6., ..., 5., 6., 7.],
...,
[ 0., 6., 7., ..., 5., 5., 6.],
[ 0., 2., 5., ..., 5., 5., 4.],
[ 1., 1., 3., ..., 1., 1., 0.]])]]
def import_dcm(dicom_dir_path, tc=TkContext.implicit):
"""
Creates a dicom object with metadataFrame and pixeldataFrame from a dcm file(s)
Parameters
----------
:param dicom_dir_path: (str) Local/HDFS path of the dcm file(s)
:return: (Dicom) returns a dicom object with metadata and pixeldata frames
Examples
--------
#Path can be local/hdfs to dcm file(s)
>>> dicom_path = "../datasets/dicom_uncompressed"
#use import_dcm available inside dicom module to create a dicom object from given dicom_path
>>> dicom = tc.dicom.import_dcm(dicom_path)
#Type of dicom object created
>>> type(dicom)
#Inspect metadata property to see dicom metadata xml content
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 >> pixeldata = dicom.pixeldata.take(1)
>>> pixeldata
[[0L, array([[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 7., 5., ..., 5., 7., 8.],
[ 0., 7., 6., ..., 5., 6., 7.],
...,
[ 0., 6., 7., ..., 5., 5., 6.],
[ 0., 2., 5., ..., 5., 5., 4.],
[ 1., 1., 3., ..., 1., 1., 0.]])]]
"""
if not isinstance(dicom_dir_path, basestring):
raise ValueError("dicom_dir_path parameter must be a string, but is {0}.".format(type(dicom_dir_path)))
TkContext.validate(tc)
scala_dicom = tc.sc._jvm.org.trustedanalytics.sparktk.dicom.internal.constructors.Import.importDcm(tc.jutils.get_scala_sc(), dicom_dir_path)
from sparktk.dicom.dicom import Dicom
return Dicom._from_scala(tc, scala_dicom)
def load(
path, tc=<class 'sparktk.arguments.implicit'>)
load Dicom from given path
def load(path, tc=TkContext.implicit):
"""load Dicom from given path"""
TkContext.validate(tc)
return tc.load(path, Dicom)
Classes
class Dicom
sparktk Dicom
Represents a collection of DICOM data objects. Reference: https://en.wikipedia.org/wiki/DICOM
The metadata property is a sparktk frame which defines the metadata of the collection of DICOM objects. Its schema has a column named "id" which holds a unique integer ID for the record and another column which holds a string of XML comprised of the metadata. Users can run XQuery or invoke canned column extraction/filter operations on this frame.
The pixeldata property is a sparktk frame which defines the pixeldata of the collection of DICOM objects. Its schema has a column named "id" which holds a unique integer ID for the record and another column which holds a matrix(internally it is a numpy.ndarray) comprised of the pixeldata. Users can run numpy supported transformations on it.
dcm4che-3.x dependencies are used to support various operations on dicom images. It is available as java library Reference: https://github.com/dcm4che/dcm4che
Note: Currently sparktk Dicom supports only uncompressed dicom images
Load a set of uncompressed sample .dcm files from path (integration-tests/datasets/dicom_uncompressed) and create a dicom object. The below examples helps you to understand how to access dicom object properties.
Examples:
#Path can be local/hdfs to dcm file(s)
>>> dicom_path = "../datasets/dicom_uncompressed"
#use import_dcm available inside dicom module to create a dicom object from given dicom_path
>>> dicom = tc.dicom.import_dcm(dicom_path)
#Type of dicom object created
>>> type(dicom)
<class 'sparktk.dicom.dicom.Dicom'>
>>> dicom.metadata.count()
3
>>> dicom.pixeldata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
#pixeldata property is sparktk frame
>>> pixeldata = dicom.pixeldata.take(1)
#Display
>>> pixeldata
[[0L, array([[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 125., 103., ..., 120., 213., 319.],
[ 0., 117., 94., ..., 135., 223., 325.],
...,
[ 0., 62., 21., ..., 896., 886., 854.],
[ 0., 63., 23., ..., 941., 872., 897.],
[ 0., 60., 30., ..., 951., 822., 906.]])]]
