def test_lt_diff_class(self):
"""Test __lt__ of two classes with different type."""
self.assertTrue(BaseTag(0x00000000) < 1)
self.assertFalse(BaseTag(0x00000001) < 1)
self.assertFalse(BaseTag(0x00000001) < 0)
def test_le_tuple(self):
"""Test __le__ of tuple with BaseTag."""
assert BaseTag(0x00010001) <= (0x0001, 0x0002)
assert BaseTag(0x00010002) <= (0x0001, 0x0002)
assert not BaseTag(0x00010002) <= (0x0001, 0x0001)
def test_raise():
BaseTag(0x00010002) <= 'Somethin'
def test_hash(self):
"""Test hash of BaseTag class."""
assert hash(BaseTag(0x00010001)) == hash(BaseTag(0x00010001))
assert hash(BaseTag(0x00010001)) != hash(BaseTag(0x00010002))
assert hash(BaseTag(0x00020001)) != hash(BaseTag(0x00010002))
def test_group(self):
"""Test BaseTag.group returns correct values."""
assert BaseTag(0x00000001).group == 0x0000
assert BaseTag(0x00020001).group == 0x0002
assert BaseTag(0xFFFF0001).group == 0xFFFF
def test_ne_same_class(self):
"""Test __ne__ of two classes with same type."""
assert not BaseTag(0x00000000) != BaseTag(0x00000000)
assert BaseTag(0x00000001) != BaseTag(0x00000000)
def test_ne_diff_class(self):
"""Test __ne__ of two classes with different type."""
assert not BaseTag(0x00000000) != 0
assert BaseTag(0x00000001) != 0
def test_group(self):
"""Test BaseTag.group returns correct values."""
assert 0x0000 == BaseTag(0x00000001).group
assert 0x0002 == BaseTag(0x00020001).group
assert 0xFFFF == BaseTag(0xFFFF0001).group
def read_dataset(fp,
is_implicit_VR,
is_little_endian,
bytelength=None,
stop_when=None,
defer_size=None,
parent_encoding=default_encoding,
specific_tags=None,
at_top_level=True):
"""Return a :class:`~pydicom.dataset.Dataset` instance containing the next
dataset in the file.
Parameters
----------
fp : file-like
An opened file-like object.
is_implicit_VR : bool
``True`` if file transfer syntax is implicit VR.
is_little_endian : bool
``True`` if file has little endian transfer syntax.
bytelength : int, None, optional
``None`` to read until end of file or ItemDeliterTag, else a fixed
number of bytes to read
stop_when : None, optional
Optional call_back function which can terminate reading. See help for
:func:`data_element_generator` for details
defer_size : int, None, optional
Size to avoid loading large elements in memory. See :func:`dcmread` for
more parameter info.
parent_encoding :
Optional encoding to use as a default in case (0008,0005) *Specific
Character Set* isn't specified.
specific_tags : list or None
See :func:`dcmread` for parameter info.
at_top_level: bool
If dataset is top level (not within a sequence).
Used to turn off explicit VR heuristic within sequences
Returns
-------
dataset.Dataset
A Dataset instance.
See Also
--------
dataset.Dataset
A collection (dictionary) of DICOM ``DataElement`` instances.
"""
raw_data_elements = dict()
fp_start = fp.tell()
if at_top_level:
is_implicit_VR = _is_implicit_vr(fp, is_implicit_VR, is_little_endian,
stop_when)
fp.seek(fp_start)
de_gen = data_element_generator(fp, is_implicit_VR, is_little_endian,
stop_when, defer_size, parent_encoding,
specific_tags)
try:
while (bytelength is None) or (fp.tell() - fp_start < bytelength):
raw_data_element = next(de_gen)
# Read data elements. Stop on some errors, but return what was read
tag = raw_data_element.tag
# Check for ItemDelimiterTag --dataset is an item in a sequence
if tag == BaseTag(0xFFFEE00D):
break
raw_data_elements[tag] = raw_data_element
except StopIteration:
pass
except EOFError as details:
if config.enforce_valid_values:
raise
msg = str(details) + " in file " + getattr(fp, "name", "<no filename>")
warnings.warn(msg, UserWarning)
except NotImplementedError as details:
logger.error(details)
ds = Dataset(raw_data_elements)
if 0x00080005 in raw_data_elements:
char_set = DataElement_from_raw(raw_data_elements[0x00080005])
encoding = convert_encodings(char_set)
else:
encoding = parent_encoding
ds.set_original_encoding(is_implicit_VR, is_little_endian, encoding)
return ds
def test_ne_diff_class(self):
"""Test __ne__ of two classes with different type."""
