def save(self, config, configs_child):
if config.experiment.task_params is None or config.experiment.task_params['type'] == 'run':
if isinstance(config.experiment_data['fig'], list):
for fig_id, fig in enumerate(config.experiment_data['fig']):
if config.experiment.method == Method.scatter:
item = config.experiment_data['item'][fig_id]
config.experiment.method_params.pop('items', None)
config.experiment.method_params.pop('x_ranges', None)
config.experiment.method_params.pop('y_ranges', None)
config.experiment.method_params['item'] = item
file_name = config.experiment_data['item'][fig_id]
fn = get_save_path(config) + '/' + file_name
save_figure(fn, fig)
else:
if config.experiment.method == Method.scatter_comparison:
config.experiment.method_params.pop('items', None)
config.experiment.method_params.pop('aux', None)
config.experiment.method_params.pop('data_bases', None)
config.experiment.method_params.pop('x_ranges', None)
config.experiment.method_params.pop('y_ranges', None)
file_name = 'regular'
fn = get_save_path(config) + '/' + file_name
save_figure(fn, config.experiment_data['fig'])
elif config.experiment.task_params['type'] == 'prepare':
pass
def load_table_dict(config):
fn_pkl = get_save_path(config) + '/' + get_file_name(config) + '.pkl'
fn_xlsx = get_save_path(config) + '/' + get_file_name(config) + '.xlsx'
if os.path.isfile(fn_pkl):
table_dict = load_table_dict_pkl(config)
elif os.path.isfile(fn_xlsx):
table_dict = load_table_dict_xlsx(config)
return table_dict
def save_figure(config, fig, prefix=''):
if prefix == '':
fn = get_save_path(config) + '/' + \
config.setup.get_file_name()
else:
fn = get_save_path(config) + '/' + \
prefix + '_' + config.setup.get_file_name()
plotly.offline.plot(fig, filename=fn + '.html', auto_open=False)
pio.write_image(fig, fn + '.png')
pio.write_image(fig, fn + '.pdf')
def is_result_exist(self, config, configs_child):
fn = get_save_path(config) + '/' + \
get_file_name(config) + '.*'
if glob.glob(fn):
return True
else:
return False
def is_result_exist(self, config, configs_child):
file_name = 'regular'
fn = get_save_path(config) + '/' + \
file_name + '.pdf'
if Path(fn).is_file():
return True
else:
return False
def load_table_dict_pkl(config):
fn = get_save_path(config) + '/' + get_file_name(config) + '.pkl'
if os.path.isfile(fn):
f = open(fn, 'rb')
table_dict = pickle.load(f)
f.close()
return table_dict
else:
raise IOError(f'No such file: {fn}')
def save_table_dict(config, table_dict):
fn = get_save_path(config) + '/' + \
config.setup.get_file_name() + '.xlsx'
df = pd.DataFrame(table_dict)
writer = pd.ExcelWriter(fn, engine='xlsxwriter')
df.to_excel(writer, index=False)
writer.save()
fn = get_save_path(config) + '/' + \
config.setup.get_file_name() + '.csv'
with open(fn, 'w') as csvfile:
writer = csv.DictWriter(csvfile,
fieldnames=table_dict.keys(),
lineterminator='\n')
writer.writeheader()
for id in range(0, len(list(table_dict.values())[0])):
tmp_dict = {}
for key, values in table_dict.items():
tmp_dict[key] = values[id]
writer.writerow(tmp_dict)
def load_table_dict(config):
fn = get_save_path(config) + '/' + get_file_name(config) + '.xlsx'
if os.path.isfile(fn):
df = pd.read_excel(fn)
tmp_dict = df.to_dict()
table_dict = {}
for key in tmp_dict:
curr_dict = tmp_dict[key]
table_dict[key] = list(curr_dict.values())
return table_dict
else:
raise IOError(f'No such file: {fn}')
def cpg_proc_special_clock_linreg_dev(
data,
annotations,
attributes,
file,
params=None,
):
if os.path.isfile(file):
head, tail = os.path.split(file)
fn = os.path.splitext(tail)[0]
ext = os.path.splitext(tail)[1]
config_root = Config(data=copy.deepcopy(data),
experiment=Experiment(
type=DataType.cpg,
task=Task.clock,
method=Method.linreg,
params=copy.deepcopy(params)),
annotations=copy.deepcopy(annotations),
attributes=copy.deepcopy(attributes),
is_run=True,
is_root=True)
root = Node(name=str(config_root), config=config_root)
config_child = Config(data=copy.deepcopy(data),
experiment=Experiment(type=DataType.cpg,
task=Task.table,
method=Method.special,
params={'file_name': fn}),
annotations=copy.deepcopy(annotations),
attributes=copy.deepcopy(attributes),
is_run=False,
is_root=False)
Node(name=str(config_child), config=config_child, parent=root)
build_tree(root)
new_file = get_save_path(config_child) + '/' + \
get_file_name(config_child) + ext
copyfile(file, new_file)
calc_tree(root)
else:
raise FileNotFoundError(f'File {file} not found.')
