Represent the content of an annotation.
sppasLabel allows to store a set of sppasTags with their scores. This class is using a list of lists, i.e. a list of pairs (tag, score). This is the best compromise between memory usage, speed and readability.
A label is a list of possible sppasTag(), represented as a UNICODE string. A data type can be associated, as sppasTag() can be 'int', 'float' or 'bool'.
Create a new sppasLabel instance.
def __init__(self, tag, score=None):
"""Create a new sppasLabel instance.
:param tag: (sppasTag or list of sppasTag or None)
:param score: (float or list of float or None)
"""
self.__key = None
self.__tags = None
if tag is not None:
if isinstance(tag, list):
if isinstance(score, list) and len(tag) == len(score):
for t, s in zip(tag, score):
self.append(t, s)
else:
for t in tag:
self.append(t)
else:
if isinstance(score, list):
score = None
self.append(tag, score)
Return the label identifier key.
def get_key(self):
"""Return the label identifier key."""
return self.__key
Set the label key.
def set_key(self, key=''):
"""Set the label key.
:param key: (str or None) A label key that can be used to identify the label.
"""
if key is None:
self.__key = None
else:
try:
key = str(key)
if len(key) == 0:
self.__key = None
else:
self.__key = key
except:
raise AnnDataTypeError(key, 'str')
Add a text into the list.
def append_content(self, content, data_type='str', score=None):
"""Add a text into the list.
:param content: (str)
:param data_type: (str): The type of this text content.\\
One of: (str, int, float, bool)
:param score: (float)
"""
tag = sppasTag(content, data_type)
self.append(tag, score)
Add a sppasTag into the list.
def append(self, tag, score=None, add=True):
"""Add a sppasTag into the list.
:param tag: (sppasTag)
:param score: (float)
:param add: (bool) Do not add the tag if this alternative is already
inside the list, but add the scores.
:return: True of the tag was appended.
"""
if not isinstance(tag, sppasTag):
raise AnnDataTypeError(tag, 'sppasTag')
if self.__tags is None:
self.__tags = list()
if len(self.__tags) > 0:
if self.__tags[0][0].get_type() != tag.get_type():
raise AnnDataTypeError(tag, self.__tags[0][0].get_type())
if add is True:
for i in range(len(self.__tags)):
current_tag, current_score = self.__tags[i]
if tag.get_typed_content() == current_tag.get_typed_content():
if score is not None:
if current_score is None:
self.__tags[i][1] = score
else:
self.__tags[i][1] += score
return False
self.__tags.append([tag, score])
return True
Remove a tag of the list.
def remove(self, tag):
"""Remove a tag of the list.
:param tag: (sppasTag) the tag to be removed of the list.
"""
if not isinstance(tag, sppasTag):
raise AnnDataTypeError(tag, 'sppasTag')
if self.__tags is not None:
if len(self.__tags) == 1:
self.__tags = None
else:
for t in self.__tags:
if t[0] == tag:
self.__tags.remove(t)
Return the score of a tag or None if tag is not in the label.
def get_score(self, tag):
"""Return the score of a tag or None if tag is not in the label.
:param tag: (sppasTag)
:returns: score: (float)
"""
if not isinstance(tag, sppasTag):
raise AnnDataTypeError(tag, 'sppasTag')
if self.__tags is not None:
for t in self.__tags:
if t[0] == tag:
return t[1]
return None
Set a score to a given tag.
def set_score(self, tag, score):
"""Set a score to a given tag.
:param tag: (sppasTag)
:param score: (float)
"""
if not isinstance(tag, sppasTag):
raise AnnDataTypeError(tag, 'sppasTag')
if self.__tags is not None:
for i, t in enumerate(self.__tags):
if t[0] == tag:
self.__tags[i][1] = score
Return the best sppasTag, i.e. the one with the better score.
def get_best(self):
"""Return the best sppasTag, i.e. the one with the better score.
:returns: (sppasTag or None)
"""
if self.__tags is None:
return None
if len(self.__tags) == 1:
return self.__tags[0][0]
_max_tag = self.__tags[0][0]
_max_score = self.__tags[0][1]
for t, s in reversed(self.__tags):
if _max_score is None or (s is not None and s > _max_score):
_max_score = s
_max_tag = t
return _max_tag
Return the type of the tags content.
def get_type(self):
"""Return the type of the tags content."""
if self.__tags is None:
return 'str'
return self.__tags[0][0].get_type()
Return False if no tag is set.
def is_tagged(self):
"""Return False if no tag is set."""
if self.__tags is None:
return False
if len(self.__tags) == 0:
return False
return True
Return True if tags are string or unicode.
Return False if no tag is set.
def is_string(self):
"""Return True if tags are string or unicode.