# Make sure to test BaseTag.__ne__() not int.__ne__()
assert not BaseTag(0x00000000) != 0
assert BaseTag(0x00000001) != 0
def test_ne_tuple(self):
"""Test __ne__ of tuple with BaseTag."""
# Make sure to test BaseTag.__ne__() not tuple.__ne__()
assert not BaseTag(0x00010002) != (0x0001, 0x0002)
assert BaseTag(0x00010001) != (0x0001, 0x0002)
def test_ge_diff_class(self):
"""Test __ge__ of two classes with different type."""
self.assertFalse(BaseTag(0x00000000) >= 1)
self.assertTrue(BaseTag(0x00000001) >= 1)
self.assertTrue(BaseTag(0x00000001) >= 0)
def test_ge_same_class(self):
"""Test __ge__ of two classes with same type."""
self.assertFalse(BaseTag(0x00000000) >= BaseTag(0x00000001))
self.assertTrue(BaseTag(0x00000001) >= BaseTag(0x00000001))
self.assertTrue(BaseTag(0x00000001) >= BaseTag(0x00000000))
def test_lt_tuple(self):
"""Test __lt__ of tuple with BaseTag."""
self.assertTrue(BaseTag(0x00010001) < (0x0001, 0x0002))
self.assertFalse(BaseTag(0x00010002) < (0x0001, 0x0002))
self.assertFalse(BaseTag(0x00010002) < (0x0001, 0x0001))
def test_eq_tuple(self):
"""Test __eq__ of tuple with BaseTag."""
assert BaseTag(0x00010002) == (0x0001, 0x0002)
assert not BaseTag(0x00010001) == (0x0001, 0x0002)
def read_dataset(fp,
is_implicit_VR,
is_little_endian,
bytelength=None,
stop_when=None,
defer_size=None,
parent_encoding=default_encoding,
specific_tags=None):
"""Return a Dataset instance containing the next dataset in the file.
Parameters
----------
fp : an opened file object
is_implicit_VR : boolean
True if file transfer syntax is implicit VR.
is_little_endian : boolean
True if file has little endian transfer syntax.
bytelength : int, None, optional
None to read until end of file or ItemDeliterTag, else
a fixed number of bytes to read
stop_when : None, optional
optional call_back function which can terminate reading.
See help for data_element_generator for details
defer_size : int, None, optional
Size to avoid loading large elements in memory.
See ``dcmread`` for more parameter info.
parent_encoding :
optional encoding to use as a default in case
a Specific Character Set (0008,0005) isn't specified
specific_tags : list or None
See ``dcmread`` for parameter info.
Returns
-------
a Dataset instance
See Also
--------
pydicom.dataset.Dataset
A collection (dictionary) of Dicom `DataElement` instances.
"""
raw_data_elements = dict()
fp_start = fp.tell()
is_implicit_VR = _is_implicit_vr(fp, is_implicit_VR, is_little_endian,
stop_when)
fp.seek(fp_start)
de_gen = data_element_generator(fp, is_implicit_VR, is_little_endian,
stop_when, defer_size, parent_encoding,
specific_tags)
try:
while (bytelength is None) or (fp.tell() - fp_start < bytelength):
raw_data_element = next(de_gen)
# Read data elements. Stop on some errors, but return what was read
tag = raw_data_element.tag
# Check for ItemDelimiterTag --dataset is an item in a sequence
if tag == BaseTag(0xFFFEE00D):
break
raw_data_elements[tag] = raw_data_element
except StopIteration:
pass
except EOFError as details:
# XXX is this error visible enough to user code with just logging?
logger.error(
str(details) + " in file " + getattr(fp, "name", "<no filename>"))
except NotImplementedError as details:
logger.error(details)
ds = Dataset(raw_data_elements)
if 0x00080005 in raw_data_elements:
char_set = DataElement_from_raw(raw_data_elements[0x00080005])
encoding = convert_encodings(char_set)
else:
encoding = parent_encoding
ds.set_original_encoding(is_implicit_VR, is_little_endian, encoding)
return ds
def test_raise():
BaseTag(0x00010002) == 'eraa'
def data_element_generator(fp,
is_implicit_VR,
is_little_endian,
stop_when=None,
defer_size=None,
encoding=default_encoding,
specific_tags=None):
"""Create a generator to efficiently return the raw data elements.
Parameters
----------
fp : file-like object
is_implicit_VR : boolean
is_little_endian : boolean
stop_when : None, callable, optional
If None (default), then the whole file is read.
A callable which takes tag, VR, length,
and returns True or False. If it returns True,
read_data_element will just return.
defer_size : int, str, None, optional
See ``dcmread`` for parameter info.
encoding :
Encoding scheme
specific_tags : list or None
See ``dcmread`` for parameter info.
