def test_evolve():
class TestIndividual(Individual):
def __init__(self, *args, **kwargs):
super(TestIndividual, self).__init__(*args, **kwargs)
for x in range(5):
self.append(random.random())
evolve(lambda x: ((sum(x), sum(x), sum(x))), TestIndividual)
def create_network(p_connect=None, p_break=None):
population = 1000
n = Network(population, 200)
plot = False
if plot:
test_plot(n)
return evolution.evolve(n, p_connect, p_break)
def main(instrument, days, popsize, strategy='trend_ema'):
print(colored("Starting evolution....", 'blue'))
evaluate = partial(evaluate_zen, days=days)
print(
blue("Evaluating ") + green(popsize) + blue(" individuals over ") +
green(days) + blue(' days in ') + green(partitions) +
blue(' partitions.'))
try:
Andividual.instruments = selectors[instrument]
except:
# if not in the list, assume it is one usable instrument
Andividual.instruments = [instrument]
Andividual.mate = cxTwoPoint
Andividual.mutate = partial(mutGaussian, mu=0, sigma=sigma, indpb=indpb)
# Andividual.strategy = strategy
strategies = parsing.strategies()
Andividual.strategies = [st for st in strategies if 'forex' not in st]
print('using strategies:', Andividual.strategies)
print(
colored(f"Mating function is ", 'blue') +
colored(Andividual.mate, 'green'))
print(
colored(f"Mutating function is ", 'blue') +
colored(Andividual.mutate, 'green'))
res = evolve(evaluate, Andividual, popsize)
return res
def main(instrument, days, popsize, strategy='trend_ema'):
print(colored("Starting evolution....", 'blue'))
evaluate = partial(evaluate_zen, days=days)
print(blue("Evaluating ")+green(popsize)+blue(" individuals over ") + green(days) + blue(' days in ') + green(partitions) + blue(' partitions.'))
Andividual.instruments = selectors[instrument]
Andividual.mate = cxTwoPoint
Andividual.mutate = partial(mutGaussian, mu=0, sigma=sigma, indpb=indpb)
Andividual.strategy = strategy
print(colored(f"Mating function is ", 'blue') + colored(Andividual.mate, 'green'))
print(colored(f"Mutating function is ", 'blue') + colored(Andividual.mutate, 'green'))
res = evolve(evaluate, Andividual, popsize)
return res
def main(instrument, days, popsize, strategy='trend_ema', nobf=None):
print(term.blue("Starting evolution...."))
evaluate = partial(evaluate_zen, days=days)
print(
term.blue("Evaluating ") + term.green(str(popsize)) +
term.blue(" individuals over ") + term.green(str(days)) +
term.blue(" days in ") + term.green(str(partitions)) +
term.blue(" partitions."))
Andividual.path = path
Andividual.instruments = selectors[instrument]
Andividual.mate = cxTwoPoint
Andividual.mutate = partial(mutGaussian, mu=0, sigma=sigma, indpb=indpb)
Andividual.strategy = strategy
print(term.blue("Mating function is ") + term.green(str(Andividual.mate)))
print(
term.blue("Mutating function is ") +
term.green(str(Andividual.mutate)))
res = evolve(evaluate, Andividual, popsize)
return res
def main():
populationSize = 200 # total size of popultion, maybe in the future reduce over time
threshold = 0.10 # number of top population that will be used for reproduction
mutationRateBest = 0.001 # chance to mutate de best solution
numberOfEvolution = 100001 # max number o time steps
functionType = input("Function ('kappa' or 'maxwellian'): "
) ## only works with kappa and maxwellian
ev_ins = str(input("Event Instant ('yyyymmdd hr:min:sec'): "))
year = int(ev_ins[0:4])
month = int(ev_ins[4:6])
day = int(ev_ins[6:8])
hour = int(ev_ins[9:11])
minute = int(ev_ins[12:14])
second = int(ev_ins[15:17])
event_instant = datetime.datetime(year, month, day, hour, minute,
second) ## instant for data selection
#13 42 40
#10 00 00
xt, yt = data.flux_values(event_instant)
index = []
#removing nan from arrays
for i in xrange(len(xt)):
if np.isnan(xt[i]) or np.isnan(yt[i]) or yt[i] < 1.0:
index.append(i)
x = np.delete(xt, index)
y = np.delete(yt, index)
# x = x[0:-1]
# y = y[0:-1]
#x = np.asarray(x)
#y = np.asarray(y)
#print x,y
#
# creates data for test
#
# kappa is defined here as
# y = A x ( 1 + B x )^-C
#
# and maxwellian as
# y = A x exp(-B x)
#
# Use three coeficients for kappa and two for maxwellian
# if vector of y is given ignore this
if len(y) == 0:
if functionType == "kappa":
coefs = [100, 0.001, 12]
y, mean = evo.kappa(x, coefs)
elif functionType == "maxwellian":
coefs = [100, 0.001]
print "aqui"
y, mean = evo.function(x, coefs)
else:
mean = sum(y) / len(y)
####
# main evolution of the code
evo.evolve(functionType, populationSize, threshold, mutationRateBest,
numberOfEvolution, x, y, mean, event_instant)
import population
import fitness
import evolution
fitness.set_solution('{}{}{}{}'.format('11', '0000', '0' * (8 * 63), '11'))
pop = population.Population.initial(100)
gen_count = 0
while pop.get_fittest().fitness() < fitness.max_fitness():
gen_count += 1
print('Generation #{}: Fittest - {}'.format(gen_count, pop.get_fittest().fitness()))
pop = evolution.evolve(pop)
print('Solution found!')
print('Generation #{}'.format(gen_count))
print('Genes:\n\t{}'.format(''.join([format(byte, '0>8b') for byte in pop.get_fittest().genes])))
#return best_validation_loss
#----<mockup model function>----
#-----<evolution>------------
from evolution import evolve
popsize = 20
generations = 10
parameter_ranges = {
"alma": [1, 100],
"korte": [0.0, 1.0],
"barack": [-10.0, 0.0],
"vadkorte": ["egy", "ketto", "harom"]
}
pop_memory = []
fitness_memory = []
solution, score = evolve(parameter_ranges, dummy_model, popsize, generations,
pop_memory, fitness_memory)
#-----</evolution>------------
print("Ground truth:")
print("alma: 13, korte: 0.2, barack: -8.0, vadkorte: ketto")
print(solution)
print(score)
from evolution import evolve evolve()
