def rbfopt_test_interface(args, black_box):
"""Interface for test functions.
Optimize the specified objective function using the algorithmic
options given on the command line.
Parameters
----------
args : Dict[string]
A dictionary containing the values of the parameters in a
format args['name'] = value.
black_box : :class:`rbfopt_black_box.RbfoptBlackBox`
The black box to be optimized.
"""
if (not isinstance(black_box, RbfoptBlackBox)):
raise ValueError('The specified module does not contain a ' +
'valid BlackBox instance')
settings = RbfoptSettings.from_dictionary(args)
settings.print(sys.stdout)
alg = RbfoptAlgorithm(settings = settings, black_box = black_box)
result = alg.optimize()
print('RbfoptAlgorithm.optimize() returned ' +
'function value {:.15f}'.format(result[0]))
for (i, val) in enumerate(result[1]):
print('x{:<4d}: {:16.6f}'.format(i, val))
def rbfopt_test_interface(args, black_box):
"""Interface for test functions.
Optimize the specified objective function using the algorithmic
options given on the command line.
Parameters
----------
args : Dict[string]
A dictionary containing the values of the parameters in a
format args['name'] = value.
black_box : :class:`rbfopt_black_box.RbfoptBlackBox`
The black box to be optimized.
"""
if (not isinstance(black_box, RbfoptBlackBox)):
raise ValueError('The specified module does not contain a ' +
'valid BlackBox instance')
settings = RbfoptSettings.from_dictionary(args)
settings.print(sys.stdout)
alg = RbfoptAlgorithm(settings = settings, black_box = black_box)
result = alg.optimize()
print('RbfoptAlgorithm.optimize() returned ' +
'function value {:.15f}'.format(result[0]))
for (i, val) in enumerate(result[1]):
print('x{:<4d}: {:16.6f}'.format(i, val))
def register_options(parser):
"""Add options to the command line parser.
Register all the options for the optimization algorithm.
Parameters
----------
parser : argparse.ArgumentParser
The parser.
See also
--------
:class:`rbfopt_settings.RbfoptSettings` for a detailed description of
all the command line options.
"""
# Algorithmic settings
algset = parser.add_argument_group('Algorithmic settings')
# Get default values from here
default = RbfoptSettings()
attrs = vars(default)
docstring = default.__doc__
param_docstring = docstring[docstring.find('Parameters'):docstring.
find('Attributes')].split(' : ')
param_name = [val.split(' ')[-1].strip() for val in param_docstring[:-1]]
param_type = [val.split('\n')[0].strip() for val in param_docstring[1:]]
param_help = [
' '.join(line.strip() for line in val.split('\n')[1:-2])
for val in param_docstring[1:]
]
# We extract the default from the docstring in case it is
# necessary, but we use the actual default from the object above.
param_default = [
val.split(' ')[-1].rstrip('.').strip('\'') for val in param_help
]
for i in range(len(param_name)):
if (param_type[i] == 'float'):
type_fun = float
elif (param_type[i] == 'int'):
type_fun = int
elif (param_type[i] == 'bool'):
type_fun = ast.literal_eval
else:
type_fun = str
algset.add_argument('--' + param_name[i],
action='store',
dest=param_name[i],
type=type_fun,
help=param_help[i],
default=getattr(default, param_name[i]))
def register_options(parser):
"""Add options to the command line parser.
Register all the options for the optimization algorithm.
Parameters
----------
parser : argparse.ArgumentParser
The parser.
See also
--------
:class:`rbfopt_settings.RbfoptSettings` for a detailed description of
all the command line options.
"""
# Algorithmic settings
algset = parser.add_argument_group('Algorithmic settings')
# Get default values from here
default = RbfoptSettings()
attrs = vars(default)
docstring = default.__doc__
param_docstring = docstring[docstring.find('Parameters'):
docstring.find('Attributes')].split(' : ')
param_name = [val.split(' ')[-1].strip() for val in param_docstring[:-1]]
param_type = [val.split('\n')[0].strip() for val in param_docstring[1:]]
param_help = [' '.join(line.strip() for line in val.split('\n')[1:-2])
for val in param_docstring[1:]]
