def xy_fitted_joint_likelihood(xy_model_and_datalist):
model, data = xy_model_and_datalist
jl = JointLikelihood(model, data)
res_frame, like_frame = jl.fit()
return jl, res_frame, like_frame
def test_fit():
model, datalist= get_model_and_datalist()
jl = JointLikelihood(model, datalist)
jl.fit()
_ = display_photometry_model_magnitudes(jl)
def restore_best_fit_model(self, interval):
# Get sub-frame containing the results for the requested interval
sub_frame = self._data_frame.loc[interval]
# Get the model for this interval
this_model = self._get_model(interval)
# Get data for this interval
this_data = self._get_data(interval)
# Instance a useless joint likelihood object so that plugins have the chance to add nuisance parameters to the
# model
_ = JointLikelihood(this_model, this_data)
# Restore best fit parameters
for parameter in this_model.free_parameters:
this_model[parameter].value = sub_frame["value"][parameter]
return this_model, this_data
def fit(self, function, minimizer='minuit', verbose=False):
"""
Fit the data with the provided function (an astromodels function)
:param function: astromodels function
:param minimizer: the minimizer to use
:param verbose: print every step of the fit procedure
:return: best fit results
"""
# This is a wrapper to give an easier way to fit simple data without having to go through the definition
# of sources
pts = PointSource("source", 0.0, 0.0, function)
model = Model(pts)
self.set_model(model)
self._joint_like_obj = JointLikelihood(model, DataList(self), verbose=verbose)
self._joint_like_obj.set_minimizer(minimizer)
return self._joint_like_obj.fit()
class XYLike(PluginPrototype):
def __init__(self, name, x, y, yerr=None, poisson_data=False, quiet=False, source_name=None):
nuisance_parameters = {}
super(XYLike, self).__init__(name, nuisance_parameters)
# Make x and y always arrays so we can handle them always in the same way
# even if they have only one element
self._x = np.array(x, ndmin=1)
self._y = np.array(y, ndmin=1)
# If there are specified errors, use those (assume Gaussian statistic)
# otherwise make sure that the user specified poisson_error = True and use
# Poisson statistic
if yerr is not None:
self._yerr = np.array(yerr, ndmin=1)
assert np.all(self._yerr > 0), "Errors cannot be negative or zero."
if not quiet:
print("Using Gaussian statistic (equivalent to chi^2) with the provided errors.")
self._is_poisson = False
self._has_errors = True
elif not poisson_data:
self._yerr = np.ones_like(self._y)
self._is_poisson = False
self._has_errors = False
if not quiet:
print("Using unweighted Gaussian (equivalent to chi^2) statistic.")
else:
if not quiet:
print("Using Poisson log-likelihood")
self._is_poisson = True
self._yerr = None
self._has_errors = True
# This will keep track of the simulated datasets we generate
self._n_simulated_datasets = 0
# This will contain the JointLikelihood object after a call to .fit()
self._joint_like_obj = None
self._likelihood_model = None
# currently not used by XYLike, but needed for subclasses
self._mask = np.ones(self._x.shape, dtype=bool)
# This is the name of the source this SED refers to (if it is a SED)
self._source_name = source_name
@classmethod
def from_function(cls, name, function, x, yerr, **kwargs):
"""
Generate an XYLike plugin from an astromodels function instance
:param name: name of plugin
:param function: astromodels function instance
:param x: where to simulate
:param yerr: y errors or None for Poisson data
:param kwargs: kwargs from xylike constructor
:return: XYLike plugin
"""
y = function(x)
xyl_gen = XYLike("generator", x, y, yerr, **kwargs)
pts = PointSource("fake", 0.0, 0.0, function)
model = Model(pts)
xyl_gen.set_model(model)
return xyl_gen.get_simulated_dataset(name)
@classmethod
def from_dataframe(cls, name, dataframe, x_column='x', y_column='y', err_column='yerr', poisson=False):
"""
Generate a XYLike instance from a Pandas.DataFrame instance
:param name: the name for the XYLike instance
:param dataframe: the input data frame
:param x_column: name of the column to be used as x (default: 'x')
:param y_column: name of the column to be used as y (default: 'y')
:param err_column: name of the column to be used as error on y (default: 'yerr')
:param poisson: if True, then the err_column is ignored and data are treated as Poisson distributed
:return: a XYLike instance
"""
x = dataframe[x_column]
y = dataframe[y_column]
if poisson is False:
yerr = dataframe[err_column]
if np.all(yerr == -99):
# This is a dataframe generate with the to_dataframe method, which uses -99 to indicate that the
# data are Poisson
return cls(name, x=x, y=y, poisson_data=True)
else:
# A dataset with errors
return cls(name, x=x, y=y, yerr=yerr)
else:
return cls(name, x=x, y=y, poisson_data=True)
@classmethod
def from_text_file(cls, name, filename):
"""
Instance the plugin starting from a text file generated with the .to_txt() method. Note that a more general
way of creating a XYLike instance from a text file is to read the file using pandas.DataFrame.from_csv, and
then use the .from_dataframe method of the XYLike plugin:
> df = pd.DataFrame.from_csv(filename, ...)
