def test_normal_refraction(self):
node = Node(name="node")
ctx = Context(n1=1.0, n2=1.5, normal_node=node, normal=(0.0, 0.0, 1.0), kind=Kind.SURFACE, end_path=(10, 10, 10), container=None)
ray = Ray(position=(0.0, 0.0, 0.0), direction=(0.0, 0.0, 1.0), wavelength=None)
interaction = FresnelRefraction(ctx)
new_ray = interaction.transform(ray)
assert ray == new_ray
def test_normal_reflection(self):
n1 = 1.0
n2 = 1.5
normal = (0.0, 0.0, 1.0)
angle = 0.0
ray = Ray(position=(0.0, 0.0, 0.0),
direction=(0.0, 0.0, 1.0),
wavelength=None)
fresnel = FresnelRefraction()
new_ray = fresnel.transform(ray, {
"n1": n1,
"n2": n2,
"normal": normal
})
assert np.allclose(ray.direction, new_ray.direction)
def test_antinormal_reflection(self):
""" FresnelReflection takes the smallest angle between the ray direction and
the normal. Thus the flipped normal will also work.
"""
n1 = 1.0
n2 = 1.5
normal = (0.0, 0.0, -1.0)
angle = 0.0
ray = Ray(position=(0.0, 0.0, 0.0),
direction=(0.0, 0.0, 1.0),
wavelength=None)
fresnel = FresnelRefraction()
new_ray = fresnel.transform(ray, {
"n1": n1,
"n2": n2,
"normal": normal
})
assert np.allclose(ray.direction, new_ray.direction)
class Dielectric(Refractive, Material):
""" A material with a refractive index.
Notes
-----
The material is unphysical in the sense that it does not absorb or emit light.
But it is useful in development and testing to have material which just
interacts with ray without in a purely refractive way.
"""
def __init__(self, refractive_index):
super(Dielectric, self).__init__(refractive_index)
self._transit_mechanism = FresnelRefraction()
self._return_mechanism = FresnelReflection()
self._path_mechanism = TravelPath()
self._emit_mechanism = None
def trace_path(self, local_ray: "Ray", container_geometry: "Geometry",
distance: float):
""" Dielectric material does not have any absorption; this moves ray full dist.
"""
new_ray = self._path_mechanism.transform(local_ray,
{"distance": distance})
yield new_ray, Decision.TRAVEL
def trace_surface(
self,
local_ray: "Ray",
container_geometry: "Geometry",
to_geometry: "Geometry",
surface_geometry: "Geometry",
) -> Tuple[Decision, dict]:
"""
"""
# Get reflectivity for the ray
normal = surface_geometry.normal(local_ray.position)
n1 = container_geometry.material.refractive_index(local_ray.wavelength)
n2 = to_geometry.material.refractive_index(local_ray.wavelength)
# Be flexible with how the normal is defined
if np.dot(normal, local_ray.direction) < 0.0:
normal = flip(normal)
angle = angle_between(normal, np.array(local_ray.direction))
if angle < 0.0 or angle > 0.5 * np.pi:
raise TraceError("The incident angle must be between 0 and pi/2.")
incident = local_ray.direction
reflectivity = self._return_mechanism.reflectivity(angle, n1, n2)
#print("Reflectivity: {}, n1: {}, n2: {}, angle: {}".format(reflectivity, n1, n2, angle))
gamma = np.random.uniform()
info = {"normal": normal, "n1": n1, "n2": n2}
# Pick between reflection (return) and transmission (transit)
if gamma < reflectivity:
new_ray = self._return_mechanism.transform(local_ray, info)
decision = Decision.RETURN
yield new_ray, decision
else:
new_ray = self._transit_mechanism.transform(local_ray, info)
decision = Decision.TRANSIT
yield new_ray, decision
def trace_surface(
self,
local_ray: "Ray",
container_geometry: "Geometry",
to_geometry: "Geometry",
surface_geometry: "Geometry",
) -> Tuple[Decision, dict]:
"""
"""
# Get reflectivity for the ray
normal = surface_geometry.normal(local_ray.position)
n1 = container_geometry.material.refractive_index(local_ray.wavelength)
n2 = to_geometry.material.refractive_index(local_ray.wavelength)
# Be flexible with how the normal is defined
if np.dot(normal, local_ray.direction) < 0.0:
normal = flip(normal)
angle = angle_between(normal, np.array(local_ray.direction))
if angle < 0.0 or angle > 0.5 * np.pi:
raise TraceError("The incident angle must be between 0 and pi/2.")
