|
1
|
|
|
""" |
|
2
|
|
|
Main classes to define High-level Analysis Config and the Analysis Steps. |
|
3
|
|
|
""" |
|
4
|
|
|
|
|
5
|
|
|
from enum import Enum |
|
6
|
|
|
|
|
7
|
|
|
from astropy import units as u |
|
8
|
|
|
from gammapy.datasets import Datasets, FluxPointsDataset |
|
9
|
|
|
from gammapy.estimators import FluxPointsEstimator |
|
10
|
|
|
from gammapy.modeling import Fit |
|
11
|
|
|
from gammapy.utils.metadata import CreatorMetaData |
|
12
|
|
|
|
|
13
|
|
|
from asgardpy.analysis.step_base import AnalysisStepBase |
|
14
|
|
|
from asgardpy.base import BaseConfig, EnergyRangeConfig |
|
15
|
|
|
from asgardpy.version import __public_version__ |
|
16
|
|
|
|
|
17
|
|
|
__all__ = [ |
|
18
|
|
|
"FitAnalysisStep", |
|
19
|
|
|
"FitConfig", |
|
20
|
|
|
"FluxPointsAnalysisStep", |
|
21
|
|
|
"FluxPointsConfig", |
|
22
|
|
|
] |
|
23
|
|
|
|
|
24
|
|
|
|
|
25
|
|
|
# Defining various components of High-level Analysis Config |
|
26
|
|
|
class BackendEnum(str, Enum): |
|
27
|
|
|
"""Config section for a list Fitting backend methods.""" |
|
28
|
|
|
|
|
29
|
|
|
minuit = "minuit" |
|
30
|
|
|
scipy = "scipy" |
|
31
|
|
|
|
|
32
|
|
|
|
|
33
|
|
|
class FitConfig(BaseConfig): |
|
34
|
|
|
"""Config section for parameters to use for Fit function.""" |
|
35
|
|
|
|
|
36
|
|
|
fit_range: EnergyRangeConfig = EnergyRangeConfig() |
|
37
|
|
|
backend: BackendEnum = BackendEnum.minuit |
|
38
|
|
|
optimize_opts: dict = {} |
|
39
|
|
|
covariance_opts: dict = {} |
|
40
|
|
|
confidence_opts: dict = {} |
|
41
|
|
|
store_trace: bool = True |
|
42
|
|
|
|
|
43
|
|
|
|
|
44
|
|
|
class FluxPointsConfig(BaseConfig): |
|
45
|
|
|
"""Config section for parameters to use for FluxPointsEstimator function.""" |
|
46
|
|
|
|
|
47
|
|
|
parameters: dict = {"selection_optional": "all"} |
|
48
|
|
|
reoptimize: bool = False |
|
49
|
|
|
|
|
50
|
|
|
|
|
51
|
|
|
# The main Analysis Steps |
|
52
|
|
|
class FitAnalysisStep(AnalysisStepBase): |
|
53
|
|
|
""" |
|
54
|
|
|
Using the Fitting parameters as defined in the Config, with the given |
|
55
|
|
|
datasets perform the fit of the models to the updated list of datasets. |
|
56
|
|
|
""" |
|
57
|
|
|
|
|
58
|
|
|
tag = "fit" |
|
59
|
|
|
|
|
60
|
|
|
def _run(self): |
|
61
|
|
|
self.fit_params = self.config.fit_params |
|
62
|
|
|
|
|
63
|
|
|
self._setup_fit() |
|
64
|
|
|
final_dataset = self._set_datasets() |
|
65
|
|
|
self.fit_result = self.fit.run(datasets=final_dataset) |
|
66
|
|
|
|
|
67
|
|
|
self.log.info(self.fit_result) |
|
68
|
|
|
|
|
69
|
|
|
def _setup_fit(self): |
|
70
|
|
|
""" |
|
71
|
|
|
Setup the Gammapy Fit function with all the provided parameters from |
|
72
|
|
|
the config. |
|
73
|
|
|
""" |
|
74
|
|
|
self.fit = Fit( |
|
75
|
|
|
backend=self.fit_params.backend, |
|
76
|
|
|
optimize_opts=self.fit_params.optimize_opts, |
|
77
|
|
|
covariance_opts=self.fit_params.covariance_opts, |
|
78
|
|
|
confidence_opts=self.fit_params.confidence_opts, |
|
79
|
|
|
store_trace=self.fit_params.store_trace, |
|
80
|
|
|
) |
|
81
|
|
|
|
|
82
|
|
|
def _set_datasets(self): |
|
83
|
|
|
""" |
|
84
|
|
|
Prepare each dataset for running the Fit function, by setting the |
|
85
|
|
|
energy range. |
|
86
|
|
|
""" |
|
87
|
|
|
en_min = u.Quantity(self.fit_params.fit_range.min) |
|
88
|
|
|
en_max = u.Quantity(self.fit_params.fit_range.max) |
|
89
|
|
|
|
|
90
|
|
|
final_dataset = Datasets() |
|
91
|
|
|
for data in self.datasets: |
|
92
|
|
|