# We extract the default from the docstring in case it is
# necessary, but we use the actual default from the object above.
param_default = [val.split(' ')[-1].rstrip('.').strip('\'')
for val in param_help]
for i in range(len(param_name)):
if (param_type[i] == 'float'):
type_fun = float
elif (param_type[i] == 'int'):
type_fun = int
elif (param_type[i] == 'bool'):
type_fun = ast.literal_eval
else:
type_fun = str
algset.add_argument('--' + param_name[i], action='store',
dest=param_name[i],
type=type_fun,
help=param_help[i],
default=getattr(default, param_name[i]))
intset = parser.add_argument_group('Execution settings')
intset.add_argument('black_box_file', action='store',
metavar='FILE_NAME', type=str,
help='Name of python file containing black box ' +
'function and the description of its ' +
'characteristics. This file should implement a ' +
'BlackBox class derived from ' +
'rbfopt_black_box.BlackBox, with name ' +
'RbfoptBlackBox. Note: the directory containing ' +
'it will be added to the Python path.')
intset.add_argument('--load', '-l', action='store', dest='load_state',
help='File to read state to resume optimization')
intset.add_argument('--log', '-o', action='store',
metavar='LOG_FILE_NAME', dest='output_stream',
help='Name of log file for output redirection')
intset.add_argument('--pause', '-p', action='store', dest='pause',
default=sys.maxsize, type=int,
help='Number of iterations after which ' +
'the optimization process should be paused')
intset.add_argument('--points_from_file', '-f', action='store',
metavar='POINTS_FILE_NAME', dest='points_file',
type=str, default=None,
help='Name of a file containing coordinates of ' +
'points that have already been evaluated or must ' +
'be evaluated. The points must be given one per ' +
'line with each value separated by space. Each ' +
'row can optionally terminate with the objective ' +
'function value at the point (if such value is ' +
'not given, it will be evaluated by the algorithm).')
intset.add_argument('--print_solution', '-ps', action='store',
dest='print_solution',
default=True, type=ast.literal_eval,
help='Print solution at the end of the ' +
'optimization (or after a pause).')
intset.add_argument('--save', '-s', action='store', dest='dump_state',
help='File to save state after optimization. ' +
'Note that this is different from the options ' +
'save_state_interval and save_state_file because ' +
'here the state is only saved at the end of ' +
'the optimization (or after a pause).')
def rbfopt_cl_interface(args, black_box):
"""Command-line interface.
Optimize the specified objective function using the algorithmic
options given on the command line.
Parameters
----------
args : Dict[string]
A dictionary containing the values of the parameters in a
format args['name'] = value.
black_box : :class:`rbfopt_black_box.RbfoptBlackBox`
The black box to be optimized.
"""
if (not isinstance(black_box, RbfoptBlackBox)):
raise ValueError('The specified module does not contain a ' +
'valid BlackBox instance')
# Open output stream if necessary
if (args['output_stream'] is None):
output_stream = sys.stdout
else:
try:
output_stream = open(args['output_stream'], 'w')
except IOError as e:
print('Error while opening log file', file=sys.stderr)
print(e, file=sys.stderr)