#Access ndarray
>>> image_ndarray= pixeldata[0][1]
>>> type(image_ndarray)
<type 'numpy.ndarray'>
#Dimesions of the image matrix stored
>>> image_ndarray.shape
(320, 320)
#Use python matplot lib package to verify image visually
>>> import pylab
>>> pylab.imshow(image_ndarray, cmap=pylab.cm.bone)
>>> pylab.show()
#Save method persists the dicom object to disk
>>> dicom.save("sandbox/dicom_data")
#loads the saved dicom object
>>> load_dicom = tc.load("sandbox/dicom_data")
#Re-check whether we loaded back the dicom object or not
>>> type(load_dicom)
<class 'sparktk.dicom.dicom.Dicom'>
#Again access pixeldata and perform same operations as above
>>> load_pixeldata = load_dicom.pixeldata.take(1)
#Order may differ when you load back dicom object
>>> load_pixeldata
[[0L, array([[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 125., 103., ..., 120., 213., 319.],
[ 0., 117., 94., ..., 135., 223., 325.],
...,
[ 0., 62., 21., ..., 896., 886., 854.],
[ 0., 63., 23., ..., 941., 872., 897.],
[ 0., 60., 30., ..., 951., 822., 906.]])]]
>>> load_image_ndarray= load_pixeldata[0][1]
>>> type(load_image_ndarray)
<type 'numpy.ndarray'>
>>> load_image_ndarray.shape
(320, 320)
#Inspect metadata property to see dicom metadata xml content
>>> load_dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Using to built-in xml libraries to run xquery on metadata
>>> import xml.etree.ElementTree as ET
#Performing add_columns operation.
#Add xml tag as column in dicom metadata frame
#Here we add SOPInstanceUID as column to metadaframe
#sample function to apply on row - add_columns
>>> def extractor(tag_name):
... def _extractor(row):
... root = ET.fromstring(row["metadata"])
... for attribute in root.findall('DicomAttribute'):
... keyword = attribute.get('keyword')
... value = None
... if attribute.find('Value') is not None:
... value = attribute.find('Value').text
... if keyword == tag_name:
... return value
... return _extractor
>>> tag_name = "SOPInstanceUID"
>>> dicom.metadata.add_columns(extractor(tag_name), (tag_name, str))
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata SOPInstanceUID
=======================================================================
[0] 0 <?xml version="1.0" encodin... 1.3.6.1.4.1.14519.5.2.1.730...
[1] 1 <?xml version="1.0" encodin... 1.3.6.1.4.1.14519.5.2.1.730...
[2] 2 <?xml version="1.0" encodin... 1.3.6.1.4.1.14519.5.2.1.730...
class Dicom(object):
"""
sparktk Dicom
Represents a collection of DICOM data objects. Reference: [https://en.wikipedia.org/wiki/DICOM](https://en.wikipedia.org/wiki/DICOM)
The metadata property is a sparktk frame which defines the metadata of the collection of DICOM objects.
Its schema has a column named "id" which holds a unique integer ID for the record and another column which
holds a string of XML comprised of the metadata. Users can run XQuery or invoke canned column extraction/filter
operations on this frame.
The pixeldata property is a sparktk frame which defines the pixeldata of the collection of DICOM objects.
Its schema has a column named "id" which holds a unique integer ID for the record and another column which
holds a matrix(internally it is a numpy.ndarray) comprised of the pixeldata. Users can run numpy supported transformations on it.
dcm4che-3.x dependencies are used to support various operations on dicom images. It is available as java library
Reference: [https://github.com/dcm4che/dcm4che](https://github.com/dcm4che/dcm4che)
Note: Currently sparktk Dicom supports only uncompressed dicom images
Load a set of uncompressed sample .dcm files from path (integration-tests/datasets/dicom_uncompressed)
and create a dicom object. The below examples helps you to understand how to access dicom object properties.
Examples
--------
#Path can be local/hdfs to dcm file(s)
>>> dicom_path = "../datasets/dicom_uncompressed"
#use import_dcm available inside dicom module to create a dicom object from given dicom_path
>>> dicom = tc.dicom.import_dcm(dicom_path)
#Type of dicom object created
>>> type(dicom)
>>> dicom.metadata.count()
3
>>> dicom.pixeldata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0
AAE= 1.2.840.10008.5.1.4.1.1.4 1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685
#Dimesions of the image matrix stored
>>> image_ndarray.shape
(320, 320)
#Use python matplot lib package to verify image visually
>>> import pylab
>>> pylab.imshow(image_ndarray, cmap=pylab.cm.bone)
>>> pylab.show()
#Save method persists the dicom object to disk
>>> dicom.save("sandbox/dicom_data")
#loads the saved dicom object
>>> load_dicom = tc.load("sandbox/dicom_data")
#Re-check whether we loaded back the dicom object or not
>>> type(load_dicom)
#Again access pixeldata and perform same operations as above
>>> load_pixeldata = load_dicom.pixeldata.take(1)
#Order may differ when you load back dicom object
>>> load_pixeldata
[[0L, array([[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 125., 103., ..., 120., 213., 319.],
[ 0., 117., 94., ..., 135., 223., 325.],
...,
[ 0., 62., 21., ..., 896., 886., 854.],
[ 0., 63., 23., ..., 941., 872., 897.],
[ 0., 60., 30., ..., 951., 822., 906.]])]]
