import numpy as np
from pymoo.core.individual import Individual
[docs]
class Population(np.ndarray):
def __new__(cls, individuals=[]):
if isinstance(individuals, Individual):
individuals = [individuals]
return np.array(individuals).view(cls)
def has(self, key):
return all([ind.has(key) for ind in self])
def collect(self, func, to_numpy=True):
val = []
for i in range(len(self)):
val.append(func(self[i]))
if to_numpy:
val = np.array(val)
return val
def apply(self, func):
self.collect(func, to_numpy=False)
def set(self, *args, **kwargs):
# if population is empty just return
if self.size == 0:
return
# done for the old interface with the interleaving variable definition
kwargs = interleaving_args(*args, kwargs=kwargs)
# for each entry in the dictionary set it to each individual
for key, values in kwargs.items():
is_iterable = hasattr(values, '__len__') and not isinstance(values, str)
if is_iterable and len(values) != len(self):
raise Exception("Population Set Attribute Error: Number of values and population size do not match!")
for i in range(len(self)):
val = values[i] if is_iterable else values
# check for view and make copy to prevent memory leakage (#455)
if isinstance(val, np.ndarray) and not val.flags["OWNDATA"]:
val = val.copy()
self[i].set(key, val)
return self
def get(self, *args, to_numpy=True, **kwargs):
val = {}
for c in args:
val[c] = []
# for each individual
for i in range(len(self)):
# for each argument
for c in args:
val[c].append(self[i].get(c, **kwargs))
# convert the results to a list
res = [val[c] for c in args]
# to numpy array if desired - default true
if to_numpy:
res = [np.array(e) for e in res]
# return as tuple or single value
if len(args) == 1:
return res[0]
else:
return tuple(res)
@classmethod
def merge(cls, a, b, *args):
# do the regular merge between first and second element
m = merge(a, b)
# process the list of others and merge as well
others = list(args)
while len(others) > 0:
m = merge(m, others.pop(0))
return m
@classmethod
def create(cls, *args):
return Population.__new__(cls, args)
@classmethod
def empty(cls, size=0):
individuals = [Individual() for _ in range(size)]
return Population.__new__(cls, individuals)
@classmethod
def new(cls, *args, **kwargs):
kwargs = interleaving_args(*args, kwargs=kwargs)
if len(kwargs) > 0:
sizes = np.unique(np.array([len(v) for _, v in kwargs.items()]))
if len(sizes) == 1:
size = sizes[0]
else:
raise Exception(f"Population.new needs to be called with same-sized inputs, but the sizes are {sizes}")
else:
size = 0
pop = Population.empty(size)
pop.set(**kwargs)
return pop
def pop_from_array_or_individual(array, pop=None):
# the population type can be different - (different type of individuals)
if pop is None:
pop = Population.empty()
# provide a whole population object - (individuals might be already evaluated)
if isinstance(array, Population):
pop = array
elif isinstance(array, np.ndarray):
pop = pop.new("X", np.atleast_2d(array))
elif isinstance(array, Individual):
pop = Population.empty(1)
pop[0] = array
else:
return None
return pop
def merge(a, b):
if a is None:
return b
elif b is None:
return a
a, b = pop_from_array_or_individual(a), pop_from_array_or_individual(b)
if len(a) == 0:
return b
elif len(b) == 0:
return a
else:
obj = np.concatenate([a, b]).view(Population)
return obj
def interleaving_args(*args, kwargs=None):
if len(args) % 2 != 0:
raise Exception(f"Even number of arguments are required but {len(args)} arguments were provided.")
if kwargs is None:
kwargs = {}
for i in range(int(len(args) / 2)):
key, values = args[i * 2], args[i * 2 + 1]
kwargs[key] = values
return kwargs
def calc_cv(pop, config=None):
if config is None:
config = Individual.default_config()
G, H = pop.get("G", "H")
from pymoo.core.individual import calc_cv as func
CV = np.array([func(g, h, config) for g, h in zip(G, H)])
return CV
