def test_multiple_variables():
import numpy as np
from EvoDAG.population import Inputs
from EvoDAG.cython_utils import SelectNumbers
from EvoDAG import EvoDAG
from SparseArray import SparseArray
y = cl.copy()
gp = EvoDAG(classifier=True,
multiple_outputs=True,
popsize=5,
share_inputs=True)
gp.X = X
gp.X[-1]._eval_tr = SparseArray.fromlist(
[0 for x in range(gp.X[-1].hy.size())])
gp.nclasses(y)
gp.y = y
inputs = Inputs(gp, SelectNumbers([x for x in range(len(gp.X))]))
inputs._func = [inputs._func[-1]]
inputs._nfunc = 1
v = inputs.input()
assert v is not None
mask = np.array(gp._mask[0].full_array(), dtype=np.bool)
D = np.array([x.hy.full_array() for x in gp.X]).T
b = np.array(gp._ytr[0].full_array())
coef = np.linalg.lstsq(D[mask], b[mask])[0]
for a, b in zip(coef, v.weight[0]):
assert_almost_equals(a, b)
def test_inputs():
from EvoDAG.population import Inputs
from EvoDAG.cython_utils import SelectNumbers
from EvoDAG import EvoDAG
y = cl.copy()
gp = EvoDAG(classifier=True,
multiple_outputs=True,
popsize=5,
share_inputs=True)
gp.X = X
gp.nclasses(y)
gp.y = y
inputs = Inputs(gp, SelectNumbers([x for x in range(len(gp.X))]))
func = inputs._func
for f in func:
inputs._func = [f]
inputs._nfunc = 1
v = inputs.input()
assert v is not None
inputs = Inputs(gp, SelectNumbers([x for x in range(len(gp.X))]))
def test_all_variables_inputs():
from EvoDAG.population import Inputs
from EvoDAG.cython_utils import SelectNumbers
from EvoDAG import EvoDAG
y = cl.copy()
gp = EvoDAG(classifier=True,
multiple_outputs=True,
use_all_vars_input_functions=True,
popsize=5,
share_inputs=True)
gp.X = X
gp.nclasses(y)
gp.y = y
inputs = Inputs(gp, SelectNumbers([x for x in range(len(gp.X))]))
func = inputs._func
print(func, gp.nvar)
for f in func:
v = inputs.all_variables()
assert v is not None
assert isinstance(v, f)
assert inputs._all_variables_index == len(func)
def test_restrictions():
from EvoDAG.population import Inputs
from EvoDAG.cython_utils import SelectNumbers
from EvoDAG import EvoDAG
from SparseArray import SparseArray
y = cl.copy()
gp = EvoDAG(classifier=True,
multiple_outputs=True,
popsize=5,
share_inputs=True)
gp.X = X
gp.X[-1]._eval_tr = SparseArray.fromlist(
[0 for x in range(gp.X[-1].hy.size())])
gp.nclasses(y)
gp.y = y
for c in [True, False]:
gp._classifier = c
tag = 'classification' if c else 'regression'
inputs = Inputs(gp, SelectNumbers([x for x in range(len(gp.X))]))
for x in inputs._funcs:
print(x, tag)
assert getattr(x, tag)
def test_selectNumbers():
from EvoDAG.cython_utils import SelectNumbers
s = SelectNumbers([x for x in range(100)])
a = s.get(10)
print(a, len(a))
assert len(a) == 10
a = s.get(91)
assert len(a) == 90
s.pos -= 10
a = s.get(10)
assert len(a) == 10
assert s.pos == 100
assert s.empty()
s.pos = 0
while not s.empty():
s.get_one()
s = SelectNumbers([])
assert s.empty()