def read_colors (path_color_file):
"""
Reads user colors for each data_type
:param None path_color_file: :py:func:`str` path to read user color for data_types
:returns: d_user_color dictionary {'data_type_1': 'orange', 'data_type_2':'blue'}
"""
check_path(path_color_file)
comment_tag_t = "#"
d_user_color = {}
with open(path_color_file) as f:
for row in f:
if(row.startswith(comment_tag_t) or row.startswith('\n')):
continue
row_split = row.rstrip('\n').split('\t')
(data_type, color) = row_split
# colors are checked inside tracks.assign_color
d_user_color[data_type] = color
return d_user_color
def read_colors(path_color_file):
"""
Reads user colors for each data_type
:param None path_color_file: :py:func:`str` path to read user color for data_types
:returns: d_user_color dictionary {'data_type_1': 'orange', 'data_type_2':'blue'}
"""
check_path(path_color_file)
comment_tag_t = "#"
d_user_color = {}
with open(path_color_file) as f:
for row in f:
if (row.startswith(comment_tag_t) or row.startswith('\n')):
continue
row_split = row.rstrip('\n').split('\t')
(data_type, color) = row_split
# colors are checked inside tracks.assign_color
d_user_color[data_type] = color
return d_user_color
def __init__(self, path, map_dict, header=True, **kwargs):
self.path = check_path(path)
self._in_file = open(self.path, "rb")
self.delimiter = self._check_delimiter(self.path, kwargs.get('delimiter', "\t"))
self.header = header
self._reader = self._reader_data()
self.fieldsB = self._set_fields_b(kwargs.get('fields_names', None))
self.fieldsG_dict = self._set_fields_g(map_dict)
self.fieldsG = self.fieldsG_dict.keys() #here before I added the new fields
self.min = self.max = 0
self.range_values = 0
self.data = self._simple_read()
# self.data = self._read(multiply_t = kwargs.get('multiply_t', 1), intervals=kwargs.get('intervals', False))
self.data_types = self.get_field_items(field ="data_types", data = self.data, default="a")
self.tracks = self.get_field_items(field="track", data = self.data, default="1")#TODO maybe this function will be more general if instead of giving field name
def __init__(self, path, map_dict, header=True, **kwargs):
self.path = check_path(path)
self._in_file = open(self.path, "rb")
self.delimiter = self._check_delimiter(self.path,
kwargs.get('delimiter', "\t"))
self.header = header
self._reader = self._reader_data()
self.fieldsB = self._set_fields_b(kwargs.get('fields_names', None))
self.fieldsG_dict = self._set_fields_g(map_dict)
self.fieldsG = self.fieldsG_dict.keys(
) #here before I added the new fields
self.min = self.max = 0
self.range_values = 0
self.data = self._simple_read()
# self.data = self._read(multiply_t = kwargs.get('multiply_t', 1), intervals=kwargs.get('intervals', False))
self.data_types = self.get_field_items(field="data_types",
data=self.data,
default="a")
self.tracks = self.get_field_items(
field="track", data=self.data, default="1"
) #TODO maybe this function will be more general if instead of giving field name
def jaaba_scores_to_csv(input_file, name_file="JAABA_scores", mode="w", delimiter="\t", path_w=None, norm=False, data_type="a"):
"""
Creates a csv file from a scores file produced using JAABA and in matlab format
:param mode: :py:func:`str` mode to use by default write
:param "\t" delimiter: :py:func:`str` Character use to separate values of
the same record in file (default "\t").
