def create_kfold_datasets(dataset, args,
selecting_file_list,
common_options, resume=False):
"""Calling the bigmler procedure to create the k-fold datasets
"""
args.output_dir = os.path.normpath(os.path.join(args.output_dir, "test"))
output_dir = args.output_dir
global subcommand_list
# creating the selecting datasets
for index in range(0, len(selecting_file_list)):
command = COMMANDS["selection"] % (
dataset, selecting_file_list[index],
output_dir)
command_args = command.split()
common_options_list = u.get_options_list(args, common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command = rebuild_command(command_args)
if resume:
next_command = subcommand_list.pop()
if different_command(next_command, command):
resume = False
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
datasets_file = os.path.normpath(os.path.join(output_dir, "dataset_gen"))
return datasets_file, resume
def create_kfold_datasets(dataset, args,
selecting_file_list,
common_options, resume=False):
"""Calling the bigmler procedure to create the k-fold datasets
"""
args.output_dir = os.path.normpath(os.path.join(args.output_dir, "test"))
output_dir = args.output_dir
global subcommand_list
# creating the selecting datasets
for index in range(0, len(selecting_file_list)):
command = COMMANDS["selection"] % (
dataset, selecting_file_list[index],
output_dir)
command_args = command.split()
common_options_list = u.get_options_list(args, common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command = rebuild_command(command_args)
if resume:
next_command = subcommand_list.pop()
if different_command(next_command, command):
resume = False
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
datasets_file = os.path.normpath(os.path.join(output_dir, "dataset_gen"))
return datasets_file, resume
def create_kfold_evaluations(datasets_file,
args,
common_options,
resume=False,
counter=0):
""" Create k-fold cross-validation from a datasets file
"""
global subcommand_list
output_dir = os.path.normpath(
u.check_dir(
os.path.join(u"%s%s" % (args.output_dir, counter),
u"evaluation.json")))
model_fields = args.model_fields
name_suffix = "_subset_%s" % counter
name_max_length = NAME_MAX_LENGTH - len(name_suffix)
name = "%s%s" % (args.name[0:name_max_length], name_suffix)
dataset_id = u.read_datasets(datasets_file)[0]
model_dataset = os.path.normpath(
os.path.join(u.check_dir(datasets_file), dataset_id.replace("/", "_")))
command = COMMANDS["create_cv"] % (datasets_file, output_dir, name,
model_dataset)
command_args = command.split()
if model_fields:
command_args.append("--model-fields")
command_args.append(model_fields)
command_args.append("--objective")
command_args.append(args.objective_field)
command_args = add_model_options(command_args, args)
common_options_list = u.get_options_list(args,
common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command = rebuild_command(command_args)
if resume:
next_command = subcommand_list.pop()
if different_command(next_command, command):
resume = False
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
evaluation_file = os.path.normpath(
os.path.join(output_dir, "evaluation.json"))
try:
with open(evaluation_file) as evaluation_handler:
evaluation = json.loads(evaluation_handler.read())
return evaluation, resume
except (ValueError, IOError):
sys.exit("Failed to retrieve evaluation.")
def create_kfold_evaluations(datasets_file, args, common_options,
resume=False, counter=0):
""" Create k-fold cross-validation from a datasets file
"""
global subcommand_list
output_dir = os.path.normpath(
u.check_dir(os.path.join(u"%s%s" % (args.output_dir, counter),
u"evaluation.json")))
model_fields = args.model_fields
name_suffix = "_subset_%s" % counter
name_max_length = NAME_MAX_LENGTH - len(name_suffix)
name = "%s%s" % (args.name[0: name_max_length], name_suffix)
dataset_id = u.read_datasets(datasets_file)[0]
model_dataset = os.path.normpath(
os.path.join(u.check_dir(datasets_file), dataset_id.replace("/", "_")))
command = COMMANDS["create_cv"] % (datasets_file, output_dir, name,
model_dataset)
command_args = command.split()
if model_fields:
command_args.append("--model-fields")
command_args.append(model_fields)
command_args.append("--objective")
command_args.append(args.objective_field)
command_args = add_model_options(command_args, args)
common_options_list = u.get_options_list(args, common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command = rebuild_command(command_args)
if resume:
next_command = subcommand_list.pop()
if different_command(next_command, command):
resume = False
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
evaluation_file = os.path.normpath(os.path.join(output_dir,
"evaluation.json"))
try:
with open(evaluation_file) as evaluation_handler:
evaluation = json.loads(evaluation_handler.read())
return evaluation, resume
except (ValueError, IOError):
sys.exit("Failed to retrieve evaluation.")
def create_node_th_evaluations(datasets_file,
args,
common_options,
resume=False,
node_threshold=DEFAULT_MIN_NODES):
""" Create node_threshold evaluations
"""
global subcommand_list
output_dir = os.path.normpath(
u.check_dir(
os.path.join(u"%s%s" % (args.output_dir, node_threshold),
"evaluation.json")))
command = COMMANDS["node_threshold"] % (datasets_file, node_threshold,
output_dir)
command_args = command.split()
common_options_list = u.get_options_list(args,
common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command_args.append("--objective")
command_args.append(args.objective_field)
command_args = add_model_options(command_args, args)
command = rebuild_command(command_args)
if resume:
next_command = subcommand_list.pop()
if different_command(next_command, command):
resume = False
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
evaluation_file = os.path.normpath(
os.path.join(output_dir, "evaluation.json"))
try:
with open(evaluation_file, u.open_mode("r")) as evaluation_handler:
evaluation = json.loads(evaluation_handler.read())
return evaluation, resume
except (ValueError, IOError):
sys.exit("Failed to retrieve evaluation.")
