def optimize_mixture_model(index, init=None):
# Select indices and get data.
indices = finite_indices[npm.query_around_point(kdt, X[index], **kdt_kwds)]
y = np.array([data[ln][indices]
for ln in config["predictor_label_names"]]).T
if init is None:
init = npm.get_initialization_point(y)
opt_kwds = dict(init=init, data=dict(y=y, N=y.shape[0], D=y.shape[1]))
opt_kwds.update(default_opt_kwds)
# Do optimization.
with stan.suppress_output():
try:
p_opt = npm._check_params_dict(model.optimizing(**opt_kwds))
except:
p_opt = None
return (index, p_opt, indices)
def optimize_mixture_model(index, init=None):
# Select indices and get data.
indices = finite_indices[npm.query_around_point(kdt, X[index], **kdt_kwds)]
y = np.array([data[ln][indices] for ln in config["predictor_label_names"]]).T
if init is None:
init = npm.get_initialization_point(y)
data_dict = dict(y=y,
N=y.shape[0],
D=y.shape[1],
max_log_y=np.log(np.max(y)))
for k, v in bounds.items():
data_dict["{}_bounds".format(k)] = v
trial_results = []
for j, init_dict in enumerate((init, npm.get_initialization_point(y), "random")):
# CHeck that the parameters are bounded?
if isinstance(init_dict, dict):
if bounds is not None:
for k, (lower, upper) in bounds.items():
if not (upper > init_dict[k] > lower):
logging.info("Clipping initial value of {} from {} to within ({}, {})".format(
k, init_dict[k], lower, upper))
offset = 0.01 * (upper - lower)
init_dict[k] = np.clip(init_dict[k], lower + offset, upper - offset)
opt_kwds = dict(
init=init_dict,
data=data_dict)
opt_kwds.update(default_opt_kwds)
# Do optimization.
with stan.suppress_output() as sm:
try:
p_opt = model.optimizing(**opt_kwds)
except:
p_opt = None
# TODO: Consider relaxing the optimization tolerances!
if p_opt is None:
# Capture stdout and stderr so we can read it later.
stdout, stderr = sm.stdout, sm.stderr
trial_results.append(p_opt)
if p_opt is None:
logging.warning("Exception when optimizing on index {} from "\
"initial point {}:".format(index, init_dict))
logging.warning(stdout)
logging.warning(stderr)
raise
else:
tolerance = 1e-2
for k, v in p_opt.items():
if k not in bounds \
or (k == "theta" and (v >= 1-tolerance)): continue
lower, upper = bounds[k]
if np.abs(v - lower) <= tolerance \
or np.abs(v - upper) <= tolerance:
logging.warning("Optimised {} at edge of grid ({} < {} < {})"
" - ignoring".format(k, lower, v, upper))
break
else:
break
else:
# TODO: Consider relaxing optimization tolerances!
logging.warning("Optimization did not converge from any initial point "\
"trialled. Consider relaxing optimization tolerances!")
import matplotlib.pyplot as plt
fig, ax = plt.subplots()
ax.hist(y, bins=100, facecolor="#cccccc", normed=True)
# Plot a model at some optimization point.
init = npm.get_initialization_point(y)
N, D = y.shape
xi = np.linspace(0, max(y), 1000)
ax.plot(xi, npm.norm_pdf(xi, init["mu_single"], init["sigma_single"],
init["theta"]), c="r")
ax.plot(xi, npm.lognorm_pdf(xi, init["mu_multiple"], init["sigma_multiple"],
init["theta"]), c="r", linestyle=":")
ax.plot(xi, npm.norm_pdf(xi, default_init["mu_single"], default_init["sigma_single"],
default_init["theta"]), c="b")
ax.plot(xi, npm.lognorm_pdf(xi, default_init["mu_multiple"], default_init["sigma_multiple"],
default_init["theta"]), c="b", linestyle=":")
raise a
p_opt = npm._check_params_dict(p_opt)
return (index, p_opt, indices)
def optimize_mixture_model(index, inits=None, scalar=5):
# Select indices and get data.
d, nearby_idx, meta = npm.query_around_point(
kdt, X[index], **kdt_kwds)
y = Y[nearby_idx]
ball = X[nearby_idx]
if inits is None:
inits = npm._get_1d_initialisation_point(
y, scalar=scalar, bounds=model_config["bounds"])