Returns
-------
VR : None if implicit VR, otherwise the VR read from the file
length :
the length as in the DICOM data element (could be
DICOM "undefined length" 0xffffffffL)
value_bytes :
the raw bytes from the DICOM file
(not parsed into python types)
is_little_endian : boolean
True if transfer syntax is little endian; else False.
"""
# Summary of DICOM standard PS3.5-2008 chapter 7:
# If Implicit VR, data element is:
# tag, 4-byte length, value.
# The 4-byte length can be FFFFFFFF (undefined length)*
#
# If Explicit VR:
# if OB, OW, OF, SQ, UN, or UT:
# tag, VR, 2-bytes reserved (both zero), 4-byte length, value
# For all but UT, the length can be FFFFFFFF (undefined length)*
# else: (any other VR)
# tag, VR, (2 byte length), value
# * for undefined length, a Sequence Delimitation Item marks the end
# of the Value Field.
# Note, except for the special_VRs, both impl and expl VR use 8 bytes;
# the special VRs follow the 8 bytes with a 4-byte length
# With a generator, state is stored, so we can break down
# into the individual cases, and not have to check them again for each
# data element
if is_little_endian:
endian_chr = "<"
else:
endian_chr = ">"
if is_implicit_VR:
element_struct = Struct(endian_chr + "HHL")
else: # Explicit VR
# tag, VR, 2-byte length (or 0 if special VRs)
element_struct = Struct(endian_chr + "HH2sH")
extra_length_struct = Struct(endian_chr + "L") # for special VRs
extra_length_unpack = extra_length_struct.unpack # for lookup speed
# Make local variables so have faster lookup
fp_read = fp.read
fp_tell = fp.tell
logger_debug = logger.debug
debugging = config.debugging
element_struct_unpack = element_struct.unpack
defer_size = size_in_bytes(defer_size)
tag_set = set()
if specific_tags is not None:
for tag in specific_tags:
if isinstance(tag, (str, compat.text_type)):
tag = Tag(tag_for_keyword(tag))
if isinstance(tag, BaseTag):
tag_set.add(tag)
tag_set.add(Tag(0x08, 0x05))
has_tag_set = len(tag_set) > 0
while True:
# Read tag, VR, length, get ready to read value
bytes_read = fp_read(8)
if len(bytes_read) < 8:
return # at end of file
if debugging:
debug_msg = "{0:08x}: {1}".format(fp.tell() - 8,
bytes2hex(bytes_read))
if is_implicit_VR:
# must reset VR each time; could have set last iteration (e.g. SQ)
VR = None
group, elem, length = element_struct_unpack(bytes_read)
else: # explicit VR
group, elem, VR, length = element_struct_unpack(bytes_read)
if not in_py2:
VR = VR.decode(default_encoding)
if VR in extra_length_VRs:
bytes_read = fp_read(4)
length = extra_length_unpack(bytes_read)[0]
if debugging:
debug_msg += " " + bytes2hex(bytes_read)
if debugging:
debug_msg = "%-47s (%04x, %04x)" % (debug_msg, group, elem)
if not is_implicit_VR:
debug_msg += " %s " % VR
if length != 0xFFFFFFFF:
debug_msg += "Length: %d" % length
else:
debug_msg += "Length: Undefined length (FFFFFFFF)"
logger_debug(debug_msg)
# Positioned to read the value, but may not want to -- check stop_when
value_tell = fp_tell()
tag = TupleTag((group, elem))
if stop_when is not None:
# XXX VR may be None here!! Should stop_when just take tag?
if stop_when(tag, VR, length):
if debugging:
logger_debug("Reading ended by stop_when callback. "
"Rewinding to start of data element.")
rewind_length = 8
if not is_implicit_VR and VR in extra_length_VRs:
rewind_length += 4
fp.seek(value_tell - rewind_length)
return
# Reading the value
# First case (most common): reading a value with a defined length
if length != 0xFFFFFFFF:
# don't defer loading of Specific Character Set value as it is
# needed immediately to get the character encoding for other tags
if has_tag_set and tag not in tag_set:
# skip the tag if not in specific tags
fp.seek(fp_tell() + length)
continue
if (defer_size is not None and length > defer_size
and tag != BaseTag(0x00080005)):
# Flag as deferred by setting value to None, and skip bytes
value = None
logger_debug("Defer size exceeded. "
"Skipping forward to next data element.")
fp.seek(fp_tell() + length)
else:
value = fp_read(length)
if debugging:
dotdot = " "
if length > 12:
dotdot = "..."