> xyl = XYLike.from_dataframe("my instance", df)
:param name: the name for the new instance
:param filename: path to the file
:return:
"""
df = pd.read_csv(filename, sep=" ")
return cls.from_dataframe(name, df)
def to_dataframe(self):
"""
Returns a pandas.DataFrame instance with the data in the 'x', 'y', and 'yerr' column. If the data are Poisson,
the yerr column will be -99 for every entry
:return: a pandas.DataFrame instance
"""
x_series = pd.Series(self.x, name='x')
y_series = pd.Series(self.y, name='y')
if self._is_poisson:
# Since DataFrame does not support metadata, there is no way to save the information that the data
# are Poisson distributed. We use instead a value of -99 for the error, to indicate that the data
# are Poisson
yerr_series = pd.Series(np.ones_like(self.x) * (-99), name='yerr')
else:
yerr_series = pd.Series(self.yerr, name='yerr')
df = pd.concat((x_series, y_series, yerr_series), axis=1)
return df
def to_txt(self, filename):
"""
Save the dataset in a text file. You can read the content back in a dataframe using:
> df = pandas.DataFrame.from_csv(filename, sep=' ')
and recreate the XYLike instance as:
> xyl = XYLike.from_dataframe(df)
:param filename: Name of the output file
:return: none
"""
df = self.to_dataframe() # type: pd.DataFrame
df.to_csv(filename, sep=" ")
def to_csv(self, *args, **kwargs):
"""
Save the data in a comma-separated-values file (CSV) file. All keywords arguments are passed to the
pandas.DataFrame.to_csv method (see the documentation from pandas for all possibilities). This gives a very
high control on the format of the output
All arguments are forwarded to pandas.DataFrame.to_csv
:return: none
"""
df = self.to_dataframe()
df.to_csv(**kwargs)
def assign_to_source(self, source_name):
"""
Assign these data to the given source (instead of to the sum of all sources, which is the default)
:param source_name: name of the source (must be contained in the likelihood model)
:return: none
"""
if self._likelihood_model is not None and source_name is not None:
assert source_name in self._likelihood_model.point_sources, "Source %s is not a point source in " \
"the likelihood model" % source_name
self._source_name = source_name
@property
def x(self):
return self._x
@property
def y(self):
return self._y
@property
def yerr(self):
return self._yerr
@property
def is_poisson(self):
return self._is_poisson
@property
def has_errors(self):
return self._has_errors
def set_model(self, likelihood_model_instance):
"""
Set the model to be used in the joint minimization. Must be a LikelihoodModel instance.
:param likelihood_model_instance: instance of Model
:type likelihood_model_instance: astromodels.Model
"""
if likelihood_model_instance is None:
return
if self._source_name is not None:
# Make sure that the source is in the model
assert self._source_name in likelihood_model_instance.point_sources, \
"This XYLike plugin refers to the source %s, " \
"but that source is not a point source in the likelihood model" % (self._source_name)
self._likelihood_model = likelihood_model_instance
def _get_total_expectation(self):
if self._source_name is None:
n_point_sources = self._likelihood_model.get_number_of_point_sources()
assert n_point_sources > 0, "You need to have at least one point source defined"
assert self._likelihood_model.get_number_of_extended_sources() == 0, "XYLike does not support extended sources"
# Make a function which will stack all point sources (XYLike do not support spatial dimension)
expectation = np.sum([source(self._x, tag=self._tag) for source in list(self._likelihood_model.point_sources.values())],
axis=0)
else:
# This XYLike dataset refers to a specific source
# Note that we checked that self._source_name is in the model when the model was set
if self._source_name in self._likelihood_model.point_sources:
expectation = self._likelihood_model.point_sources[self._source_name](self._x)
else:
raise KeyError("This XYLike plugin has been assigned to source %s, "
"which is not a point soure in the current model" % self._source_name)
return expectation
def get_log_like(self):
"""
Return the value of the log-likelihood with the current values for the
parameters
"""
expectation = self._get_total_expectation()
if self._is_poisson:
# Poisson log-likelihood
return np.sum(poisson_log_likelihood_ideal_bkg(self._y, np.zeros_like(self._y), expectation))
else:
# Chi squared
chi2_ = half_chi2(self._y, self._yerr, expectation)
assert np.all(np.isfinite(chi2_))
return np.sum(chi2_) * (-1)
def get_simulated_dataset(self, new_name=None):
assert self._has_errors, "You cannot simulate a dataset if the original dataset has no errors"
self._n_simulated_datasets += 1
# unmask the data
old_mask = copy.copy(self._mask)
self._mask = np.ones(self._x.shape, dtype=bool)
if new_name is None:
new_name = "%s_sim%i" % (self.name, self._n_simulated_datasets)
# Get total expectation from model
expectation = self._get_total_expectation()
if self._is_poisson:
new_y = np.random.poisson(expectation)
else:
new_y = np.random.normal(expectation, self._yerr)
# remask the data BEFORE creating the new plugin
self._mask = old_mask
return self._new_plugin(new_name, self._x, new_y, yerr=self._yerr)
def _new_plugin(self, name, x, y, yerr):
"""
construct a new plugin. allows for returning a new plugin
from simulated data set while customizing the constructor
further down the inheritance tree
:param name: new name
:param x: new x
:param y: new y
:param yerr: new yerr
:return: new XYLike
"""
new_xy = type(self)(name, x, y, yerr, poisson_data=self._is_poisson, quiet=True)
# apply the current mask
new_xy._mask = copy.copy(self._mask)
return new_xy
def plot(self, x_label='x', y_label='y', x_scale='linear', y_scale='linear'):
fig, sub = plt.subplots(1,1)
sub.errorbar(self.x, self.y, yerr=self.yerr, fmt='.')