if not isinstance(data, FluxPointsDataset): |
|
93
|
|
|
geom = data.counts.geom |
|
94
|
|
|
data.mask_fit = geom.energy_mask(en_min, en_max) |
|
95
|
|
|
final_dataset.append(data) |
|
96
|
|
|
|
|
97
|
|
|
return final_dataset |
|
98
|
|
|
|
|
99
|
|
|
|
|
100
|
|
|
class FluxPointsAnalysisStep(AnalysisStepBase): |
|
101
|
|
|
""" |
|
102
|
|
|
Using the Flux Points Estimator parameters in the config, and the given |
|
103
|
|
|
datasets and instrument_spectral_info perform the Flux Points Estimation |
|
104
|
|
|
and store the result in a list of flux points for each dataset. |
|
105
|
|
|
""" |
|
106
|
|
|
|
|
107
|
|
|
tag = "flux-points" |
|
108
|
|
|
|
|
109
|
|
|
def _run(self): |
|
110
|
|
|
self.flux_points = [] |
|
111
|
|
|
datasets, energy_edges = self._sort_datasets_info() |
|
112
|
|
|
|
|
113
|
|
|
for dataset, energy_edge in zip(datasets, energy_edges, strict=True): |
|
114
|
|
|
self._set_fpe(energy_edge) |
|
115
|
|
|
flux_points = self.fpe.run(datasets=dataset) |
|
116
|
|
|
flux_points.name = dataset.names |
|
117
|
|
|
|
|
118
|
|
|
flux_points.meta["creation"] = CreatorMetaData() |
|
119
|
|
|
flux_points.meta["creation"].creator += f", Asgardpy {__public_version__}" |
|
120
|
|
|
flux_points.meta["optional"] = { |
|
121
|
|
|
"instrument": flux_points.name, |
|
122
|
|
|
} |
|
123
|
|
|
|
|
124
|
|
|
self.flux_points.append(flux_points) |
|
125
|
|
|
|
|
126
|
|
|
def _set_fpe(self, energy_bin_edges): |
|
127
|
|
|
""" |
|
128
|
|
|
Setup the Gammapy FluxPointsEstimator function with all the |
|
129
|
|
|
provided parameters. |
|
130
|
|
|
""" |
|
131
|
|
|
fpe_settings = self.config.flux_points_params.parameters |
|
132
|
|
|
|
|
133
|
|
|
self.fpe = FluxPointsEstimator( |
|
134
|
|
|
energy_edges=energy_bin_edges, |
|
135
|
|
|
source=self.config.target.source_name, |
|
136
|
|
|
n_jobs=self.config.general.n_jobs, |
|
137
|
|
|
parallel_backend=self.config.general.parallel_backend, |
|
138
|
|
|
reoptimize=self.config.flux_points_params.reoptimize, |
|
139
|
|
|
**fpe_settings, |
|
140
|
|
|
) |
|
141
|
|
|
|
|
142
|
|
|
def _sort_datasets_info(self): |
|
143
|
|
|
""" |
|
144
|
|
|
The given list of datasets may contain sub-instrument level datasets. |
|
145
|
|
|
With the help of the dict information for instrument specific name and |
|
146
|
|
|
spectral energy edges, this function, sorts the datasets and returns |
|
147
|
|
|
them to be passed to the Flux Points Estimator function. |
|
148
|
|
|
|
|
149
|
|
|
Returns |
|
150
|
|
|
------- |
|
151
|
|
|
sorted_datasets: List of Datasets object. |
|
152
|
|
|
sorted_energy_edges: List of energy edges for flux points estimation |
|
153
|
|
|
for respective instruments' datasets |
|
154
|
|
|
""" |
|
155
|
|
|
dataset_name_list = self.datasets.names |
|
156
|
|
|
sorted_datasets = [] |
|
157
|
|
|
sorted_energy_edges = [] |
|
158
|
|
|
|
|
159
|
|
|
for i, name in enumerate(self.instrument_spectral_info["name"]): |
|
160
|
|
|
dataset_list = [] |
|
161
|
|
|
for j, dataset_names in enumerate(dataset_name_list): |
|
162
|
|
|
if name in dataset_names: |
|
163
|
|
|
dataset_list.append(self.datasets[j]) |
|
164
|
|
|
if len(dataset_list) != 0: |
|
165
|
|
|
sorted_energy_edges.append(self.instrument_spectral_info["spectral_energy_ranges"][i]) |
|
166
|
|
|
dataset_list = Datasets(dataset_list) |
|
167
|
|
|
sorted_datasets.append(dataset_list) |
|
168
|
|
|
|
|
169
|
|
|
return sorted_datasets, sorted_energy_edges |
|
170
|
|
|
|
chaimain /
asgardpy