def create_kfold_evaluations(datasets_file, args, common_options,
resume=False, counter=0):
""" Create k-fold cross-validation from a datasets file
"""
global subcommand_list
output_dir = u.check_dir(os.path.join("%s%s" % (args.output_dir, counter),
"evaluation.json"))
model_fields = args.model_fields
name_suffix = "_subset_%s" % counter
name_max_length = NAME_MAX_LENGTH - len(name_suffix)
name = "%s%s" % (args.name[0: name_max_length] , name_suffix)
command = COMMANDS["create_cv"] % (datasets_file, output_dir, name)
command_args = command.split()
if model_fields:
command_args.append("--model-fields")
command_args.append(model_fields)
common_options_list = u.get_options_list(args, common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command = " ".join(command_args)
if resume:
next_command = subcommand_list.pop().strip()
if next_command != command:
resume = False
u.log_message("%s\n" % command, log_file=subcommand_file,
console=False)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.log_message("%s\n" % command, log_file=subcommand_file,
console=False)
main_dispatcher(args=command_args)
evaluation_file = os.path.join(output_dir, "evaluation.json")
try:
with open(evaluation_file) as evaluation_handler:
evaluation = json.loads(evaluation_handler.read())
return evaluation, resume
except (ValueError, IOError):
sys.exit("Failed to retrieve evaluation.")
def create_node_th_evaluations(datasets_file, args, common_options,
resume=False,
node_threshold=DEFAULT_MIN_NODES):
""" Create node_threshold evaluations
"""
global subcommand_list
output_dir = os.path.normpath(u.check_dir(
os.path.join(u"%s%s" % (args.output_dir, node_threshold),
"evaluation.json")))
command = COMMANDS["node_threshold"] % (
datasets_file, node_threshold, output_dir)
command_args = command.split()
common_options_list = u.get_options_list(args, common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command_args.append("--objective")
command_args.append(args.objective_field)
command_args = add_model_options(command_args, args)
command = rebuild_command(command_args)
if resume:
next_command = subcommand_list.pop()
if different_command(next_command, command):
resume = False
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.sys_log_message(command, log_file=subcommand_file)
main_dispatcher(args=command_args)
evaluation_file = os.path.normpath(os.path.join(output_dir,
"evaluation.json"))
try:
with open(evaluation_file, u.open_mode("r")) as evaluation_handler:
evaluation = json.loads(evaluation_handler.read())
return evaluation, resume
except (ValueError, IOError):
sys.exit("Failed to retrieve evaluation.")
def create_node_th_evaluations(datasets_file, args, common_options,
resume=False,
node_threshold=DEFAULT_MIN_NODES):
""" Create node_threshold evaluations
"""
global subcommand_list
output_dir = u.check_dir(
os.path.join("%s%s" % (args.output_dir, node_threshold),
"evaluation.json"))
command = COMMANDS["node_threshold"] % (
datasets_file, node_threshold, output_dir)
command_args = command.split()
common_options_list = u.get_options_list(args, common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command = " ".join(command_args)
if resume:
next_command = subcommand_list.pop().strip()
if next_command != command:
resume = False
u.log_message("%s\n" % command, log_file=subcommand_file,
console=False)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.log_message("%s\n" % command, log_file=subcommand_file,
console=False)
main_dispatcher(args=command_args)
evaluation_file = os.path.join(output_dir, "evaluation.json")
try:
with open(evaluation_file) as evaluation_handler:
evaluation = json.loads(evaluation_handler.read())
return evaluation, resume
except (ValueError, IOError):
sys.exit("Failed to retrieve evaluation.")
def create_kfold_datasets(dataset, args,
selecting_file_list, objective, kfold_field,
common_options, resume=False):
"""Calling the bigmler procedure to create the k-fold datasets
"""
args.output_dir = os.path.join(args.output_dir, "test")
output_dir = args.output_dir
k = args.k_folds
global subcommand_list
# creating the selecting datasets
for index in range(0, len(selecting_file_list)):
command = COMMANDS["selection"] % (
dataset, selecting_file_list[index],
output_dir)
command_args = command.split()
common_options_list = u.get_options_list(args, common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command = " ".join(command_args)
if resume:
next_command = subcommand_list.pop().strip()
if next_command != command:
resume = False
u.log_message("%s\n" % command, log_file=subcommand_file,
console=False)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.log_message("%s\n" % command, log_file=subcommand_file,
console=False)
main_dispatcher(args=command_args)
# updating the datasets to set the objective field
datasets_file = os.path.join(output_dir, "dataset_gen")
with open(datasets_file) as datasets_handler:
for line in datasets_handler:
dataset_id = line.strip()
command = COMMANDS["objective"] % (dataset_id,
"dataset_%s" % index, output_dir)
command_args = command.split()
command_args.append("--objective")
command_args.append(objective)
common_options_list = u.get_options_list(args, common_options,
prioritary=command_args)
command_args.extend(common_options_list)
command = " ".join(command_args)
if resume:
next_command = subcommand_list.pop().strip()
if next_command != command:
resume = False
u.log_message("%s\n" % command, log_file=subcommand_file,
console=False)
main_dispatcher(args=command_args)
elif not subcommand_list:
main_dispatcher(args=['main', '--resume'])
resume = False
else:
u.log_message("%s\n" % command, log_file=subcommand_file,
console=False)
main_dispatcher(args=command_args)
return datasets_file, resume