def save(self, config, configs_child):
fn = get_save_path(config) + '/' + \
get_file_name(config)
save_figure(fn, config.experiment_data['fig'])
def save(self, config, configs_child):
fn = get_save_path(config) + '/' + \
get_file_name(config)
save_table_dict(fn, config.metrics)
def betas_clock_special(
data,
annotations,
attributes,
file,
method_params=None,
):
"""
Producing epigentic clock, using best CpGs which are provided in input file.
Epigentic clock represents as table:
Each row corresponds to clocks, which are built on all CpGs from the previous rows including the current row.
Columns:
* item: CpG id.
* aux: gene, on which CpG is mapped.
* R2: determination coefficient of linear regression between real and predicted target observable.
A statistical measure of how well the regression line approximates the data points.
* r: correlation coefficient of linear regression between real and predicted target observable.
* evs: explained variance regression score.
* mae: mean absolute error regression loss.
* rmse: root mean square error
Possible parameters of experiment:
* ``'type'``: type of clocks. \n
Possible options: \n
``'all'``: iterative building of clocks starting from one element in the model,
ending with ``'size'`` elements in the model. \n
``'single '``: building of clocks only with ``'size'`` elements in the model. \n
``'deep'``: iterative building of clocks starting from one element in the model,
ending with ``'size'`` elements in the model, but choosing all possible combinations from ``'size'`` elements.
* ``'part'``: the proportion of considered number of subject in the test set. From ``0.0`` to ``1.0``.
* ``'size'``: maximum number of exogenous variables in a model.
* ``'runs'`` number of bootstrap runs in model
:param data: pdm.Data instance, which specifies information about dataset.
:param annotations: pdm.Annotations instance, which specifies subset of CpGs.
:param attributes: pdm.Attributes instance, which specifies information about subjects.
:param method_params: parameters of experiment.
"""
if os.path.isfile(file):
head, tail = os.path.split(file)
fn = os.path.splitext(tail)[0]
ext = os.path.splitext(tail)[1]
config_root = Config(data=copy.deepcopy(data),
experiment=Experiment(
data=DataType.betas,
task=Task.clock,
method=Method.linreg,
method_params=copy.deepcopy(method_params)),
annotations=copy.deepcopy(annotations),
attributes=copy.deepcopy(attributes),
is_run=True,
is_root=True)
root = Node(name=str(config_root), config=config_root)
config_child = Config(data=copy.deepcopy(data),
experiment=Experiment(
data=DataType.betas,
task=Task.table,
method=Method.special,
method_params={'file_name': fn}),
annotations=copy.deepcopy(annotations),
attributes=copy.deepcopy(attributes),
is_run=False,
is_root=False)
Node(name=str(config_child), config=config_child, parent=root)
build_tree(root)
new_file = get_save_path(config_child) + '/' + \
get_file_name(config_child) + ext
copyfile(file, new_file)
calc_tree(root)
else:
raise FileNotFoundError(f'File {file} not found.')
