def test_trackers():
""" test whether simple trackers can be used """
times = []
def store_time(state, t):
times.append(t)
def get_data(state):
return {"integral": state.integral}
devnull = open(os.devnull, "w")
data = trackers.DataTracker(get_data, interval=0.1)
tracker_list = [
trackers.PrintTracker(interval=0.1, stream=devnull),
trackers.CallbackTracker(store_time, interval=0.1),
None, # should be ignored
data,
]
if module_available("matplotlib"):
tracker_list.append(trackers.PlotTracker(interval=0.1, show=False))
grid = UnitGrid([16, 16])
state = ScalarField.random_uniform(grid, 0.2, 0.3)
pde = DiffusionPDE()
pde.solve(state, t_range=1, dt=0.005, tracker=tracker_list)
devnull.close()
assert times == data.times
if module_available("pandas"):
df = data.dataframe
np.testing.assert_allclose(df["time"], times)
np.testing.assert_allclose(df["integral"], state.integral)
def test_steady_state_tracker():
""" test the SteadyStateTracker """
storage = MemoryStorage()
c0 = ScalarField.random_uniform(UnitGrid([5]))
pde = DiffusionPDE()
tracker = trackers.SteadyStateTracker(atol=1e-2, rtol=1e-2, progress=True)
pde.solve(c0, 1e3, dt=0.1, tracker=[tracker, storage.tracker(interval=1e2)])
assert len(storage) < 9 # finished early
def test_small_tracker_dt():
"""test the case where the dt of the tracker is smaller than the dt
of the simulation"""
storage = MemoryStorage()
pde = DiffusionPDE()
c0 = ScalarField.random_uniform(UnitGrid([4, 4]), 0.1, 0.2)
pde.solve(
c0, 1e-2, dt=1e-3, method="explicit", tracker=storage.tracker(interval=1e-4)
)
assert len(storage) == 11
def test_plot_movie_tracker(tmp_path):
""" test whether the plot tracker creates files without errors """
output_file = tmp_path / "movie.mov"
grid = UnitGrid([4, 4])
state = ScalarField.random_uniform(grid)
pde = DiffusionPDE()
tracker = trackers.PlotTracker(movie=output_file, interval=0.1, show=False)
pde.solve(state, t_range=0.5, dt=0.005, tracker=tracker, backend="numpy")
assert output_file.stat().st_size > 0
def test_runtime_tracker():
""" test the RuntimeTracker """
s = ScalarField.random_uniform(UnitGrid([128]))
tracker = trackers.RuntimeTracker("0:01")
sol = ExplicitSolver(DiffusionPDE())
con = Controller(sol, t_range=1e4, tracker=["progress", tracker])
con.run(s, dt=1e-3)
def test_consistency_tracker():
""" test the ConsistencyTracker """
s = ScalarField.random_uniform(UnitGrid([128]))
sol = ExplicitSolver(DiffusionPDE(1e3))
con = Controller(sol, t_range=1e5, tracker=["consistency"])
with np.errstate(all="ignore"):
con.run(s, dt=1)
assert con.info["t_final"] < con.info["t_end"]
def test_plot_tracker(tmp_path):
""" test whether the plot tracker creates files without errors """
output_file = tmp_path / "img.png"
def get_title(state, t):
return f"{state.integral:g} at {t:g}"
grid = UnitGrid([4, 4])
state = ScalarField.random_uniform(grid)
pde = DiffusionPDE()
tracker = trackers.PlotTracker(
output_file=output_file, title=get_title, interval=0.1, show=False
)
pde.solve(state, t_range=0.5, dt=0.005, tracker=tracker, backend="numpy")
assert output_file.stat().st_size > 0
def test_data_tracker(tmp_path):
""" test the DataTracker """
field = ScalarField(UnitGrid([4, 4]))
eq = DiffusionPDE()
path = tmp_path / "test_data_tracker.pickle"
data1 = trackers.DataTracker(lambda f: f.average, filename=path)
data2 = trackers.DataTracker(lambda f: {"avg": f.average, "int": f.integral})
eq.solve(field, 10, tracker=[data1, data2])
with path.open("br") as fp:
time, data = pickle.load(fp)
np.testing.assert_allclose(time, np.arange(11))
assert isinstance(data, list)
assert len(data) == 11
assert path.stat().st_size > 0
def test_movie_scalar(movie_func, tmp_path):
"""test Movie class"""
# create some data
state = ScalarField.random_uniform(UnitGrid([4, 4]))
eq = DiffusionPDE()
storage = MemoryStorage()
tracker = storage.tracker(interval=1)
eq.solve(state, t_range=2, dt=1e-2, backend="numpy", tracker=tracker)
# check creating the movie
path = tmp_path / "test_movie.mov"
try:
movie_func(storage, filename=path, progress=False)
except RuntimeError:
pass # can happen when ffmpeg is not installed
else:
assert path.stat().st_size > 0
def test_callback_tracker():
""" test trackers that support a callback """
data = []
def store_mean_data(state):
data.append(state.average)
def get_mean_data(state):
return state.average
grid = UnitGrid([4, 4])
state = ScalarField.random_uniform(grid, 0.2, 0.3)
pde = DiffusionPDE()
data_tracker = trackers.DataTracker(get_mean_data, interval=0.1)
callback_tracker = trackers.CallbackTracker(store_mean_data, interval=0.1)
tracker_list = [data_tracker, callback_tracker]
pde.solve(state,
t_range=0.5,
dt=0.005,
tracker=tracker_list,
backend="numpy")
np.testing.assert_array_equal(data, data_tracker.data)
data = []
def store_time(state, t):
data.append(t)
def get_time(state, t):
return t
grid = UnitGrid([4, 4])
state = ScalarField.random_uniform(grid, 0.2, 0.3)
pde = DiffusionPDE()
data_tracker = trackers.DataTracker(get_time, interval=0.1)
tracker_list = [
trackers.CallbackTracker(store_time, interval=0.1), data_tracker
]
pde.solve(state,
t_range=0.5,
dt=0.005,
tracker=tracker_list,
backend="numpy")
ts = np.arange(0, 0.55, 0.1)
np.testing.assert_allclose(data, ts, atol=1e-2)
np.testing.assert_allclose(data_tracker.data, ts, atol=1e-2)