sub.set_xscale(x_scale)
sub.set_yscale(y_scale)
sub.set_xlabel(x_label)
sub.set_ylabel(y_label)
if self._likelihood_model is not None:
flux = self._get_total_expectation()
sub.plot(self.x, flux, '--', label='model')
sub.legend(loc=0)
return fig
def inner_fit(self):
"""
This is used for the profile likelihood. Keeping fixed all parameters in the
LikelihoodModel, this method minimize the logLike over the remaining nuisance
parameters, i.e., the parameters belonging only to the model for this
particular detector. If there are no nuisance parameters, simply return the
logLike value.
"""
return self.get_log_like()
def get_model(self):
return self._get_total_expectation()
def fit(self, function, minimizer='minuit', verbose=False):
"""
Fit the data with the provided function (an astromodels function)
:param function: astromodels function
:param minimizer: the minimizer to use
:param verbose: print every step of the fit procedure
:return: best fit results
"""
# This is a wrapper to give an easier way to fit simple data without having to go through the definition
# of sources
pts = PointSource("source", 0.0, 0.0, function)
model = Model(pts)
self.set_model(model)
self._joint_like_obj = JointLikelihood(model, DataList(self), verbose=verbose)
self._joint_like_obj.set_minimizer(minimizer)
return self._joint_like_obj.fit()
def goodness_of_fit(self, n_iterations=1000, continue_of_failure=False):
"""
Returns the goodness of fit of the performed fit
:param n_iterations: number of Monte Carlo simulations to generate
:param continue_of_failure: whether to continue or not if a fit fails (default: False)
:return: tuple (goodness of fit, frame with all results, frame with all likelihood values)
"""
g = GoodnessOfFit(self._joint_like_obj)
return g.by_mc(n_iterations, continue_of_failure)
def get_number_of_data_points(self):
"""
returns the number of active data points
:return:
"""
# the sum of the mask should be the number of data points in use
return self._mask.sum()
def worker(self, interval):
# Get the dataset for this interval
this_data = self._data_getter(interval) # type: DataList
# Get the model for this interval
this_models = self._model_getter(interval)
# Apply preprocessor (if any)
if self._preprocessor is not None:
self._preprocessor(this_models, this_data)
n_models = len(this_models)
# Fit all models and collect the results
parameters_frames = []
like_frames = []
analysis_results = []
for this_model in this_models:
# Prepare a joint likelihood and fit it
with warnings.catch_warnings():
warnings.simplefilter("ignore", RuntimeWarning)
jl = JointLikelihood(this_model, this_data)
this_parameter_frame, this_like_frame = self._fitter(jl)
# Append results
parameters_frames.append(this_parameter_frame)
like_frames.append(this_like_frame)
analysis_results.append(jl.results)
# Now merge the results in one data frame for the parameters and one for the likelihood
# values
if n_models > 1:
# Prepare the keys so that the first model will be indexed with model_0, the second model_1 and so on
keys = ["model_%i" % x for x in range(n_models)]
# Concatenate all results in one frame for parameters and one for likelihood
frame_with_parameters = pd.concat(parameters_frames, keys=keys)
frame_with_like = pd.concat(like_frames, keys=keys)
else:
frame_with_parameters = parameters_frames[0]
frame_with_like = like_frames[0]
return frame_with_parameters, frame_with_like, analysis_results
def test_energy_time_fit():
# Let's generate our dataset of 4 spectra with a normalization that follows
# a powerlaw in time
def generate_one(K):
# Let's generate some data with y = Powerlaw(x)
gen_function = Powerlaw()
gen_function.K = K
# Generate a dataset using the power law, and a
# constant 30% error
x = np.logspace(0, 2, 50)
xyl_generator = XYLike.from_function("sim_data",
function=gen_function,
x=x,
yerr=0.3 * gen_function(x))
y = xyl_generator.y
y_err = xyl_generator.yerr
# xyl = XYLike("data", x, y, y_err)
# xyl.plot(x_scale='log', y_scale='log')
return x, y, y_err
time_tags = np.array([1.0, 2.0, 5.0, 10.0])
# This is the power law that defines the normalization as a function of time
normalizations = 0.23 * time_tags**(-1.2)
datasets = list(map(generate_one, normalizations))
# Now set up the fit and fit it
time = IndependentVariable("time", 1.0, u.s)
plugins = []