>>> load_image_ndarray= load_pixeldata[0][1]
>>> type(load_image_ndarray)
>>> load_image_ndarray.shape
(320, 320)
#Inspect metadata property to see dicom metadata xml content
>>> load_dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 >> import xml.etree.ElementTree as ET
#Performing add_columns operation.
#Add xml tag as column in dicom metadata frame
#Here we add SOPInstanceUID as column to metadaframe
#sample function to apply on row - add_columns
>>> def extractor(tag_name):
... def _extractor(row):
... root = ET.fromstring(row["metadata"])
... for attribute in root.findall('DicomAttribute'):
... keyword = attribute.get('keyword')
... value = None
... if attribute.find('Value') is not None:
... value = attribute.find('Value').text
... if keyword == tag_name:
... return value
... return _extractor
>>> tag_name = "SOPInstanceUID"
>>> dicom.metadata.add_columns(extractor(tag_name), (tag_name, str))
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata SOPInstanceUID
=======================================================================
[0] 0 Ancestors (in MRO)
- Dicom
- __builtin__.object
Instance variables
var metadata
var pixeldata
Methods
def __init__(
self, tc, scala_dicom)
def __init__(self, tc, scala_dicom):
self._tc = tc
from sparktk.frame.frame import Frame
self._metadata = Frame(self._tc, scala_dicom.metadata())
self._pixeldata = Frame(self._tc, scala_dicom.pixeldata())
def drop_rows(
self, predicate)
Drop the rows of dicom metadata and pixeldata frames using given predicate
Parameters:
| predicate: | predicate to apply on filter |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
>>> import xml.etree.ElementTree as ET
#sample custom filter function
>>> def drop_meta(tag_name, tag_value):
... def _drop_meta(row):
... root = ET.fromstring(row["metadata"])
... for attribute in root.findall('DicomAttribute'):
... keyword = attribute.get('keyword')
... if attribute.get('keyword') is not None:
... if attribute.find('Value') is not None:
... value = attribute.find('Value').text
... if keyword == tag_name and value == tag_value:
... return True
... return _drop_meta
>>> tag_name = "SOPInstanceUID"
>>> tag_value = "1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685"
>>> dicom.drop_rows(drop_meta(tag_name, tag_value))
>>> dicom.metadata.count()
2
#After filter
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 1 <?xml version="1.0" encodin...
[1] 2 <?xml version="1.0" encodin...
>>> dicom.pixeldata.inspect(truncate=30)
[#] id imagematrix
===========================================================
[0] 1 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 70. 85. ..., 215. 288. 337.]
[ 0. 63. 72. ..., 228. 269. 317.]
...,
[ 0. 42. 40. ..., 966. 919. 871.]
[ 0. 42. 33. ..., 988. 887. 860.]
[ 0. 46. 38. ..., 983. 876. 885.]]
[1] 2 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 111. 117. ..., 159. 148. 135.]
[ 0. 116. 111. ..., 152. 138. 139.]
...,
[ 0. 49. 18. ..., 1057. 965. 853.]
[ 0. 42. 20. ..., 1046. 973. 891.]
[ 0. 48. 26. ..., 1041. 969. 930.]]
def drop_rows(self, predicate):
"""
Drop the rows of dicom metadata and pixeldata frames using given predicate
Parameters
----------
:param predicate: predicate to apply on filter
Examples
--------
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0
AAE= 1.2.840.10008.5.1.4.1.1.4 1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685 def drop_rows_by_keywords(
self, keywords_values_dict)
Drop the rows based on dictionary of {"keyword":"value"}(applying 'and' operation on dictionary) from column holding xml string.