Return False if no tag is set.
"""
if self.is_tagged() is False:
return False
return self.__tags[0][0].get_type() == 'str'
Return True if tags are of type "float".
Return False if no tag is set.
def is_float(self):
"""Return True if tags are of type "float".
Return False if no tag is set.
"""
if self.is_tagged() is False:
return False
return self.__tags[0][0].get_type() == 'float'
Return True if tags are of type "int".
Return False if no tag is set.
def is_int(self):
"""Return True if tags are of type "int".
Return False if no tag is set.
"""
if self.is_tagged() is False:
return False
return self.__tags[0][0].get_type() == 'int'
Return True if tags are of type "bool".
Return False if no tag is set.
def is_bool(self):
"""Return True if tags are of type "bool".
Return False if no tag is set.
"""
if self.is_tagged() is False:
return False
return self.__tags[0][0].get_type() == 'bool'
Return True if tags are of type "point".
Return False if no tag is set.
def is_point(self):
"""Return True if tags are of type "point".
Return False if no tag is set.
"""
if self.is_tagged() is False:
return False
return self.__tags[0][0].get_type() == 'point'
Return True if tags are of type "rect".
Return False if no tag is set.
def is_rect(self):
"""Return True if tags are of type "rect".
Return False if no tag is set.
"""
if self.is_tagged() is False:
return False
return self.__tags[0][0].get_type() == 'rect'
Return a deep copy of the label.
def copy(self):
"""Return a deep copy of the label."""
return copy.deepcopy(self)
Return True if a tag matches all or any of the functions.
(bool)
function: a function in python with 2 arguments: tag/value
> Search if a tag is exactly matching "R":
>>> l.match([(exact, "R", False)])
> Search if a tag is starting with "p" or starting with "t":
>>> l.match([(startswith, "p", False),
>>> (startswith, "t", False), ], logic_bool="or")
def match(self, tag_functions, logic_bool='and'):
"""Return True if a tag matches all or any of the functions.
:param tag_functions: list of (function, value, logical_not)
:param logic_bool: (str) Apply a logical "and" or a logical "or" between the functions.
:returns: (bool)
- function: a function in python with 2 arguments: tag/value
- value: the expected value for the tag
- logical_not: boolean
:Example: Search if a tag is exactly matching "R":
>>> l.match([(exact, "R", False)])
:Example: Search if a tag is starting with "p" or starting with "t":
>>> l.match([(startswith, "p", False),
>>> (startswith, "t", False), ], logic_bool="or")
"""
is_matching = False
for tag, score in self.__tags:
matches = list()
for func, value, logical_not in tag_functions:
if logical_not is True:
matches.append(not func(tag, value))
else:
matches.append(func(tag, value))
if logic_bool == 'and':
is_matching = all(matches)
else:
is_matching = any(matches)
if is_matching is True:
return True
return is_matching
Convert the label into a string, include or not alternative tags.
@DeprecationWarning Use aioutils.serialize_label() instead.
def serialize(self, empty='', alt=True):
"""Convert the label into a string, include or not alternative tags.
@DeprecationWarning
Use aioutils.serialize_label() instead.
"""
if self.__tags is None:
return empty
if len(self.__tags) == 0:
return empty
if self.get_best() is None:
return empty
if alt is False or len(self.__tags) == 1:
best = self.get_best()
if best.is_empty():
return empty
return best.get_content()
tag_contents = list()
for tag, score in self.__tags:
content = tag.get_content()
if len(content) > 0:
tag_contents.append(content)
else:
tag_contents.append(empty)
return '{' + '|'.join(tag_contents) + '}'
def __format__(self, fmt):
return str(self).__format__(fmt)
def __repr__(self):
st = ''
if self.__tags is not None:
for t, s in self.__tags:
st += 'sppasTag({!s:s}, score={!s:s}), '.format(t, s)
return st
def __str__(self):
st = ''
if self.__tags is not None:
for t, s in self.__tags:
st += '{!s:s}, {!s:s} ; '.format(t, s)
return st
def __iter__(self):
if self.__tags is not None:
for t in self.__tags:
yield t
def __getitem__(self, i):
if self.__tags is not None:
return self.__tags[i]
else:
raise IndexError(i)
def __len__(self):
if self.__tags is not None:
return len(self.__tags)
return 0
def __eq__(self, other):
if self.__tags is not None:
if other is None:
return False
if isinstance(other, sppasLabel) is False:
return False
if len(self.__tags) != len(other):
return False
for tag1, tag2 in zip(self.__tags, other):
if tag1[0] != tag2[0]:
return False
if tag1[1] != tag2[1]:
return False
return True
elif other is None:
return True
return False
def __ne__(self, other):
return not self == other