for i, dataset in enumerate(datasets):
x, y, y_err = dataset
xyl = XYLike("data%i" % i, x, y, y_err)
xyl.tag = (time, time_tags[i])
assert xyl.tag == (time, time_tags[i], None)
plugins.append(xyl)
data = DataList(*plugins)
spectrum = Powerlaw()
spectrum.K.bounds = (0.01, 1000.0)
src = PointSource("test", 0.0, 0.0, spectrum)
model = Model(src)
model.add_independent_variable(time)
time_po = Powerlaw()
time_po.K.bounds = (0.01, 1000)
time_po.K.value = 2.0
time_po.index = -1.5
model.link(spectrum.K, time, time_po)
jl = JointLikelihood(model, data)
jl.set_minimizer("minuit")
best_fit_parameters, likelihood_values = jl.fit()
# Make sure we are within 10% of the expected result
assert np.allclose(
best_fit_parameters["value"].values,
[0.25496115, -1.2282951, -2.01508341],
rtol=0.1,
)
class XYLike(PluginPrototype):
def __init__(self, name, x, y, yerr=None, poisson_data=False, quiet=False, source_name=None):
nuisance_parameters = {}
super(XYLike, self).__init__(name, nuisance_parameters)
# Make x and y always arrays so we can handle them always in the same way
# even if they have only one element
self._x = np.array(x, ndmin=1)
self._y = np.array(y, ndmin=1)
# If there are specified errors, use those (assume Gaussian statistic)
# otherwise make sure that the user specified poisson_error = True and use
# Poisson statistic
if yerr is not None:
self._yerr = np.array(yerr, ndmin=1)
assert np.all(self._yerr > 0), "Errors cannot be negative or zero."
if not quiet:
print("Using Gaussian statistic (equivalent to chi^2) with the provided errors.")
self._is_poisson = False
self._has_errors = True
elif not poisson_data:
self._yerr = np.ones_like(self._y)
self._is_poisson = False
self._has_errors = False
if not quiet:
print("Using unweighted Gaussian (equivalent to chi^2) statistic.")
else:
if not quiet:
print("Using Poisson log-likelihood")
self._is_poisson = True
self._yerr = None
self._has_errors = True
# This will keep track of the simulated datasets we generate
self._n_simulated_datasets = 0
# This will contain the JointLikelihood object after a call to .fit()
self._joint_like_obj = None
self._likelihood_model = None
# currently not used by XYLike, but needed for subclasses
self._mask = np.ones(self._x.shape, dtype=bool)
# This is the name of the source this SED refers to (if it is a SED)
self._source_name = source_name
@classmethod
def from_function(cls, name, function, x, yerr, **kwargs):
"""
Generate an XYLike plugin from an astromodels function instance
:param name: name of plugin
:param function: astromodels function instance
:param x: where to simulate
:param yerr: y errors or None for Poisson data
:param kwargs: kwargs from xylike constructor
:return: XYLike plugin
"""
y = function(x)
xyl_gen = XYLike("generator", x, y, yerr, **kwargs)
pts = PointSource("fake", 0.0, 0.0, function)
model = Model(pts)
xyl_gen.set_model(model)
return xyl_gen.get_simulated_dataset(name)
@classmethod
def from_dataframe(cls, name, dataframe, x_column='x', y_column='y', err_column='yerr', poisson=False):
"""
Generate a XYLike instance from a Pandas.DataFrame instance
:param name: the name for the XYLike instance
:param dataframe: the input data frame
:param x_column: name of the column to be used as x (default: 'x')
:param y_column: name of the column to be used as y (default: 'y')
:param err_column: name of the column to be used as error on y (default: 'yerr')
:param poisson: if True, then the err_column is ignored and data are treated as Poisson distributed
:return: a XYLike instance
"""
x = dataframe[x_column]
y = dataframe[y_column]
if poisson is False:
yerr = dataframe[err_column]
if np.all(yerr == -99):
# This is a dataframe generate with the to_dataframe method, which uses -99 to indicate that the
# data are Poisson
return cls(name, x=x, y=y, poisson_data=True)
else:
# A dataset with errors
return cls(name, x=x, y=y, yerr=yerr)
else:
return cls(name, x=x, y=y, poisson_data=True)
@classmethod
def from_text_file(cls, name, filename):
"""
Instance the plugin starting from a text file generated with the .to_txt() method. Note that a more general
way of creating a XYLike instance from a text file is to read the file using pandas.DataFrame.from_csv, and
then use the .from_dataframe method of the XYLike plugin:
> df = pd.DataFrame.from_csv(filename, ...)