# Update meta dictionary with things about the data.
meta = dict(max_log_y=np.log(np.max(y)),
N=nearby_idx.size,
y_percentiles=np.percentile(y, [16, 50, 84]),
ball_ptps=np.ptp(ball, axis=0),
ball_medians=np.median(ball, axis=0),
init_points=inits,
kdt_indices=nearby_idx)
data_dict = dict(y=y, N=y.size, scalar=scalar)
data_dict.update(bounds)
#for k, v in model_config["parameter_bounds"].items():
# data_dict["{}_bounds".format(k)] = v
p_opts = []
ln_probs = []
for j, init_dict in enumerate(inits):
opt_kwds = dict(init=init_dict, data=data_dict)
opt_kwds.update(default_opt_kwds)
# Do optimization.
# TODO: Suppressing output is always dangerous.
with stan.suppress_output(
config.get("suppress_stan_output", True)) as sm:
try:
p_opt = model.optimizing(**opt_kwds)
except:
logging.exception(f"Exception occurred when optimizing index {index}"\
f" from {init_dict}:")
else:
if p_opt is not None:
p_opts.append(p_opt)
ln_probs.append(
npm.ln_prob(y, 1, *npm._pack_params(**p_opt)))
try:
p_opt
except UnboundLocalError:
logging.warning("Stan failed. STDOUT & STDERR:")
logging.warning("\n".join(sm.outputs))
if p_opt is None:
stdout, stderr = sm.outputs
logging.warning(f"STDOUT:\n{stdout}\nSTDERR:\n{stderr}")
if len(p_opts) < 1:
logging.warning("Optimization on index {} did not converge from any "\
"initial point trialled. Consider relaxing the "\
"optimization tolerances! If this occurs regularly "\
"then something is very wrong!".format(index))
return (index, None, meta)
# evaluate best.
else:
idx = np.argmax(ln_probs)
p_opt = p_opts[idx]
meta["init_idx"] = idx
"""
if sum(done) > 550 and sum(done) < 570:
theta, mu_single, sigma_single, mu_multiple, sigma_multiple = npm._pack_params(**p_opt)
fig, ax = plt.subplots()
xi = np.linspace(0, 20, 1000)
y_s = npm.norm_pdf(xi, mu_single, sigma_single, theta)
y_m = npm.lognorm_pdf(xi, mu_multiple, sigma_multiple, theta)
ax.plot(xi, y_s, c="tab:blue")
ax.plot(xi, y_m, c="tab:red")
p_single = np.exp(np.log(y_s) - logsumexp([np.log(y_s), np.log(y_m)], axis=0))
ax.plot(xi, p_single, c="k")
ax.set_title(f"{index}: {theta:.1e} {mu_single:.2f} {sigma_single:.2f} {sigma_multiple:.2f}")
ax.hist(y, bins=np.linspace(0, 20, 20), alpha=0.5, facecolor="#666666", normed=True)
if sum(done) > 570:
raise a
"""
return (index, p_opt, meta)
def optimize_mixture_model(index, inits=None, debug=False):
suppress = config.get("suppress_stan_output", True)
# Select indices and get data.
d, nearby_idx, meta = npm.query_around_point(
kdt, X[index], **kdt_kwds)
y = Y[nearby_idx]
ball = X[nearby_idx]
if inits is None:
inits = npm._get_1d_initialisation_point(
y, scalar=mu_multiple_scalar, bounds=bounds)
# Update meta dictionary with things about the data.
meta = dict(max_log_y=np.log(np.max(y)),
N=nearby_idx.size,
y_percentiles=np.percentile(y, [16, 50, 84]),
ball_ptps=np.ptp(ball, axis=0),
ball_medians=np.median(ball, axis=0),
init_points=inits,
kdt_indices=nearby_idx)
data_dict = dict(y=y, N=y.size, scalar=mu_multiple_scalar)
data_dict.update(stan_bounds)
p_opts = []
ln_probs = []
for j, init_dict in enumerate(inits):
opt_kwds = dict(init=init_dict,
data=data_dict,
as_vector=False)
opt_kwds.update(default_opt_kwds)
