ObsData

ObsData#

class jetset.data_loader.ObsData(cosmo=None, data_table=None, dupl_filter=False, data_set_filter=None, UL_filtering=False, UL_value=None, UL_CL=0.95, **keywords)[source]#

Bases: object

Legacy-compatible observational SED container for fitting workflows.

Notes

Wraps an input table, applies optional filtering (datasets, duplicates, upper limits), computes derived log-space columns, and exposes arrays used by minimization and plotting utilities.

Attributes Summary

`gammapy_table`	Return data as a Gammapy-compatible flux-points table.
`metadata`	Return metadata dictionary excluding internal bookkeeping keys.
`table`	Return the working observational table.

Methods Summary

`add_systematics`(syst[, nu_range, dataset])	Add fractional systematic uncertainty in quadrature.
`filter_UL`([val])	Remove rows considered upper limits by error threshold.
`filter_data_set`(filters[, exclude, silent])	Filter rows by dataset label.
`filter_freq`([nu_min, nu_max, exclude])	Filter rows by frequency interval.
`filter_time`([T_min, T_max, exclude])	Filter rows by time interval.
`find_time_span`([dataset, silent, get_values])	Compute time-span summary for selected rows.
`get_data_points`([log_log, skip_UL, frame, ...])	Return data arrays ready for plotting or fitting.
`get_data_sets`()	Return unique dataset labels present in the table.
`get_gammapy_table`()	Build a Gammapy flux-points table from the current data.
`get_time_span`([dataset])	Return time-span summary for one dataset or all data.
`group_data`([N_bin, bin_width, ...])	Rebin data in logarithmic-frequency bins.
`lin_to_log`([val, err])	Wrapper to convert linear values/errors to log10 space.
`load`(file_name)	Load a serialized `ObsData` instance from disk.
`log_to_lin`([log_val, log_err])	Wrapper to convert log10 values/errors back to linear space.
`plot_sed`([plot_obj, frame, color, fmt, ms, ...])	Plot observational SED points.
`plot_time_spans`([save_as])	Plot time coverage per dataset on a single figure.
`remove_dupl_entries`(data)	remove duplicate entries
`reset_data`()	Reset working data to the original loaded dataset.
`save`(file_name)	Serialize this `ObsData` instance to disk.
`set_UL`([val])	Flag upper limits from an error-threshold criterion.
`set_error`(error_value[, nu_range, dataset, ...])	Set relative statistical errors on selected rows.
`set_fake_error`(val)	Set fallback relative errors used in fits and plotting.
`set_zero_error`([val, replace_zero])	Mark and optionally replace non-positive uncertainty values.
`show_data_sets`()	Print currently available dataset labels.
`show_time_span`([dataset])	Print time-span summary for one dataset or all data.

Attributes Documentation

gammapy_table#: Return data as a Gammapy-compatible flux-points table.

metadata#

Return metadata dictionary excluding internal bookkeeping keys.

Parameters:: skip (list, optional) – Keys to exclude from the returned dictionary.
Returns:: Metadata key/value pairs currently set on the object.
Return type:: dict

table#: Return the working observational table.

Methods Documentation

add_systematics(syst, nu_range=None, dataset=None)[source]#

Add fractional systematic uncertainty in quadrature.

Parameters:

syst (float) – Relative systematic term (for example 0.1 for 10%).
nu_range (tuple, optional) – Frequency interval (nu_min, nu_max) where the systematic is applied.
dataset (str, optional) – Dataset label where the systematic is applied.

filter_UL(val=None)[source]#

Remove rows considered upper limits by error threshold.

Parameters:: val (float, optional) – Error threshold; defaults to self.UL_value.

filter_data_set(filters, exclude=False, silent=False)[source]#

Filter rows by dataset label.

Parameters:

filters (str) – Comma-separated dataset names.
exclude (bool, optional) – If False keep only listed datasets; if True remove them.
silent (bool, optional) – Suppress informational prints when True.

filter_freq(nu_min=None, nu_max=None, exclude=False)[source]#

Filter rows by frequency interval.

Parameters:

nu_min (float, optional) – Minimum frequency in Hz to include.
nu_max (float, optional) – Maximum frequency in Hz to include.
exclude (bool, optional) – If True, remove selected rows instead of keeping them.

filter_time(T_min=None, T_max=None, exclude=False)[source]#

Filter rows by time interval.

Parameters:

T_min (float, optional) – Minimum T_start (MJD) to include.
T_max (float, optional) – Maximum T_stop (MJD) to include.
exclude (bool, optional) – If True, remove selected rows instead of keeping them.

find_time_span(dataset=None, silent=True, get_values=False)[source]#

Compute time-span summary for selected rows.

Parameters:

dataset (str, optional) – Dataset label. If omitted, uses all rows.
silent (bool, optional) – If False, print the summary.
get_values (bool, optional) – If True, return summary values.

Returns:

Returns (T_start, T_stop, DT, n_points) when get_values=True, otherwise None.

Return type:

tuple or None

get_data_points(log_log=False, skip_UL=False, frame='obs', density=False)[source]#

Return data arrays ready for plotting or fitting.