> xyl = XYLike.from_dataframe("my instance", df)
:param name: the name for the new instance
:param filename: path to the file
:return:
"""
df = pd.DataFrame.from_csv(filename, sep=" ")
return cls.from_dataframe(name, df)
def to_dataframe(self):
"""
Returns a pandas.DataFrame instance with the data in the 'x', 'y', and 'yerr' column. If the data are Poisson,
the yerr column will be -99 for every entry
:return: a pandas.DataFrame instance
"""
x_series = pd.Series.from_array(self.x, name='x')
y_series = pd.Series.from_array(self.y, name='y')
if self._is_poisson:
# Since DataFrame does not support metadata, there is no way to save the information that the data
# are Poisson distributed. We use instead a value of -99 for the error, to indicate that the data
# are Poisson
yerr_series = pd.Series.from_array(np.ones_like(self.x) * (-99), name='yerr')
else:
yerr_series = pd.Series.from_array(self.yerr, name='yerr')
df = pd.concat((x_series, y_series, yerr_series), axis=1)
return df
def to_txt(self, filename):
"""
Save the dataset in a text file. You can read the content back in a dataframe using:
> df = pandas.DataFrame.from_csv(filename, sep=' ')
and recreate the XYLike instance as:
> xyl = XYLike.from_dataframe(df)
:param filename: Name of the output file
:return: none
"""
df = self.to_dataframe() # type: pd.DataFrame
df.to_csv(filename, sep=" ")
def to_csv(self, *args, **kwargs):
"""
Save the data in a comma-separated-values file (CSV) file. All keywords arguments are passed to the
pandas.DataFrame.to_csv method (see the documentation from pandas for all possibilities). This gives a very
high control on the format of the output
All arguments are forwarded to pandas.DataFrame.to_csv
:return: none
"""
df = self.to_dataframe()
df.to_csv(**kwargs)
def assign_to_source(self, source_name):
"""
Assign these data to the given source (instead of to the sum of all sources, which is the default)
:param source_name: name of the source (must be contained in the likelihood model)
:return: none
"""
if self._likelihood_model is not None and source_name is not None:
assert source_name in self._likelihood_model.point_sources, "Source %s is not a point source in " \
"the likelihood model" % source_name
self._source_name = source_name
@property
def x(self):
return self._x
@property
def y(self):
return self._y
@property
def yerr(self):
return self._yerr
@property
def is_poisson(self):
return self._is_poisson
@property
def has_errors(self):
return self._has_errors
def set_model(self, likelihood_model_instance):
"""
Set the model to be used in the joint minimization. Must be a LikelihoodModel instance.
:param likelihood_model_instance: instance of Model
:type likelihood_model_instance: astromodels.Model
"""
if likelihood_model_instance is None:
return
if self._source_name is not None:
# Make sure that the source is in the model
assert self._source_name in likelihood_model_instance.point_sources, \
"This XYLike plugin refers to the source %s, " \
"but that source is not a point source in the likelihood model" % (self._source_name)
self._likelihood_model = likelihood_model_instance
def _get_total_expectation(self):
if self._source_name is None:
n_point_sources = self._likelihood_model.get_number_of_point_sources()
assert n_point_sources > 0, "You need to have at least one point source defined"
assert self._likelihood_model.get_number_of_extended_sources() == 0, "XYLike does not support extended sources"
# Make a function which will stack all point sources (XYLike do not support spatial dimension)
expectation = np.sum(map(lambda source: source(self._x, tag=self._tag),
self._likelihood_model.point_sources.values()),
axis=0)
else:
# This XYLike dataset refers to a specific source
# Note that we checked that self._source_name is in the model when the model was set
if self._source_name in self._likelihood_model.point_sources:
expectation = self._likelihood_model.point_sources[self._source_name](self._x)
else:
raise KeyError("This XYLike plugin has been assigned to source %s, "
"which is not a point soure in the current model" % self._source_name)
return expectation
def get_log_like(self):
"""
Return the value of the log-likelihood with the current values for the
parameters
"""
expectation = self._get_total_expectation()
if self._is_poisson:
# Poisson log-likelihood
return np.sum(poisson_log_likelihood_ideal_bkg(self._y, np.zeros_like(self._y), expectation))
else:
# Chi squared
chi2_ = half_chi2(self._y, self._yerr, expectation)
assert np.all(np.isfinite(chi2_))
return np.sum(chi2_) * (-1)
def get_simulated_dataset(self, new_name=None):
assert self._has_errors, "You cannot simulate a dataset if the original dataset has no errors"
self._n_simulated_datasets += 1
# unmask the data
old_mask = copy.copy(self._mask)
self._mask = np.ones(self._x.shape, dtype=bool)
if new_name is None:
new_name = "%s_sim%i" % (self.name, self._n_simulated_datasets)
# Get total expectation from model
expectation = self._get_total_expectation()
if self._is_poisson:
new_y = np.random.poisson(expectation)
else:
new_y = np.random.normal(expectation, self._yerr)
# remask the data BEFORE creating the new plugin
self._mask = old_mask
return self._new_plugin(new_name, self._x, new_y, yerr=self._yerr)
def _new_plugin(self, name, x, y, yerr):
"""
construct a new plugin. allows for returning a new plugin
from simulated data set while customizing the constructor
further down the inheritance tree
:param name: new name
:param x: new x
:param y: new y
:param yerr: new yerr
:return: new XYLike
"""
new_xy = type(self)(name, x, y, yerr, poisson_data=self._is_poisson, quiet=True)
# apply the current mask
new_xy._mask = copy.copy(self._mask)
return new_xy
def plot(self, x_label='x', y_label='y', x_scale='linear', y_scale='linear'):
fig, sub = plt.subplots(1,1)
sub.errorbar(self.x, self.y, yerr=self.yerr, fmt='.')