# Do optimization.
# TODO: Suppressing output is always dangerous.
with stan.suppress_output(suppress) as sm:
try:
p_opt = model.optimizing(**opt_kwds)
except:
logger.exception(f"Exception occurred when optimizing index {index}"\
f" from {init_dict}:")
else:
if p_opt is not None:
p_opts.append(p_opt["par"])
ln_probs.append(
utils.ln_prob(
y,
1,
*utils._pack_params(**p_opt["par"]),
bounds=bounds))
assert abs(ln_probs[-1] - p_opt["value"]) < 1e-8
try:
p_opt
except UnboundLocalError:
logger.warning("Stan failed. STDOUT & STDERR:")
logger.warning("\n".join(sm.outputs))
else:
if p_opt is None:
stdout, stderr = sm.outputs
logger.warning("Stan only returned p_opt = None")
logger.warning(f"STDOUT:\n{stdout}\nSTDERR:\n{stderr}")
if len(p_opts) < 1:
logger.warning(f"Optimization on index {index} did not converge"\
"from any initial point trialled. Consider "\
"relaxing the optimization tolerances! If this "\
"occurs regularly then something is very wrong!")
return (index, None, meta)
else:
# evaluate best.
idx = np.argmax(ln_probs)
p_opt = p_opts[idx]
meta["init_idx"] = idx
"""
# Calculate uncertainties.
op_bounds = ()
def nlp(p):
w, mu_s, sigma_s, sigma_m = p
mu_m = np.log(mu_s + mu_multiple_scalar * sigma_s) + sigma_m**2
if not (bounds["theta"][1] >= w >= bounds["theta"][0]) \
or not (bounds["mu_single"][1] >= mu_s >= bounds["mu_single"][0]) \
or not (bounds["sigma_multiple"][1] >= sigma_m >= bounds["sigma_multiple"][0]):
return np.inf
return -utils.ln_likelihood(y, w, mu_s, sigma_s, mu_m, sigma_m)
op_bounds = [bounds["theta"],
bounds["mu_single"],
bounds["sigma_single"],
bounds["sigma_multiple"],
]
#x0 = utils._pack_params(**p_opt)
x0 = (p_opt["theta"], p_opt["mu_single"], p_opt["sigma_single"], p_opt["sigma_multiple"])
p_opt2 = op.minimize(nlp, x0, bounds=op_bounds, method="L-BFGS-B")
"""
# Create a three-panel figure showing:
# (1) a log-density of the HRD + the selected ball points
# (2) a log-density of colour vs apparent magnitude + the selected ball points
# (3) the jitter + fitted parameters
if sampling:
chains = 2 # TODO: move to config file.
sampling_kwds = dict(data=opt_kwds["data"],
init=[p_opt] * chains,
chains=chains)
try:
samples = model.sampling(**sampling_kwds)
except:
None
else:
extracted = samples.extract()
chains = np.array(
[extracted[k] for k in samples.flatnames]).T
latex_labels = dict(
theta=r"$w$",
mu_single=r"$\mu_\mathrm{single}$",
sigma_single=r"$\sigma_\mathrm{single}$",
mu_multiple=r"$\mu_\mathrm{multiple}$",
sigma_multiple=r"$\sigma_\mathrm{multiple}$")
corner_fig = corner.corner(
chains,
labels=[
latex_labels[k] for k in samples.flatnames
])
source_id = S[index]
figure_path = os.path.join(
figures_dir,
f"{model_name}-{source_id}-samples.png")
corner_fig.savefig(figure_path, dpi=150)
chains_path = os.path.join(
figures_dir,
f"{model_name}-{source_id}-chains.pkl")
dump = dict(names=samples.flatnames,
chains=chains,
y=y,
ball=ball,
X=X[index])
with open(chains_path, "wb") as fp:
pickle.dump(dump, fp)
plt.close("all")
if plot_mixture_model_figures:
source_id = S[index]
figure_path = os.path.join(
figures_dir, f"{model_name}-{source_id}.png")
x_upper = 2 * config["models"][model_name]["bounds"][
"mu_single"][1]
bins = np.linspace(0, x_upper, 51)
xi = np.linspace(0, x_upper, 1000)
y_s = utils.norm_pdf(xi, p_opt["mu_single"],
p_opt["sigma_single"], p_opt["theta"])
y_m = utils.lognorm_pdf(xi, p_opt["mu_multiple"],
p_opt["sigma_multiple"],
p_opt["theta"])
items_for_deletion = [
axes[0].scatter(ball.T[0],
ball.T[1],
c="tab:blue",
s=1,
zorder=10,
alpha=0.5),
axes[1].scatter(ball.T[0],
ball.T[2],
c="tab:blue",
s=1,
zorder=10,
alpha=0.5),
axes[2].hist(y,
bins=bins,
facecolor="#cccccc",
density=True,
zorder=-1)[-1],
axes[2].axvline(Y[index], c="#666666"),
axes[2].plot(xi, y_s, c="tab:blue"),
axes[2].fill_between(xi,
np.zeros_like(y_s),
y_s,
facecolor="tab:blue",
alpha=0.25),
axes[2].plot(xi, y_m, c="tab:red"),
axes[2].fill_between(xi,
np.zeros_like(y_m),
y_m,
facecolor="tab:red",
alpha=0.25),
]
