def test_CSVAttributesProbe_3():
"""Changing the order of the attributes list changes the order of the columns."""
# Set up a population
# TODO simplify this with a test fixture
pop = data.test_population
pop, _ = op.evaluate(pop) # Evaluate its fitness
# Assign distinct alues to an attribute on each individual
attrs = [('foo', ['GREEN', 15, 'BLUE', 72.81]), \
('bar', ['Colorless', 'green', 'ideas', 'sleep']), \
('baz', [['a', 'b', 'c'], [1, 2, 3], [None, None, None], [0.1, 0.2, 0.3]])]
for attr, vals in attrs:
for (ind, val) in zip(pop, vals):
ind.attributes[attr] = val
# Setup a probe that writes to a str in memory
stream = io.StringIO()
probe = CSVAttributesProbe(
stream,
['bar', 'foo'
]) # Passing params in reverse order from the other test above
probe.set_step(10)
# Execute
probe(pop, None)
result = stream.getvalue()
stream.close()
# Test
expected = "step, bar, foo\n" + \
"10, Colorless, GREEN\n" + \
"10, green, 15\n" + \
"10, ideas, BLUE\n" + \
"10, sleep, 72.81\n"
assert (result == expected)
def test_CSVAttributesProbe_1():
"""When recording the attribute 'my_value' and leaving other arguments at their default,
running CSVProbe on a population of individuals with just the attribute 'my_value' should
produce the correct CSV-formatted output."""
# Set up a population
# TODO simplify this with a test fixture
pop = data.test_population
pop, _ = op.evaluate(pop) # Evaluate its fitness
# Assign distinct alues to an attribute on each individual
attr = 'my_value'
vals = ['GREEN', 15, 'BLUE', 72.81]
for (ind, val) in zip(pop, vals):
ind.attributes[attr] = val
# Setup a probe that writes to a str in memory
stream = io.StringIO()
probe = CSVAttributesProbe(stream, ['my_value'])
probe.set_step(10)
# Execute
probe(pop, None)
result = stream.getvalue()
stream.close()
# Test
expected = "step, my_value\n" + \
"10, GREEN\n" + \
"10, 15\n" + \
"10, BLUE\n" + \
"10, 72.81\n"
assert (result == expected)
def test_CSVAttributesProbe_2():
"""When recording the attribute 'my_value' and leaving other arguments at their default,
running CSVProbe on a population of individuals with several attributes should
produce CSV-formatted output that only records 'my_value'."""
# Set up a population
# TODO simplify this with a test fixture
pop = data.test_population
pop, _ = op.evaluate(pop) # Evaluate its fitness
# Assign distinct alues to an attribute on each individual
attrs = [('foo', ['GREEN', 15, 'BLUE', 72.81]), \
('bar', ['Colorless', 'green', 'ideas', 'sleep']), \
('baz', [['a', 'b', 'c'], [1, 2, 3], [None, None, None], [0.1, 0.2, 0.3]])]
for attr, vals in attrs:
for (ind, val) in zip(pop, vals):
ind.attributes[attr] = val
# Setup a probe that writes to a str in memory
stream = io.StringIO()
probe = CSVAttributesProbe(stream, ['foo', 'bar'])
probe.set_step(10)
# Execute
probe(pop, None)
result = stream.getvalue()
stream.close()
# Test
expected = "step, foo, bar\n" + \
"10, GREEN, Colorless\n" + \
"10, 15, green\n" + \
"10, BLUE, ideas\n" + \
"10, 72.81, sleep\n"
assert (result == expected)
def _build_test_pop():
"""Construct a synthetic population for illustrating example operations."""
pop = [
core.Individual([1, 0, 1, 1, 0], core.IdentityDecoder(),
binary.MaxOnes()),
core.Individual([0, 0, 1, 0, 0], core.IdentityDecoder(),
binary.MaxOnes()),
core.Individual([0, 1, 1, 1, 1], core.IdentityDecoder(),
binary.MaxOnes()),
core.Individual([1, 0, 0, 0, 1], core.IdentityDecoder(),
binary.MaxOnes())
]
pop, _ = op.evaluate(pop)
# Assign distinct values to an attribute on each individual
attrs = [('foo', ['GREEN', 15, 'BLUE', 72.81]),
('bar', ['Colorless', 'green', 'ideas', 'sleep']),
('baz', [['a', 'b', 'c'], [1, 2, 3], [None, None, None],
[0.1, 0.2, 0.3]])]
for attr, vals in attrs:
for (ind, val) in zip(pop, vals):
ind.attributes[attr] = val
return pop