sub.set_xscale(x_scale)
sub.set_yscale(y_scale)
sub.set_xlabel(x_label)
sub.set_ylabel(y_label)
if self._likelihood_model is not None:
flux = self._get_total_expectation()
sub.plot(self.x, flux, '--', label='model')
sub.legend(loc=0)
return fig
def inner_fit(self):
"""
This is used for the profile likelihood. Keeping fixed all parameters in the
LikelihoodModel, this method minimize the logLike over the remaining nuisance
parameters, i.e., the parameters belonging only to the model for this
particular detector. If there are no nuisance parameters, simply return the
logLike value.
"""
return self.get_log_like()
def get_model(self):
return self._get_total_expectation()
def fit(self, function, minimizer='minuit', verbose=False):
"""
Fit the data with the provided function (an astromodels function)
:param function: astromodels function
:param minimizer: the minimizer to use
:param verbose: print every step of the fit procedure
:return: best fit results
"""
# This is a wrapper to give an easier way to fit simple data without having to go through the definition
# of sources
pts = PointSource("source", 0.0, 0.0, function)
model = Model(pts)
self.set_model(model)
self._joint_like_obj = JointLikelihood(model, DataList(self), verbose=verbose)
self._joint_like_obj.set_minimizer(minimizer)
return self._joint_like_obj.fit()
def goodness_of_fit(self, n_iterations=1000, continue_of_failure=False):
"""
Returns the goodness of fit of the performed fit
:param n_iterations: number of Monte Carlo simulations to generate
:param continue_of_failure: whether to continue or not if a fit fails (default: False)
:return: tuple (goodness of fit, frame with all results, frame with all likelihood values)
"""
g = GoodnessOfFit(self._joint_like_obj)
return g.by_mc(n_iterations, continue_of_failure)
def get_number_of_data_points(self):
"""
returns the number of active data points
:return:
"""
# the sum of the mask should be the number of data points in use
return self._mask.sum()
def unbinned_polyfit(events: Iterable[float], grade: int, t_start: float, t_stop: float, exposure: float, bayes: bool) -> Tuple[Polynomial, float]:
"""
function to fit a polynomial to unbinned event data.
not a member to allow parallel computation
:param events: the events to fit
:param grade: the polynomical order or grade
:param t_start: the start time to fit over
:param t_stop: the end time to fit over
:param expousure: the exposure of the interval
:param bayes: to do a bayesian fit or not
"""
log.debug(f"starting unbinned_polyfit with grade {grade}")
log.debug(f"have {len(events)} events with {exposure} exposure")