# Make a copy of parameters and adjust them, deleting keys
# that are not recognized as valid by RbfoptSettings.
local_args = args.copy()
del local_args['black_box_file']
del local_args['output_stream']
del local_args['load_state']
del local_args['dump_state']
del local_args['points_file']
del local_args['pause']
del local_args['print_solution']
settings = RbfoptSettings.from_dictionary(local_args)
settings.print(output_stream=output_stream)
if (args['load_state'] is not None):
alg = RbfoptAlgorithm.load_from_file(args['load_state'])
elif (args['points_file'] is not None):
try:
init_node_pos = list()
points_file = open(args['points_file'], 'r')
for line in points_file:
init_node_pos.append([float(val) for val in line.split()])
if (len(init_node_pos[0]) == black_box.get_dimension() + 1):
# In this case the file contains function values as well,
# as the last column
init_node_val = np.array([val[-1] for val in init_node_pos])
init_node_pos = np.array([val[:-1] for val in init_node_pos])
else:
init_node_pos = np.array(init_node_pos)
init_node_val = None
except Exception as e:
print('Exception raised reading file with initialization points',
file=output_stream)
print(e, file=output_stream)
output_stream.close()
raise
alg = RbfoptAlgorithm(settings=settings, black_box=black_box,
init_node_pos=init_node_pos,
init_node_val=init_node_val)
else:
alg = RbfoptAlgorithm(settings=settings, black_box=black_box)
alg.set_output_stream(output_stream)
result = alg.optimize(args['pause'])
print('RbfoptAlgorithm.optimize() returned ' +
'function value {:.15f}'.format(result[0]),
file=output_stream)
if (args['print_solution']):
for (i, val) in enumerate(result[1]):
print('x{:<4d}: {:16.6f}'.format(i, val), file=output_stream)
if (args['dump_state'] is not None):
alg.save_to_file(args['dump_state'])
print('Dumped state to file {:s}'.format(args['dump_state']),
file=output_stream)
output_stream.close()
# Open output stream if necessary
output_stream = my_rbfopt_utils.open_output_stream(algo_args)
# Create dictionary from subparser
dict_args = vars(algo_args)
# Remove non-RBFOpt arguments from algo_arguments because we want to
# pass the dict_args with all the rbfopt setting arguments to
# construct an Rbfopt settings object
dict_args.pop('output_stream')
# Use parameters to create Black Box and remove them from dictonary
black_box = my_rbfopt_utils.construct_black_box(
parameters, my_rbfopt_utils.read_write_obj_fun)
settings = RbfoptSettings.from_dictionary(dict_args)
# Run the interface
alg = RbfoptAlgorithm(settings, black_box)
if (output_stream is not None):
alg.set_output_stream(output_stream)
alg.optimize()
# Run single-objective CMA-ES
elif args.mode == "CMAES":
assert args.param_list is not None, "Parameter string is missing!"
parameters = args.param_list
def rbfopt_cl_interface(args, black_box):
"""Command-line interface.
Optimize the specified objective function using the algorithmic
options given on the command line.
Parameters
----------
args : Dict[string]
A dictionary containing the values of the parameters in a
format args['name'] = value.
black_box : :class:`rbfopt_black_box.RbfoptBlackBox`
The black box to be optimized.
"""
if (not isinstance(black_box, RbfoptBlackBox)):
raise ValueError('The specified module does not contain a ' +
'valid BlackBox instance')
# Open output stream if necessary
if (args['output_stream'] is None):
output_stream = sys.stdout
else:
try:
output_stream = open(args['output_stream'], 'w')
except IOError as e:
print('Error while opening log file', file = sys.stderr)
print(e, file = sys.stderr)
# Make a copy of parameters and adjust them, deleting keys
# that are not recognized as valid by RbfoptSettings.
local_args = args.copy()
del local_args['black_box_file']
del local_args['output_stream']
del local_args['load_state']
del local_args['dump_state']
del local_args['points_file']
del local_args['pause']
del local_args['print_solution']
settings = RbfoptSettings.from_dictionary(local_args)
settings.print(output_stream = output_stream)
if (args['load_state'] is not None):
alg = RbfoptAlgorithm.load_from_file(args['load_state'])
elif (args['points_file'] is not None):
try:
init_node_pos = list()
points_file = open(args['points_file'], 'r')
for line in points_file:
init_node_pos.append([float(val) for val in line.split()])
if (len(init_node_pos[0]) == black_box.get_dimension() + 1):
# In this case the file contains function values as well,
# as the last column
init_node_val = np.array([val[-1] for val in init_node_pos])
init_node_pos = np.array([val[:-1] for val in init_node_pos])
else:
init_node_pos = np.array(init_node_pos)
init_node_val = None
except Exception as e:
print('Exception raised reading file with initialization points',
file = output_stream)
print(type(e), file = output_stream)
print(e, file = output_stream)
output_stream.close()
exit()
alg = RbfoptAlgorithm(settings = settings, black_box = black_box,
init_node_pos = init_node_pos,
init_node_val = init_node_val)
else:
alg = RbfoptAlgorithm(settings = settings, black_box = black_box)
alg.set_output_stream(output_stream)
result = alg.optimize(args['pause'])
print('RbfoptAlgorithm.optimize() returned ' +
'function value {:.15f}'.format(result[0]),
file = output_stream)
if (args['print_solution']):
for (i, val) in enumerate(result[1]):
print('x{:<4d}: {:16.6f}'.format(i, val), file = output_stream)
if (args['dump_state'] is not None):
alg.save_to_file(args['dump_state'])
print('Dumped state to file {:s}'.format(args['dump_state']),
file = output_stream)
output_stream.close()