[unittest] Test new Func1 features

This commit is contained in:
Ingmar Schoegl 2023-06-29 20:49:00 -06:00 committed by Ray Speth
parent 267ebb9022
commit 6471a2cd9a

View File

@ -2,6 +2,8 @@ import numpy as np
import cantera as ct
from . import utilities
import math
import pytest
class TestFunc1(utilities.CanteraTest):
def test_function(self):
@ -12,6 +14,7 @@ class TestFunc1(utilities.CanteraTest):
def test_lambda(self):
f = ct.Func1(lambda t: np.sin(t)*np.sqrt(t))
assert f.type == "functor"
for t in [0.1, 0.7, 4.5]:
self.assertNear(f(t), np.sin(t)*np.sqrt(t))
@ -24,6 +27,7 @@ class TestFunc1(utilities.CanteraTest):
m = Multiplier(8.1)
f = ct.Func1(m)
assert f.type == "functor"
for t in [0.1, 0.7, 4.5]:
self.assertNear(f(t), 8.1*t)
@ -31,6 +35,7 @@ class TestFunc1(utilities.CanteraTest):
f = ct.Func1(5)
for t in [0.1, 0.7, 4.5]:
self.assertNear(f(t), 5)
assert f.type == "constant"
def test_sequence(self):
f = ct.Func1([5])
@ -42,7 +47,9 @@ class TestFunc1(utilities.CanteraTest):
def test_numpy(self):
f = ct.Func1(np.array(5))
assert f.type == "constant"
g = ct.Func1(np.array([[5]]))
assert g.type == "constant"
for t in [0.1, 0.7, 4.5]:
self.assertNear(f(t), 5)
self.assertNear(g(t), 5)
@ -70,11 +77,58 @@ class TestFunc1(utilities.CanteraTest):
with self.assertRaises(NotImplementedError):
copy.copy(f)
def test_simple(self):
functors = {
'sin': math.sin,
'cos': math.cos,
'exp': math.exp,
'log': math.log,
}
for name, fcn in functors.items():
coeff = 2.34
func = ct.Func1.cxx_functor(name, coeff)
assert func.type == name
for val in [.1, 1., 10.]:
assert name in func.write()
assert func(val) == pytest.approx(fcn(coeff * val))
def test_compound(self):
functors = {
'sum': lambda x, y: x + y,
'diff': lambda x, y: x - y,
'product': lambda x, y: x * y,
'ratio': lambda x, y: x / y,
}
f1 = ct.Func1.cxx_functor('pow', 2)
f2 = ct.Func1.cxx_functor('sin')
for name, fcn in functors.items():
func = ct.Func1.cxx_functor(name, f1, f2)
assert func.type == name
for val in [.1, 1., 10.]:
assert name not in func.write()
assert func(val) == pytest.approx(fcn(f1(val), f2(val)))
def test_modified(self):
functors = {
'plus-constant': lambda x, y: x + y,
'times-constant': lambda x, y: x * y,
}
f1 = ct.Func1.cxx_functor('sin')
constant = 2.34
for name, fcn in functors.items():
func = ct.Func1.cxx_functor(name, f1, constant)
assert func.type == name
for val in [.1, 1., 10.]:
assert name not in func.write()
assert func(val) == pytest.approx(fcn(f1(val), constant))
def test_tabulated1(self):
# this implicitly probes advanced functors
arr = np.array([[0, 2], [1, 1], [2, 0]])
time = arr[:, 0]
fval = arr[:, 1]
fcn = ct.TabulatedFunction(time, fval)
assert fcn.type == "tabulated-linear"
for t, f in zip(time, fval):
self.assertNear(f, fcn(t))
@ -82,6 +136,7 @@ class TestFunc1(utilities.CanteraTest):
time = [0, 1, 2]
fval = [2, 1, 0]
fcn = ct.TabulatedFunction(time, fval)
assert fcn.type == "tabulated-linear"
for t, f in zip(time, fval):
self.assertNear(f, fcn(t))
@ -105,15 +160,16 @@ class TestFunc1(utilities.CanteraTest):
time = [0, 1, 2]
fval = [2, 1, 0]
fcn = ct.TabulatedFunction(time, fval, method='previous')
assert fcn.type == "tabulated-previous"
val = np.array([fcn(v) for v in [-0.5, 0, 0.5, 1.5, 2, 2.5]])
self.assertArrayNear(val, np.array([2.0, 2.0, 2.0, 1.0, 0.0, 0.0]))
def test_tabulated_failures(self):
with self.assertRaisesRegex(ValueError, 'do not match'):
with pytest.raises(ct.CanteraError, match="even number of entries"):
ct.TabulatedFunction(range(2), range(3))
with self.assertRaisesRegex(ValueError, 'must not be empty'):
with pytest.raises(ct.CanteraError, match="at least 4 entries"):
ct.TabulatedFunction([], [])
with self.assertRaisesRegex(ct.CanteraError, 'monotonically'):
with pytest.raises(ct.CanteraError, match="monotonically"):
ct.TabulatedFunction((0, 1, 0.5, 2), (2, 1, 1, 0))
with self.assertRaisesRegex(NotImplementedError, 'not implemented'):
ct.TabulatedFunction((0, 1, 1, 2), (2, 1, 1, 0), method='not implemented')
with pytest.raises(ct.CanteraError, match="No such type"):
ct.TabulatedFunction((0, 1, 1, 2), (2, 1, 1, 0), method='spam')