incident = local_ray.direction
reflectivity = self._return_mechanism.reflectivity(angle, n1, n2)
#print("Reflectivity: {}, n1: {}, n2: {}, angle: {}".format(reflectivity, n1, n2, angle))
gamma = np.random.uniform()
info = {"normal": normal, "n1": n1, "n2": n2}
# Pick between reflection (return) and transmission (transit)
if gamma < reflectivity:
new_ray = self._return_mechanism.transform(local_ray, info)
decision = Decision.RETURN
yield new_ray, decision
else:
new_ray = self._transit_mechanism.transform(local_ray, info)
decision = Decision.TRANSIT
yield new_ray, decision
@classmethod
def make_constant(cls, x_range: Tuple[float, float],
refractive_index: float):
""" Returns a dielectric material with spectrally constant refractive index.
"""
refractive_index = np.column_stack(
(x_range, [refractive_index, refractive_index]))
return cls(refractive_index)
@classmethod
def air(cls, x_range: Tuple[float, float] = (300.0, 4000.0)):
""" Returns a dielectric material with constant refractive index of 1.0 in
default range.
"""
return cls.make_constant(x_range=x_range, refractive_index=1.0)
@classmethod
def glass(cls, x_range: Tuple[float, float] = (300.0, 4000.0)):
""" Returns a dielectric material with constant refractive index of 1.5 in
default range.
"""
return cls.make_constant(x_range=x_range, refractive_index=1.5)
class Host(Refractive, Blend, Material):
""" A material with a refractive index that can host a single or multiple
Lumophores.
"""
def __init__(self, refractive_index: np.ndarray, lumophores: [Lumophore]):
super(Host, self).__init__(
refractive_index=refractive_index,
lumophores=lumophores
)
self._transit_mechanism = FresnelRefraction()
self._return_mechanism = FresnelReflection()
self._path_mechanism = Absorption()
self._emit_mechanism = Emission()
def select_lumophore(self, nanometers: float) -> Lumophore:
""" Selects, at random, one of the lumophores from the list.
Parameters
----------
nanometers : float
The wavelength of the interacting photon in nanometers.
Returns
-------
Lumophore
The lumophore which absorbed the photon.
Notes
-----
The selection is weighted by the relative absorption strength of all
materials at the given wavelength.
"""
absorptions = [l.absorption_coefficient(nanometers) for l in self.lumophores]
count = len(self.lumophores)
bins = list(range(0, count + 1))
cdf = np.cumsum(absorptions)
pdf = cdf / max(cdf)
pdf = np.hstack([0, pdf[:]])
pdfinv_lookup = np.interp(np.random.uniform(), pdf, bins)
absorber_index = int(np.floor(pdfinv_lookup))
lumophore = self.lumophores[absorber_index]
return lumophore
def trace_surface(
self,
local_ray: "Ray",
container_geometry: "Geometry",
to_geometry: "Geometry",
surface_geometry: "Geometry",
) -> Tuple[Decision, dict]:
"""
"""
# Ray both materials need a refractive index to compute Frensel reflection;
# if they are not the both refractive then just let ray cross the interface.
try:
normal = surface_geometry.normal(local_ray.position)
except Exception:
import pdb; pdb.set_trace()
if not all([isinstance(x, Refractive) for x in (container_geometry.material, to_geometry.material)]):
new_ray = CrossInterface().transform(local_ray, {"normal": normal})
yield new_ray, Decision.TRANSIT
return
# Get reflectivity for the ray
n1 = container_geometry.material.refractive_index(local_ray.wavelength)
n2 = to_geometry.material.refractive_index(local_ray.wavelength)
# Be flexible with how the normal is defined
if np.dot(normal, local_ray.direction) < 0.0:
normal = flip(normal)
angle = angle_between(normal, np.array(local_ray.direction))
if angle < 0.0 or angle > 0.5 * np.pi:
raise TraceError("The incident angle must be between 0 and pi/2.")
incident = local_ray.direction
reflectivity = self._return_mechanism.reflectivity(angle, n1, n2)
#print("Reflectivity: {}, n1: {}, n2: {}, angle: {}".format(reflectivity, n1, n2, angle))
gamma = np.random.uniform()
info = {"normal": normal, "n1": n1, "n2": n2}
# Pick between reflection (return) and transmission (transit)
if gamma < reflectivity:
new_ray = self._return_mechanism.transform(local_ray, info)
decision = Decision.RETURN
yield new_ray, decision
else:
new_ray = self._transit_mechanism.transform(local_ray, info)
decision = Decision.TRANSIT
yield new_ray, decision
def trace_path(
self,
local_ray: "Ray",
container_geometry: "Geometry",
distance: float
) -> Tuple[Decision, dict]:
# Which of the host's materials captured the ray
material = self.select_lumophore(local_ray.wavelength)