Ex: keywords_values_dict -> {"SOPInstanceUID":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "Manufacturer":"SIEMENS", "StudyDate":"20030315"}
Parameters:
| keywords_values_dict | (dict(str, str)): | dictionary of keywords and values from xml string in metadata |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
>>> keywords_values_dict = {"SOPInstanceUID":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "Manufacturer":"SIEMENS", "StudyDate":"20030315"}
>>> dicom.drop_rows_by_keywords(keywords_values_dict)
>>> dicom.metadata.count()
2
#After drop_rows
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 1 <?xml version="1.0" encodin...
[1] 2 <?xml version="1.0" encodin...
>>> dicom.pixeldata.inspect(truncate=30)
[#] id imagematrix
===========================================================
[0] 1 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 70. 85. ..., 215. 288. 337.]
[ 0. 63. 72. ..., 228. 269. 317.]
...,
[ 0. 42. 40. ..., 966. 919. 871.]
[ 0. 42. 33. ..., 988. 887. 860.]
[ 0. 46. 38. ..., 983. 876. 885.]]
[1] 2 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 111. 117. ..., 159. 148. 135.]
[ 0. 116. 111. ..., 152. 138. 139.]
...,
[ 0. 49. 18. ..., 1057. 965. 853.]
[ 0. 42. 20. ..., 1046. 973. 891.]
[ 0. 48. 26. ..., 1041. 969. 930.]]
def drop_rows_by_keywords(self, keywords_values_dict):
"""
Drop the rows based on dictionary of {"keyword":"value"}(applying 'and' operation on dictionary) from column holding xml string.
Ex: keywords_values_dict -> {"SOPInstanceUID":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "Manufacturer":"SIEMENS", "StudyDate":"20030315"}
Parameters
----------
:param keywords_values_dict: (dict(str, str)) dictionary of keywords and values from xml string in metadata
Examples
--------
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0
AAE= 1.2.840.10008.5.1.4.1.1.4 1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685 ")
#Always scala dicom is invoked, as python joins are expensive compared to serailizations.
def f(scala_dicom):
scala_dicom.dropRowsByKeywords(self._tc.jutils.convert.to_scala_map(keywords_values_dict))
self._call_scala(f)
def drop_rows_by_tags(
self, tags_values_dict)
Drop the rows based on dictionary of {"tag":"value"}(applying 'and' operation on dictionary) from column holding xml string
Ex: tags_values_dict -> {"00080018":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "00080070":"SIEMENS", "00080020":"20030315"}
Parameters:
| tags_values_dict | (dict(str, str)): | dictionary of tags and values from xml string in metadata |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
>>> tags_values_dict = {"00080018":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "00080070":"SIEMENS", "00080020":"20030315"}
>>> dicom.drop_rows_by_tags(tags_values_dict)
>>> dicom.metadata.count()
2
#After drop_rows
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 1 <?xml version="1.0" encodin...
[1] 2 <?xml version="1.0" encodin...
>>> dicom.pixeldata.inspect(truncate=30)
[#] id imagematrix
===========================================================
[0] 1 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 70. 85. ..., 215. 288. 337.]
[ 0. 63. 72. ..., 228. 269. 317.]
...,
[ 0. 42. 40. ..., 966. 919. 871.]
[ 0. 42. 33. ..., 988. 887. 860.]
[ 0. 46. 38. ..., 983. 876. 885.]]
[1] 2 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 111. 117. ..., 159. 148. 135.]
[ 0. 116. 111. ..., 152. 138. 139.]
...,
[ 0. 49. 18. ..., 1057. 965. 853.]
[ 0. 42. 20. ..., 1046. 973. 891.]
[ 0. 48. 26. ..., 1041. 969. 930.]]
def export_to_dcm(
self, path)
export_to_dcm creates .dcm image from dicom object with (metadata, imagedata) and saves to given path
Parameters:
| path | (str): | local/hdfs path |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.pixeldata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
#pixeldata property is sparktk frame
>>> pixeldata = dicom.pixeldata.take(1)
#dispaly
>>> pixeldata
[[0L, array([[ 0., 0., 0., ..., 0., 0., 0.],
[ 0., 125., 103., ..., 120., 213., 319.],
[ 0., 117., 94., ..., 135., 223., 325.],
...,
[ 0., 62., 21., ..., 896., 886., 854.],
[ 0., 63., 23., ..., 941., 872., 897.],
[ 0., 60., 30., ..., 951., 822., 906.]])]]