:param None path_w: :py:func:`str` path to dump the files, by default None
:param False norm: set whether data should be normalize (-1,1) using normalization
factor contained in the file
:param data_type: :py:func:`str` data type in the file "behavior" e.g. chase
"""
path = ""
header = ["animal", "startTime", "endTime", "value", "dataType"]
if not path_w:
path = getcwd()
print >>stderr, 'CSV file will be dump into \"%s\" ' \
'as not path has been set in path_w' % (path)
else:
if exists(path_w):
path = path_w
else:
raise IOError('Provided path does not exists: %s' % path_w)
input_file = check_path(input_file)
jaaba_data = loadmat(input_file)
# Checking JAABA version
version_jaaba = hstack(hstack(hstack(jaaba_data['version'])))[0][0]
if version_jaaba != '0.5.1':
print >>stderr, 'WARNING: JAABA version is not 0.5.1 but %s, this might cause ' \
'problems if the structure of JAABA files has changed.' \
% (version_jaaba)
# Structure of the file can be find here:
# http://jaaba.sourceforge.net/ApplyingAClassifier.html#ScoresFile
start_times = jaaba_data['allScores']['t0s']
end_times = jaaba_data['allScores']['t1s']
scores = jaaba_data['allScores']['scores']
score_norm = jaaba_data['allScores']['scoreNorm']
start_times_flat = hstack(hstack(hstack(start_times)))
end_times_flat = hstack(hstack(hstack(end_times)))
scores_flat = hstack(hstack(hstack(scores)))
score_norm = hstack(hstack(score_norm))[0][0]
scoreFile = open(join(path, name_file + _csv_file_ext), mode)
scoreFile.write(delimiter.join(header) + "\n")
if norm:
# Dirty way of solving problem with ipython notebook, division was not working there
score_norm = float(score_norm)
scores_flat = divide(scores_flat, score_norm)
for idx_animal, start_times_animal in enumerate (start_times_flat):
start_times_animal= hstack(start_times_animal)
end_times_animal = hstack(end_times_flat [idx_animal])
scores_animal = hstack(scores_flat [idx_animal])
for idx_time, start_time in enumerate (start_times_animal):
end_time = end_times_animal[idx_time]
mean_score = mean(scores_animal[start_time:end_time])
# Because we use the convention that the animal is performing the behavior
# from frame t to t+1 if it is labeled/classified as performing the behavior
# at frame t, allScores.postprocessed{i}(allScores.t1s{i}(j)) will be 0 and
# allScores.postprocessed{i}(allScores.t0s{i}(j)) will be 1.
# that is why I substract one to the end_time
# In fact in the graphical interface it starts at start_time - 0.5 and ends in
# end_time - 0.5
scoreFile.write(delimiter.join('{}'.format(v) for v in [idx_animal+1, start_time, end_time -1, mean_score, data_type]) + "\n")
scoreFile.close()
def extract_jaaba_features(dir_perframe, output="csv", map_jaaba=False, delimiter="\t", feature="velmag", path_w=""):
"""
Creates a csv file or a IntData object from feature mat files dumped by JAABA in perframe directory in matlab format
:param dir_perframe: path to the JAABA directory where perframe features are dumped
:param "csv" output: :py:func:`str` sets whether data has to be extracted to a csv or an IntData object
:param map_jaaba: path to the mapping files between JAABA data and pergola ontology
:param "\t" delimiter: :py:func:`str` Character used in the csv output file to separate values of
the same record (default "\t").
:param "velmag" feature: :py:func:`str` data type (feature) to extract e.g. velmag (speed of the center of rotation)
More features can be found in http://ctrax.sourceforge.net/bmat.html
:param None path_w: :py:func:`str` path to dump the files
returns: IntData object
"""
input_path = join(dir_perframe, feature + ".mat")
input_file = check_path(input_path)
jaaba_feature = loadmat(input_file)
output_option = ["csv", "IntData"]
if output not in output_option:
raise ValueError("Option output \'%s\' not allowed. Possible values are %s"%(output_options, ', '.join(['{}'.format(m) for m in output_options])))
temp = NamedTemporaryFile()
header = ["animal", "startTime", "endTime", "value", "dataType"]
temp.write(delimiter.join(header) + "\n")
for id_animal, animal_jaaba_feature in enumerate (jaaba_feature['data'][0]):
animal_jaaba_feature= hstack(animal_jaaba_feature)
for t, v in ndenumerate(animal_jaaba_feature):
temp.write(delimiter.join('{}'.format(v) for v in [id_animal+1, t[0], t[0]+1, v, feature]) + "\n")
# rewinds the file handle
temp.seek(0)
if output == "csv":
if not path_w:
path = getcwd()
print >>stderr, 'CSV file will be dump into \"%s\" ' \
'as not path has been set in path_w' % (path)
else:
if exists(path_w):
path = path_w
else:
raise IOError('Provided path does not exists: %s' % path_w)
feature_file = open(join(path, feature + _csv_file_ext), "wb")
copyfileobj(temp, feature_file)
temp.close()
elif output == "IntData":
map_jaaba = check_path(map_jaaba)
map = MappingInfo(map_jaaba)
int_data_jaaba = IntData(temp.name, map_dict = map.correspondence)
temp.close()
return (int_data_jaaba)
def jaaba_scores_to_intData(input_file, map_jaaba, name_file="JAABA_scores", delimiter="\t", norm=False, data_type="a"):
"""
Creates a csv file from a scores file produced using JAABA and in matlab format
:param input_file: path to the JAABA file in matlab format
:param map_jaaba: path to the mapping files between JAABA data and pergola ontology
:param "\t" delimiter: :py:func:`str` Character use to separate values of
the same record in file (default "\t").