# create 3ML plugins and fit them with 3ML!
# should eventuallly allow better config
# select the model based on the grade
if threeML_config.time_series.default_fit_method is not None:
bayes = threeML_config.time_series.default_fit_method
log.debug("using a default poly fit method")
if len(events) == 0:
log.debug("no events! returning zero")
return Polynomial([0] * (grade + 1)), 0
shape = _grade_model_lookup[grade]()
with silence_console_log():
ps = PointSource("dummy", 0, 0, spectral_shape=shape)
model = Model(ps)
observation = EventObservation(events, exposure, t_start, t_stop)
xy = UnbinnedPoissonLike("series", observation=observation)
if not bayes:
# make sure the model is positive
for i, (k, v) in enumerate(model.free_parameters.items()):
if i == 0:
v.bounds = (0, None)
v.value = 10
else:
v.value = 0.0
# we actually use a line here
# because a constant is returns a
# single number
if grade == 0:
shape.b = 0
shape.b.fix = True
jl: JointLikelihood = JointLikelihood(model, DataList(xy))
grid_minimizer = GlobalMinimization("grid")
local_minimizer = LocalMinimization("minuit")
my_grid = {
model.dummy.spectrum.main.shape.a: np.logspace(0, 3, 10)}
grid_minimizer.setup(
second_minimization=local_minimizer, grid=my_grid)
jl.set_minimizer(grid_minimizer)
# if the fit falis, retry and then just accept
try:
jl.fit(quiet=True)
except(FitFailed, BadCovariance, AllFitFailed, CannotComputeCovariance):
try:
jl.fit(quiet=True)
except(FitFailed, BadCovariance, AllFitFailed, CannotComputeCovariance):
log.debug("all MLE fits failed, returning zero")
return Polynomial([0]*(grade + 1)), 0
coeff = [v.value for _, v in model.free_parameters.items()]
log.debug(f"got coeff: {coeff}")
final_polynomial = Polynomial(coeff)
final_polynomial.set_covariace_matrix(jl.results.covariance_matrix)
min_log_likelihood = xy.get_log_like()
else:
# set smart priors
for i, (k, v) in enumerate(model.free_parameters.items()):
if i == 0:
v.bounds = (0, None)
v.prior = Log_normal(mu=np.log(5), sigma=np.log(5))
v.value = 1
else:
v.prior = Gaussian(mu=0, sigma=.5)
v.value = 0.1
# we actually use a line here
# because a constant is returns a
# single number
if grade == 0:
shape.b = 0
shape.b.fix = True
ba: BayesianAnalysis = BayesianAnalysis(model, DataList(xy))
ba.set_sampler("emcee")
ba.sampler.setup(n_iterations=500, n_burn_in=200, n_walkers=20)
ba.sample(quiet=True)
ba.restore_median_fit()
coeff = [v.value for _, v in model.free_parameters.items()]
log.debug(f"got coeff: {coeff}")
final_polynomial = Polynomial(coeff)
final_polynomial.set_covariace_matrix(
ba.results.estimate_covariance_matrix())
min_log_likelihood = xy.get_log_like()
log.debug(f"-min loglike: {-min_log_likelihood}")
return final_polynomial, -min_log_likelihood
def polyfit(x: Iterable[float], y: Iterable[float], grade: int, exposure: Iterable[float], bayes: bool = False) -> Tuple[Polynomial, float]:
"""
function to fit a polynomial to data.
not a member to allow parallel computation
:param x: the x coord of the data
:param y: teh y coord of the data
:param grade: the polynomical order or grade
:param expousure: the exposure of the interval
:param bayes: to do a bayesian fit or not
"""
# Check that we have enough counts to perform the fit, otherwise
# return a "zero polynomial"
log.debug(f"starting polyfit with grade {grade} ")
if threeML_config.time_series.default_fit_method is not None:
bayes = threeML_config.time_series.default_fit_method
log.debug("using a default poly fit method")
nan_mask = np.isnan(y)
y = y[~nan_mask]
x = x[~nan_mask]
exposure = exposure[~nan_mask]
non_zero_mask = y > 0
n_non_zero = non_zero_mask.sum()
if n_non_zero == 0:
log.debug("no counts, return 0")
# No data, nothing to do!
return Polynomial([0.0]*(grade+1)), 0.0
# create 3ML plugins and fit them with 3ML!
# should eventuallly allow better config
# seelct the model based on the grade
shape = _grade_model_lookup[grade]()
ps = PointSource("_dummy", 0, 0, spectral_shape=shape)
model = Model(ps)
avg = np.mean(y/exposure)
log.debug(f"starting polyfit with avg norm {avg}")
with silence_console_log():
xy = XYLike("series", x=x, y=y, exposure=exposure,
poisson_data=True, quiet=True)
if not bayes:
# make sure the model is positive
for i, (k, v) in enumerate(model.free_parameters.items()):
if i == 0:
v.bounds = (0, None)
v.value = avg
else:
v.value = 0.0
# we actually use a line here
# because a constant is returns a
# single number
if grade == 0:
shape.b = 0
shape.b.fix = True
jl: JointLikelihood = JointLikelihood(model, DataList(xy))