>>> dicom.export_to_dcm("dicom_export")
def export_to_dcm(self, path):
"""
export_to_dcm creates .dcm image from dicom object with (metadata, imagedata) and saves to given path
Parameters
----------
:param path: (str) local/hdfs path
Examples
--------
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.pixeldata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0
AAE= 1.2.840.10008.5.1.4.1.1.4 1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685 def extract_keywords(
self, keywords)
Extracts value for each keyword from column holding xml string and adds column for each keyword to assign value For missing keyword, the value is None
Ex: keywords -> ["PatientID"]
Parameters:
| keywords | (str or list(str)): | List of keywords from metadata xml string |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
#Extract values for given keywords and add as new columns in metadata frame
>>> dicom.extract_keywords(["SOPInstanceUID", "Manufacturer", "StudyDate"])
>>> dicom.metadata.count()
3
>>> dicom.metadata.column_names
[u'id', u'metadata', u'SOPInstanceUID', u'Manufacturer', u'StudyDate']
>>> dicom.metadata.inspect(truncate=20)
[#] id metadata SOPInstanceUID Manufacturer StudyDate
============================================================================
[0] 0 <?xml version="1.... 1.3.6.1.4.1.14519... SIEMENS 20030315
[1] 1 <?xml version="1.... 1.3.6.1.4.1.14519... SIEMENS 20030315
[2] 2 <?xml version="1.... 1.3.6.1.4.1.14519... SIEMENS 20030315
def extract_keywords(self, keywords):
"""
Extracts value for each keyword from column holding xml string and adds column for each keyword to assign value
For missing keyword, the value is None
Ex: keywords -> ["PatientID"]
Parameters
----------
:param keywords: (str or list(str)) List of keywords from metadata xml string
Examples
--------
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0
AAE= 1.2.840.10008.5.1.4.1.1.4 1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685 def extract_tags(
self, tags)
Extracts value for each tag from column holding xml string and adds column for each tag to assign value. For missing tag, the value is None
Ex: tags -> ["00020001", "00020002"]
Parameters:
| tags | (str or list(str)): | List of tags from xml string of metadata column |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
#Extract value for each tag from column holding xml string
>>> dicom.extract_tags(["00080018", "00080070", "00080030"])
>>> dicom.metadata.count()
3
>>> dicom.metadata.column_names
[u'id', u'metadata', u'00080018', u'00080070', u'00080030']
>>> dicom.metadata.inspect(truncate=20)
[#] id metadata 00080018 00080070 00080030
============================================================================
[0] 0 <?xml version="1.... 1.3.6.1.4.1.14519... SIEMENS 20030315
[1] 1 <?xml version="1.... 1.3.6.1.4.1.14519... SIEMENS 20030315
[2] 2 <?xml version="1.... 1.3.6.1.4.1.14519... SIEMENS 20030315
def filter(
self, predicate)
Filter the rows of dicom metadata and pixeldata based on given predicate
Parameters:
| predicate: | predicate to apply on filter |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
>>> import xml.etree.ElementTree as ET
#sample custom filter function
>>> def filter_meta(tag_name, tag_value):
... def _filter_meta(row):
... root = ET.fromstring(row["metadata"])
... for attribute in root.findall('DicomAttribute'):
... keyword = attribute.get('keyword')
... if attribute.get('keyword') is not None:
... if attribute.find('Value') is not None:
... value = attribute.find('Value').text
... if keyword == tag_name and value == tag_value:
... return True
... return _filter_meta
>>> tag_name = "SOPInstanceUID"
>>> tag_value = "1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685"
>>> dicom.filter(filter_meta(tag_name, tag_value))
>>> dicom.metadata.count()
1
#After filter
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
>>> dicom.pixeldata.inspect(truncate=30)
[#] id imagematrix
=====================================================
[0] 0 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 125. 103. ..., 120. 213. 319.]
[ 0. 117. 94. ..., 135. 223. 325.]
...,
[ 0. 62. 21. ..., 896. 886. 854.]
[ 0. 63. 23. ..., 941. 872. 897.]