:param False norm: set whether data should be normalize (-1,1) using normalization
factor contained in the file
:param data_type: :py:func:`str` data type in the file "behavior" e.g. chase
:returns: IntData object
"""
path = ""
header = ["animal", "startTime", "endTime", "value", "dataType"]
input_file = check_path(input_file)
jaaba_data = loadmat(input_file)
# Checking JAABA version
version_jaaba = hstack(hstack(hstack(jaaba_data['version'])))[0][0]
if version_jaaba != '0.5.1':
print >>stderr, 'WARNING: JAABA version is not 0.5.1 but %s, this might cause ' \
'problems if the structure of JAABA files has changed.' \
% (version_jaaba)
# Structure of the file can be find here:
# http://jaaba.sourceforge.net/ApplyingAClassifier.html#ScoresFile
start_times = jaaba_data['allScores']['t0s']
end_times = jaaba_data['allScores']['t1s']
scores = jaaba_data['allScores']['scores']
score_norm = jaaba_data['allScores']['scoreNorm']
start_times_flat = hstack(hstack(hstack(start_times)))
end_times_flat = hstack(hstack(hstack(end_times)))
scores_flat = hstack(hstack(hstack(scores)))
score_norm = hstack(hstack(score_norm))[0][0]
temp = NamedTemporaryFile(delete=True)
temp.write(delimiter.join(header) + "\n")
if norm:
# Dirty way of solving problem with ipython notebook, division was not working there
score_norm = float(score_norm)
scores_flat = divide(scores_flat, score_norm)
for idx_animal, start_times_animal in enumerate (start_times_flat):
start_times_animal= hstack(start_times_animal)
end_times_animal = hstack(end_times_flat [idx_animal])
scores_animal = hstack(scores_flat [idx_animal])
for idx_time, start_time in enumerate (start_times_animal):
end_time = end_times_animal[idx_time]
mean_score = mean(scores_animal[start_time:end_time])
# Because we use the convention that the animal is performing the behavior
# from frame t to t+1 if it is labeled/classified as performing the behavior
# at frame t, allScores.postprocessed{i}(allScores.t1s{i}(j)) will be 0 and
# allScores.postprocessed{i}(allScores.t0s{i}(j)) will be 1.
# that is why I substract one to the end_time
# In fact in the graphical interface it starts at start_time - 0.5 and ends in
# end_time - 0.5
temp.write(delimiter.join('{}'.format(v) for v in [idx_animal+1, start_time, end_time -1, mean_score, data_type]) + "\n")
# rewinds the file handle
temp.seek(0)
map_jaaba = check_path(map_jaaba)
map = MappingInfo(map_jaaba)
int_data_jaaba = IntData(temp.name, map_dict = map.correspondence)
temp.close()
return (int_data_jaaba)
def jaaba_scores_to_csv(input_file,
name_file="JAABA_scores",
mode="w",
delimiter="\t",
path_w=None,
norm=False,
data_type="a"):
"""
Creates a csv file from a scores file produced using JAABA and in matlab format
:param mode: :py:func:`str` mode to use by default write
:param "\t" delimiter: :py:func:`str` Character use to separate values of
the same record in file (default "\t").