jl.set_minimizer("minuit")
# if the fit falis, retry and then just accept
try:
jl.fit(quiet=True)
except(FitFailed, BadCovariance, AllFitFailed, CannotComputeCovariance):
log.debug("1st fit failed")
try:
jl.fit(quiet=True)
except(FitFailed, BadCovariance, AllFitFailed, CannotComputeCovariance):
log.debug("all MLE fits failed")
pass
coeff = [v.value for _, v in model.free_parameters.items()]
log.debug(f"got coeff: {coeff}")
final_polynomial = Polynomial(coeff)
try:
final_polynomial.set_covariace_matrix(
jl.results.covariance_matrix)
except:
log.exception(f"Fit failed in channel")
raise FitFailed()
min_log_likelihood = xy.get_log_like()
else:
# set smart priors
for i, (k, v) in enumerate(model.free_parameters.items()):
if i == 0:
v.bounds = (0, None)
v.prior = Log_normal(
mu=np.log(avg), sigma=np.max([np.log(avg/2), 1]))
v.value = 1
else:
v.prior = Gaussian(mu=0, sigma=2)
v.value = 1e-2
# we actually use a line here
# because a constant is returns a
# single number
if grade == 0:
shape.b = 0
shape.b.fix = True
ba: BayesianAnalysis = BayesianAnalysis(model, DataList(xy))
ba.set_sampler("emcee")
ba.sampler.setup(n_iterations=500, n_burn_in=200, n_walkers=20)
ba.sample(quiet=True)
ba.restore_median_fit()
coeff = [v.value for _, v in model.free_parameters.items()]
log.debug(f"got coeff: {coeff}")
final_polynomial = Polynomial(coeff)
final_polynomial.set_covariace_matrix(
ba.results.estimate_covariance_matrix())
min_log_likelihood = xy.get_log_like()
log.debug(f"-min loglike: {-min_log_likelihood}")
return final_polynomial, -min_log_likelihood
def test_ubinned_poisson_full(event_observation_contiguous, event_observation_split):
s = Line()
ps = PointSource("s", 0, 0, spectral_shape=s)
s.a.bounds = (0, None)
s.a.value = .1
s.b.value = .1
s.a.prior = Log_normal(mu=np.log(10), sigma=1)
s.b.prior = Gaussian(mu=0, sigma=1)
m = Model(ps)
######
######
######
ub1 = UnbinnedPoissonLike("test", observation=event_observation_contiguous)
jl = JointLikelihood(m, DataList(ub1))
jl.fit(quiet=True)
np.testing.assert_allclose([s.a.value, s.b.value], [6.11, 1.45], rtol=.5)
ba = BayesianAnalysis(m, DataList(ub1))
ba.set_sampler("emcee")
ba.sampler.setup(n_burn_in=100, n_walkers=20, n_iterations=500)
ba.sample(quiet=True)
ba.restore_median_fit()
np.testing.assert_allclose([s.a.value, s.b.value], [6.11, 1.45], rtol=.5)
######
######
######
ub2 = UnbinnedPoissonLike("test", observation=event_observation_split)
jl = JointLikelihood(m, DataList(ub2))
jl.fit(quiet=True)
np.testing.assert_allclose([s.a.value, s.b.value], [2., .2], rtol=.5)
ba = BayesianAnalysis(m, DataList(ub2))
ba.set_sampler("emcee")
ba.sampler.setup(n_burn_in=100, n_walkers=20, n_iterations=500)
ba.sample(quiet=True)
ba.restore_median_fit()
np.testing.assert_allclose([s.a.value, s.b.value], [2., .2], rtol=.5)
def test_energy_time_fit():
# Let's generate our dataset of 4 spectra with a normalization that follows
# a powerlaw in time
def generate_one(K):
# Let's generate some data with y = Powerlaw(x)
gen_function = Powerlaw()
gen_function.K = K
# Generate a dataset using the power law, and a
# constant 30% error
x = np.logspace(0, 2, 50)
xyl_generator = XYLike.from_function("sim_data", function=gen_function,
x=x,
yerr=0.3 * gen_function(x))
y = xyl_generator.y
y_err = xyl_generator.yerr
# xyl = XYLike("data", x, y, y_err)
# xyl.plot(x_scale='log', y_scale='log')
return x, y, y_err
time_tags = np.array([1.0, 2.0, 5.0, 10.0])
# This is the power law that defines the normalization as a function of time
normalizations = 0.23 * time_tags ** (-1.2)
datasets = map(generate_one, normalizations)
# Now set up the fit and fit it
time = IndependentVariable("time", 1.0, u.s)
plugins = []
for i, dataset in enumerate(datasets):
x, y, y_err = dataset
xyl = XYLike("data%i" % i, x, y, y_err)
xyl.tag = (time, time_tags[i])
assert xyl.tag == (time, time_tags[i], None)
plugins.append(xyl)
data = DataList(*plugins)
spectrum = Powerlaw()
spectrum.K.bounds = (0.01, 1000.0)
src = PointSource("test", 0.0, 0.0, spectrum)
model = Model(src)
model.add_independent_variable(time)
time_po = Powerlaw()
time_po.K.bounds = (0.01, 1000)
time_po.K.value = 2.0
time_po.index = -1.5
model.link(spectrum.K, time, time_po)
jl = JointLikelihood(model, data)
jl.set_minimizer("minuit")
best_fit_parameters, likelihood_values = jl.fit()
# Make sure we are within 10% of the expected result
assert np.allclose(best_fit_parameters['value'].values, [0.25496115, -1.2282951 , -2.01508341], rtol=0.1)