[ 0. 60. 30. ..., 951. 822. 906.]]
def filter(self, predicate):
"""
Filter the rows of dicom metadata and pixeldata based on given predicate
Parameters
----------
:param predicate: predicate to apply on filter
Examples
--------
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0
AAE= 1.2.840.10008.5.1.4.1.1.4 1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685 def filter_by_keywords(
self, keywords_values_dict)
Filter the rows based on dictionary of {"keyword":"value"}(applying 'and' operation on dictionary) from column holding xml string
Ex: keywords_values_dict -> {"SOPInstanceUID":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "Manufacturer":"SIEMENS", "StudyDate":"20030315"}
Parameters:
| keywords_values_dict | (dict(str, str)): | dictionary of keywords and values from xml string in metadata |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
>>> keywords_values_dict = {"SOPInstanceUID":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "Manufacturer":"SIEMENS", "StudyDate":"20030315"}
>>> dicom.filter_by_keywords(keywords_values_dict)
>>> dicom.metadata.count()
1
#After filter
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
>>> dicom.pixeldata.inspect(truncate=30)
[#] id imagematrix
=====================================================
[0] 0 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 125. 103. ..., 120. 213. 319.]
[ 0. 117. 94. ..., 135. 223. 325.]
...,
[ 0. 62. 21. ..., 896. 886. 854.]
[ 0. 63. 23. ..., 941. 872. 897.]
[ 0. 60. 30. ..., 951. 822. 906.]]
def filter_by_keywords(self, keywords_values_dict):
"""
Filter the rows based on dictionary of {"keyword":"value"}(applying 'and' operation on dictionary) from column holding xml string
Ex: keywords_values_dict -> {"SOPInstanceUID":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "Manufacturer":"SIEMENS", "StudyDate":"20030315"}
Parameters
----------
:param keywords_values_dict: (dict(str, str)) dictionary of keywords and values from xml string in metadata
Examples
--------
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0
AAE= 1.2.840.10008.5.1.4.1.1.4 1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685 ")
#Always scala dicom is invoked, as python joins are expensive compared to serailizations.
def f(scala_dicom):
scala_dicom.filterByKeywords(self._tc.jutils.convert.to_scala_map(keywords_values_dict))
self._call_scala(f)
def filter_by_tags(
self, tags_values_dict)
Filter the rows based on dictionary of {"tag":"value"}(applying 'and' operation on dictionary) from column holding xml string
Ex: tags_values_dict -> {"00080018":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "00080070":"SIEMENS", "00080020":"20030315"}
Parameters:
| tags_values_dict | (dict(str, str)): | dictionary of tags and values from xml string in metadata. |
Examples:
>>> dicom_path = "../datasets/dicom_uncompressed"
>>> dicom = tc.dicom.import_dcm(dicom_path)
>>> dicom.metadata.count()
3
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
[1] 1 <?xml version="1.0" encodin...
[2] 2 <?xml version="1.0" encodin...
#Part of xml string looks as below
<?xml version="1.0" encoding="UTF-8"?>
<NativeDicomModel xml:space="preserve">
<DicomAttribute keyword="FileMetaInformationVersion" tag="00020001" vr="OB"><InlineBinary>AAE=</InlineBinary></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPClassUID" tag="00020002" vr="UI"><Value number="1">1.2.840.10008.5.1.4.1.1.4</Value></DicomAttribute>
<DicomAttribute keyword="MediaStorageSOPInstanceUID" tag="00020003" vr="UI"><Value number="1">1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685</Value></DicomAttribute>
...
>>> tags_values_dict = {"00080018":"1.3.6.1.4.1.14519.5.2.1.7308.2101.234736319276602547946349519685", "00080070":"SIEMENS", "00080020":"20030315"}
>>> dicom.filter_by_tags(tags_values_dict)
>>> dicom.metadata.count()
1
#After filter
>>> dicom.metadata.inspect(truncate=30)
[#] id metadata
=======================================
[0] 0 <?xml version="1.0" encodin...
>>> dicom.pixeldata.inspect(truncate=30)
[#] id imagematrix
=====================================================
[0] 0 [[ 0. 0. 0. ..., 0. 0. 0.]
[ 0. 125. 103. ..., 120. 213. 319.]
[ 0. 117. 94. ..., 135. 223. 325.]
...,
[ 0. 62. 21. ..., 896. 886. 854.]
[ 0. 63. 23. ..., 941. 872. 897.]
[ 0. 60. 30. ..., 951. 822. 906.]]
def save(
self, path)
Persists the dicom object to the given file path
def save(self, path):
"""Persists the dicom object to the given file path"""
self._get_new_scala().save(path)