# Ax limits.
axes[0].set_xlim(-0.5, 5)
axes[0].set_ylim(10, -15)
axes[1].set_xlim(-0.5, 5)
axes[1].set_ylim(15, 3)
axes[2].set_xlim(0, x_upper)
axes[2].set_yticks([])
fig.tight_layout()
fig.savefig(figure_path, dpi=150)
for item in items_for_deletion:
try:
item.set_visible(False)
except AttributeError:
for _ in item:
if hasattr(_, "set_visible"):
_.set_visible(False)
if debug:
# Create
raise a
return (index, p_opt, meta)
def optimize_mixture_model(index, inits=None, debug=False):
# Select indices and get data.
d, nearby_idx, meta = npm.query_around_point(kdt, X[index], **kdt_kwds)
y = Y[nearby_idx]
ball = X[nearby_idx]
if y.size < kdt_kwds.get("minimum_points", np.inf):
logger.warning(f"Danger: minimum number of points not found ({y.size})")
if inits is None:
inits = npm.get_1d_initialisation_point(y, scalar=mu_multiple_scalar, bounds=bounds)
# Update meta dictionary with things about the data.
meta = dict()
if debug:
meta.update(N=nearby_idx.size,
y_percentiles=np.percentile(y, [16, 50, 84]),
ball_ptps=np.ptp(ball, axis=0),
ball_medians=np.median(ball, axis=0),
init_points=inits,
kdt_indices=nearby_idx)
data_dict = dict(y=y, N=y.size, mu_multiple_scalar=mu_multiple_scalar)
data_dict.update(stan_bounds)
p_opts = []
ln_probs = []
for j, init_dict in enumerate(inits):
opt_kwds = dict(init=init_dict, data=data_dict, as_vector=False)
opt_kwds.update(default_opt_kwds)
with stan.suppress_output(suppress_output=(not debug)) as sm:
try:
p_opt = model.optimizing(**opt_kwds)
except:
logger.exception(f"Exception occurred when optimizing index {index}"\
f" from {init_dict}:")
else:
if p_opt is not None:
p_opts.append(p_opt["par"])
ln_probs.append(p_opt["value"])
'''
s = np.argsort(y)
fig, ax = plt.subplots()
ax.plot(y[s], p_opt["par"]["ll_s"][s], c="tab:blue")
ax.plot(y[s], p_opt["par"]["ll_m"][s], c="tab:red")
if np.random.choice(200, 1)[0] == 42:
raise a
'''
try:
p_opt
except UnboundLocalError:
stdout, stderr = sm.outputs
logger.warning("Stan failed. STDOUT:")
logger.warning(stdout)
logger.warning("STDERR:")
logger.warning(stderr)
else:
if p_opt is None:
stdout, stderr = sm.outputs
logger.warning("Stan only returned p_opt = None")
logger.warning(f"STDOUT:\n{stdout}\nSTDERR:\n{stderr}")
if len(p_opts) < 1:
logger.warning(f"Optimization on index {index} did not converge"\
"from any initial point trialled. Consider "\
"relaxing the optimization tolerances! If this "\
"occurs regularly then something is very wrong!")
return (index, None, meta)
else:
# evaluate best.
idx = np.argmax(ln_probs)
p_opt = p_opts[idx]
meta["init_idx"] = idx
return (index, p_opt, meta)