:param None path_w: :py:func:`str` path to dump the files, by default None
:param False norm: set whether data should be normalize (-1,1) using normalization
factor contained in the file
:param data_type: :py:func:`str` data type in the file "behavior" e.g. chase
"""
path = ""
header = ["animal", "startTime", "endTime", "value", "dataType"]
if not path_w:
path = getcwd()
print >>stderr, 'CSV file will be dump into \"%s\" ' \
'as not path has been set in path_w' % (path)
else:
if exists(path_w):
path = path_w
else:
raise IOError('Provided path does not exists: %s' % path_w)
input_file = check_path(input_file)
jaaba_data = loadmat(input_file)
# Checking JAABA version
version_jaaba = hstack(hstack(hstack(jaaba_data['version'])))[0][0]
if version_jaaba != '0.5.1':
print >>stderr, 'WARNING: JAABA version is not 0.5.1 but %s, this might cause ' \
'problems if the structure of JAABA files has changed.' \
% (version_jaaba)
# Structure of the file can be find here:
# http://jaaba.sourceforge.net/ApplyingAClassifier.html#ScoresFile
start_times = jaaba_data['allScores']['t0s']
end_times = jaaba_data['allScores']['t1s']
scores = jaaba_data['allScores']['scores']
score_norm = jaaba_data['allScores']['scoreNorm']
start_times_flat = hstack(hstack(hstack(start_times)))
end_times_flat = hstack(hstack(hstack(end_times)))
scores_flat = hstack(hstack(hstack(scores)))
score_norm = hstack(hstack(score_norm))[0][0]
scoreFile = open(join(path, name_file + _csv_file_ext), mode)
scoreFile.write(delimiter.join(header) + "\n")
if norm:
# Dirty way of solving problem with ipython notebook, division was not working there
score_norm = float(score_norm)
scores_flat = divide(scores_flat, score_norm)
for idx_animal, start_times_animal in enumerate(start_times_flat):
start_times_animal = hstack(start_times_animal)
end_times_animal = hstack(end_times_flat[idx_animal])
scores_animal = hstack(scores_flat[idx_animal])
for idx_time, start_time in enumerate(start_times_animal):
end_time = end_times_animal[idx_time]
mean_score = mean(scores_animal[start_time:end_time])
# Because we use the convention that the animal is performing the behavior
# from frame t to t+1 if it is labeled/classified as performing the behavior
# at frame t, allScores.postprocessed{i}(allScores.t1s{i}(j)) will be 0 and
# allScores.postprocessed{i}(allScores.t0s{i}(j)) will be 1.
# that is why I substract one to the end_time
# In fact in the graphical interface it starts at start_time - 0.5 and ends in
# end_time - 0.5
scoreFile.write(
delimiter.join('{}'.format(v) for v in [
idx_animal + 1, start_time, end_time -
1, mean_score, data_type
]) + "\n")
scoreFile.close()
def extract_jaaba_features(dir_perframe,
output="csv",
map_jaaba=False,
delimiter="\t",
feature="velmag",
path_w=""):
"""
Creates a csv file or a IntData object from feature mat files dumped by JAABA in perframe directory in matlab format
:param dir_perframe: path to the JAABA directory where perframe features are dumped
:param "csv" output: :py:func:`str` sets whether data has to be extracted to a csv or an IntData object
:param map_jaaba: path to the mapping files between JAABA data and pergola ontology
:param "\t" delimiter: :py:func:`str` Character used in the csv output file to separate values of
the same record (default "\t").