logger_debug("%08x: %-34s %s %r %s" %
(value_tell, bytes2hex(
value[:12]), dotdot, value[:12], dotdot))
# If the tag is (0008,0005) Specific Character Set, then store it
if tag == BaseTag(0x00080005):
from pydicom.values import convert_string
encoding = convert_string(value, is_little_endian)
# Store the encoding value in the generator
# for use with future elements (SQs)
encoding = convert_encodings(encoding)
yield RawDataElement(tag, VR, length, value, value_tell,
is_implicit_VR, is_little_endian)
# Second case: undefined length - must seek to delimiter,
# unless is SQ type, in which case is easier to parse it, because
# undefined length SQs and items of undefined lengths can be nested
# and it would be error-prone to read to the correct outer delimiter
else:
# Try to look up type to see if is a SQ
# if private tag, won't be able to look it up in dictionary,
# in which case just ignore it and read the bytes unless it is
# identified as a Sequence
if VR is None:
try:
VR = dictionary_VR(tag)
except KeyError:
# Look ahead to see if it consists of items
# and is thus a SQ
next_tag = TupleTag(unpack(endian_chr + "HH", fp_read(4)))
# Rewind the file
fp.seek(fp_tell() - 4)
if next_tag == ItemTag:
VR = 'SQ'
if VR == 'SQ':
if debugging:
msg = "{0:08x}: Reading/parsing undefined length sequence"
logger_debug(msg.format(fp_tell()))
seq = read_sequence(fp, is_implicit_VR, is_little_endian,
length, encoding)
if has_tag_set and tag not in tag_set:
continue
yield DataElement(tag,
VR,
seq,
value_tell,
is_undefined_length=True)
else:
delimiter = SequenceDelimiterTag
if debugging:
logger_debug("Reading undefined length data element")
value = read_undefined_length_value(fp, is_little_endian,
delimiter, defer_size)
# If the tag is (0008,0005) Specific Character Set,
# then store it
if tag == (0x08, 0x05):
from pydicom.values import convert_string
encoding = convert_string(value, is_little_endian)
# Store the encoding value in the generator for use
# with future elements (SQs)
encoding = convert_encodings(encoding)
# tags with undefined length are skipped after read
if has_tag_set and tag not in tag_set:
continue
yield RawDataElement(tag, VR, length, value, value_tell,
is_implicit_VR, is_little_endian)
def test_le_diff_class(self):
"""Test __le__ of two classes with different type."""
assert BaseTag(0x00000000) <= 1
assert BaseTag(0x00000001) <= 1
assert not BaseTag(0x00000001) <= 0
def test_raise():
BaseTag(0x00010002) >= 'AGHIJJJJ'
def test_raise():
BaseTag(0x00010002) != 'aaag'
def test_gt_diff_class(self):
"""Test __gt__ of two classes with different type."""
assert not BaseTag(0x00000000) > 1
assert not BaseTag(0x00000001) > 1
assert BaseTag(0x00000001) > 0
def test_str(self):
"""Test str(BaseTag) produces correct value."""
assert str(BaseTag(0x00000000)) == '(0000, 0000)'
assert str(BaseTag(0x00010002)) == '(0001, 0002)'
assert str(BaseTag(0x10002000)) == '(1000, 2000)'
assert str(BaseTag(0xFFFFFFFE)) == '(ffff, fffe)'
def test_gt_tuple(self):
"""Test __gt__ of tuple with BaseTag."""
assert not BaseTag(0x00010001) > (0x0001, 0x0002)
assert not BaseTag(0x00010002) > (0x0001, 0x0002)
assert BaseTag(0x00010002) > (0x0001, 0x0001)
def test_element(self):
"""Test BaseTag.element returns correct values."""
assert BaseTag(0x00010000).element == 0x0000
assert BaseTag(0x00010002).element == 0x0002
assert BaseTag(0x0001FFFF).element == 0xFFFF
def test_raise():
BaseTag(0x00010002) > 'BLUH'
def test_tuple_tag(self):
"""Test quick tag construction with TupleTag."""
assert TupleTag((0xFFFF, 0xFFee)) == BaseTag(0xFFFFFFEE)
def test_eq_same_class(self):
"""Test __eq__ of two classes with same type."""
assert BaseTag(0x00000000) == BaseTag(0x00000000)
assert not BaseTag(0x00000001) == BaseTag(0x00000000)
def test_lt_same_class(self):
"""Test __lt__ of two classes with same type."""
assert BaseTag(0x00000000) < BaseTag(0x00000001)
assert not BaseTag(0x00000001) < BaseTag(0x00000001)
assert not BaseTag(0x00000001) < BaseTag(0x00000000)
def test_tuple_tag(self):
"""Test quick tag construction with TupleTag."""
self.assertEqual(TupleTag((0xFFFF, 0xFFee)), BaseTag(0xFFFFFFEE))