Parameters:

log_log (bool, optional) – If True, return log10-transformed arrays.
skip_UL (bool, optional) – If True, exclude rows flagged as upper limits.
frame ({"obs", "src"}, optional) – Output frame of returned arrays.
density (bool, optional) – If True, return density-style values (divide by frequency).

Returns:

(x, y, dx, dy) arrays in the requested frame/scale.

Return type:

tuple

get_data_sets()[source]#

Return unique dataset labels present in the table.

Returns:: Dataset labels.
Return type:: list

get_gammapy_table()[source]#

Build a Gammapy flux-points table from the current data.

Returns:: Table with e_ref, e2dnde and e2dnde_err columns plus metadata.
Return type:: astropy.table.Table

get_time_span(dataset=None)[source]#

Return time-span summary for one dataset or all data.

Parameters:: dataset (str, optional) – Dataset label. If omitted, uses all rows.
Returns:: (T_start, T_stop, DT, n_points) in MJD units.
Return type:: tuple

group_data(N_bin=None, bin_width=None, correct_dispersions=True, nu_min=None, nu_max=None)[source]#

Rebin data in logarithmic-frequency bins.

Parameters:

N_bin (int, optional) – Number of bins. Mutually exclusive with bin_width.
bin_width (float, optional) – Log10-bin width. Mutually exclusive with N_bin.
correct_dispersions (bool, optional) – Apply weighted-mean dispersion correction inside each bin.
nu_min (float, optional) – Lower bound of rebinning interval in Hz.
nu_max (float, optional) – Upper bound of rebinning interval in Hz.

lin_to_log(val=None, err=None)[source]#

Wrapper to convert linear values/errors to log10 space.

Parameters:

val (array-like, optional) – Linear values.
err (array-like, optional) – Linear errors associated with val.

Returns:

Converted values and, when provided, propagated log errors.

Return type:

array-like or list

static load(file_name)[source]#

Load a serialized ObsData instance from disk.

Parameters:: file_name (str) – Pickle file path.
Returns:: Loaded object with data sorted by frequency.
Return type:: ObsData

log_to_lin(log_val=None, log_err=None)[source]#

Wrapper to convert log10 values/errors back to linear space.

Parameters:

log_val (array-like, optional) – Log10 values.
log_err (array-like, optional) – Errors in log10 space.

Returns:

Converted linear values and, when provided, propagated errors.

Return type:

array-like or list

plot_sed(plot_obj=None, frame='obs', color=None, fmt='o', ms=4, mew=0.5, figsize=None, show_dataset=False, density=False)[source]#

Plot observational SED points.

Parameters:

plot_obj (jetset.plot_sedfit.PlotSED, optional) – Existing plot object; if omitted a new one is created.
frame ({"obs", "src"}, optional) – Frame used for plotting.
color (str, optional) – Matplotlib color.
fmt (str, optional) – Marker format.
ms (float, optional) – Marker size.
mew (float, optional) – Marker edge width.
figsize (tuple, optional) – Figure size used only when creating a new plot.
show_dataset (bool, optional) – If True, plot each dataset label as a separate series.
density (bool, optional) – If True, plot density representation.

Returns:

Plot object containing added data traces.

Return type:

jetset.plot_sedfit.PlotSED

plot_time_spans(save_as=None)[source]#

Plot time coverage per dataset on a single figure.

Parameters:: save_as (str, optional) – Optional output path where the figure is saved.
Returns:: Generated figure.
Return type:: matplotlib.figure.Figure

remove_dupl_entries(data)[source]#

remove duplicate entries

Parameters:: data – (array) 2-dim array storing the the table of the data.
Returns msk:: a boolean array to mask the duplicated entries

Note

One entry is flagged as duplicated as each comlum in a row of the data table is equal to the corresponding elements of another row

reset_data()[source]#

Reset working data to the original loaded dataset.

Notes

Rebuilds internal tables from the unfiltered source copy.

save(file_name)[source]#

Serialize this ObsData instance to disk.

Parameters:: file_name (str) – Pickle output path.

set_UL(val=None)[source]#

Flag upper limits from an error-threshold criterion.

Parameters:: val (float, optional) – Threshold applied to dnuFnu (or log equivalent) to define UL points. If omitted, uses self.UL_value.

set_error(error_value, nu_range=None, dataset=None, data_msk=None)[source]#

Set relative statistical errors on selected rows.

Parameters:

error_value (float) – Relative error factor; applied as error_value * nuFnu.
nu_range (tuple, optional) – Frequency interval (nu_min, nu_max) to target.
dataset (str, optional) – Dataset label to target.
data_msk (array-like of bool, optional) – Additional boolean mask combined with internal selection masks.

set_fake_error(val)[source]#

Set fallback relative errors used in fits and plotting.

Parameters:: val (float) – Relative error assigned to the dnuFnu_fake columns.

set_zero_error(val=0.2, replace_zero=True)[source]#

Mark and optionally replace non-positive uncertainty values.

Parameters:

val (float, optional) – Relative replacement error used when replace_zero=True.
replace_zero (bool, optional) – If True, replace zero/negative errors with a fallback value.

show_data_sets()[source]#

Print currently available dataset labels.

Notes

Labels are read from the dataset column of the working table.

show_time_span(dataset=None)[source]#: Print time-span summary for one dataset or all data.

ObsData

Contents

ObsData#