:param "velmag" feature: :py:func:`str` data type (feature) to extract e.g. velmag (speed of the center of rotation)
More features can be found in http://ctrax.sourceforge.net/bmat.html
:param None path_w: :py:func:`str` path to dump the files
returns: IntData object
"""
input_path = join(dir_perframe, feature + ".mat")
input_file = check_path(input_path)
jaaba_feature = loadmat(input_file)
output_option = ["csv", "IntData"]
if output not in output_option:
raise ValueError(
"Option output \'%s\' not allowed. Possible values are %s" %
(output_options, ', '.join(
['{}'.format(m) for m in output_options])))
temp = NamedTemporaryFile()
header = ["animal", "startTime", "endTime", "value", "dataType"]
temp.write(delimiter.join(header) + "\n")
for id_animal, animal_jaaba_feature in enumerate(jaaba_feature['data'][0]):
animal_jaaba_feature = hstack(animal_jaaba_feature)
for t, v in ndenumerate(animal_jaaba_feature):
temp.write(
delimiter.join(
'{}'.format(v)
for v in [id_animal + 1, t[0], t[0] + 1, v, feature]) +
"\n")
# rewinds the file handle
temp.seek(0)
if output == "csv":
if not path_w:
path = getcwd()
print >>stderr, 'CSV file will be dump into \"%s\" ' \
'as not path has been set in path_w' % (path)
else:
if exists(path_w):
path = path_w
else:
raise IOError('Provided path does not exists: %s' % path_w)
feature_file = open(join(path, feature + _csv_file_ext), "wb")
copyfileobj(temp, feature_file)
temp.close()
elif output == "IntData":
map_jaaba = check_path(map_jaaba)
map = MappingInfo(map_jaaba)
int_data_jaaba = IntData(temp.name, map_dict=map.correspondence)
temp.close()
return (int_data_jaaba)
def jaaba_scores_to_intData(input_file,
map_jaaba,
name_file="JAABA_scores",
delimiter="\t",
norm=False,
data_type="a"):
"""
Creates a csv file from a scores file produced using JAABA and in matlab format
:param input_file: path to the JAABA file in matlab format
:param map_jaaba: path to the mapping files between JAABA data and pergola ontology
:param "\t" delimiter: :py:func:`str` Character use to separate values of
the same record in file (default "\t").
:param False norm: set whether data should be normalize (-1,1) using normalization
factor contained in the file
:param data_type: :py:func:`str` data type in the file "behavior" e.g. chase
:returns: IntData object
"""
path = ""
header = ["animal", "startTime", "endTime", "value", "dataType"]
input_file = check_path(input_file)
jaaba_data = loadmat(input_file)
# Checking JAABA version
version_jaaba = hstack(hstack(hstack(jaaba_data['version'])))[0][0]
if version_jaaba != '0.5.1':
print >>stderr, 'WARNING: JAABA version is not 0.5.1 but %s, this might cause ' \
'problems if the structure of JAABA files has changed.' \
% (version_jaaba)
# Structure of the file can be find here:
# http://jaaba.sourceforge.net/ApplyingAClassifier.html#ScoresFile
start_times = jaaba_data['allScores']['t0s']
end_times = jaaba_data['allScores']['t1s']
scores = jaaba_data['allScores']['scores']
score_norm = jaaba_data['allScores']['scoreNorm']
start_times_flat = hstack(hstack(hstack(start_times)))
end_times_flat = hstack(hstack(hstack(end_times)))
scores_flat = hstack(hstack(hstack(scores)))
score_norm = hstack(hstack(score_norm))[0][0]
temp = NamedTemporaryFile(delete=True)
temp.write(delimiter.join(header) + "\n")
if norm:
# Dirty way of solving problem with ipython notebook, division was not working there
score_norm = float(score_norm)
scores_flat = divide(scores_flat, score_norm)
for idx_animal, start_times_animal in enumerate(start_times_flat):
start_times_animal = hstack(start_times_animal)
end_times_animal = hstack(end_times_flat[idx_animal])
scores_animal = hstack(scores_flat[idx_animal])
for idx_time, start_time in enumerate(start_times_animal):
end_time = end_times_animal[idx_time]
mean_score = mean(scores_animal[start_time:end_time])
# Because we use the convention that the animal is performing the behavior
# from frame t to t+1 if it is labeled/classified as performing the behavior
# at frame t, allScores.postprocessed{i}(allScores.t1s{i}(j)) will be 0 and
# allScores.postprocessed{i}(allScores.t0s{i}(j)) will be 1.
# that is why I substract one to the end_time
# In fact in the graphical interface it starts at start_time - 0.5 and ends in
# end_time - 0.5
temp.write(
delimiter.join('{}'.format(v) for v in [
idx_animal + 1, start_time, end_time -
1, mean_score, data_type
]) + "\n")
# rewinds the file handle
temp.seek(0)
map_jaaba = check_path(map_jaaba)
map = MappingInfo(map_jaaba)
int_data_jaaba = IntData(temp.name, map_dict=map.correspondence)
temp.close()
return (int_data_jaaba)