# Sample the exponential distribution and get a distance at which the
# ray is absorbed.
sampled_distance = self._path_mechanism.path_length(
local_ray.wavelength, material
)
logger.debug("Host.trace_path args: {}".format((local_ray, container_geometry, distance)))
# If the sampled distance is less than the full distance the ray can travel
# then the ray is absorbed.
if sampled_distance < distance:
# Apply the absorption transformation to the ray; this updates the rays
# position to the absorption location.
info = {"distance": sampled_distance}
#print("Sampled pathlength: {}".format(info))
new_ray = self._path_mechanism.transform(local_ray, info)
# Test if ray is reemitted by comparing a random number to the quantum yield
qy = material.quantum_yield
# If the random number is less than the quantum yield then emission occurs.
if np.random.uniform() < qy:
# If ray is re-emitted generate two events: ABSORB and EMIT
yield new_ray, Decision.ABSORB
# Emission occurred
new_ray = self._emit_mechanism.transform(new_ray, {"material": material})
yield new_ray, Decision.EMIT
else:
# If the ray is not emitted generate one event: ABSORB
# Non-radiative absorption occurred
new_ray = replace(new_ray, is_alive=False)
yield new_ray, Decision.ABSORB
else:
# If not absorbed travel the full distance
info = {"distance": distance}
new_ray = self._path_mechanism.transform(local_ray, info)
yield new_ray, Decision.TRAVEL
@classmethod
def from_dataframe(cls, df):
""" Returns a Host instance initialised with data from the dataframe.
Parameters
----------
df : pandas.DataFrame
The dataframe must start with the following columns (an index column is
fine and will be ignored),
- *wavelength <anything>*, the wavelength in nanometers
- *refractive index <anything>*, the real part of the refractive index
Then the subsequent columns must be absorption coefficient, emission
spectrum, and quantum yield pairs for the individual lumophores that are
added to the host,
- *absorption coefficient <anything>*, the absorption coefficient of the
lumophore in units of 1/cm
- *emission spectrum <anything>*, the emission spectrum of lumophore in
arbitrary units (just the shape is important).
- *quantum yield <anything>*, the quantum yield of the lumophore. Note
that the wavelength dependence of the quantum yield is *not* used. The
first value in the column is used as the quantum yield.
To validate the dataframe only the start of the column names are checked
so anything after the required part is ignored and can be what you like
so that you can organise your data.
"""
columns = df.columns.tolist()
if tuple(columns[0:2]) != ("wavelength", "refractive index"):
raise AppError(
"Data frame is wrong format. The first two column names "
"be 'wavelength' and 'refractive index'."
)
wavelength = df["wavelength"].values
ior = df["refractive index"].values
col_err_msg = (
"Column {} in data frame has incorrect name. Got {} and "
"expected it to start with {}."
)
lumo_cols = columns[2:]
for idx, name in enumerate(lumo_cols):
name = name.lower()
col_idx = idx + 2
if idx % 3 == 0:
expected = "absorption coefficient"
if not name.startswith(expected):
raise AppError(col_err_msg.format(col_idx, name, expected))
elif idx % 3 == 1:
expected = "emission spectrum"
if not name.startswith(expected):
raise AppError(col_err_msg.format(col_idx, name, expected))
elif idx % 3 == 2:
expected = "quantum yield"
if not name.startswith(expected):
raise AppError(col_err_msg.format(col_idx, name, expected))
if len(lumo_cols) % 3 != 0:
raise AppError(
"Data column(s) missing. Columns after the first two should "
"be groups like (absorption coefficient, emission_spectrum "
", quantum_yield)."
)
refractive_index = np.column_stack((wavelength, ior))
from itertools import zip_longest
def grouper(iterable):
args = [iter(iterable)] * 3
return zip_longest(*args)
lumophores = []
for (alpha_name, emission_name, qy_name) in grouper(lumo_cols):
absorption_coefficient = np.column_stack(
(wavelength, df[alpha_name].values)
)
emission_spectrum = np.column_stack((wavelength, df[emission_name].values))
quantum_yield = df[qy_name].values[0]
lumo = Lumophore(absorption_coefficient, emission_spectrum, quantum_yield)
logger.debug(
"Making lumophore with max absorption coefficient {}".format(
np.max(absorption_coefficient[:, 1])
)
)
lumophores.append(lumo)
host = cls(refractive_index, lumophores)
return host