PDF¶

class turbustat.statistics.PDF(img, min_val=-inf, bins=None, weights=None, normalization_type=None)[source]¶

Bases: BaseStatisticMixIn

Create the PDF of a given array.

Parameters:

imgnumpy.ndarray or astropy.io.fits.PrimaryHDU or astropy.io.fits.ImageHDU or spectral_cube.Projection or spectral_cube.Slice or SpectralCube: A 1-3D array.
min_valfloat, optional: Minimum value to keep in the given image.
binslist or numpy.ndarray or int, optional: Bins to compute the PDF from.
weightsnumpy.ndarray or astropy.io.fits.PrimaryHDU or astropy.io.fits.ImageHDU or spectral_cube.Projection or spectral_cube.Slice or SpectralCube, optional: Weights to apply to the image. Must have the same shape as the image.
normalization_type{“standardize”, “center”, “normalize”, “normalize_by_mean”}, optional: See data_normalization.

Examples

>>> from turbustat.statistics import PDF
>>> from astropy.io import fits
>>> moment0 = fits.open("Design4_21_0_0_flatrho_0021_13co.moment0.fits")[0]  
>>> pdf_mom0 = PDF(moment0).run(verbose=True)  

Attributes Summary

`bins`	Bin centers.
`data`
`distance`
`ecdf`	ECDF values in `bins`.
`header`
`model_params`	Parameters of the fitted model.
`model_stderrs`	Standard errors of the fitted model.
`need_header_flag`
`no_data_flag`
`normalization_type`
`pdf`	PDF values in `bins`.

Methods Summary

`corner_plot`(**kwargs)	Create a corner plot from the MCMC.
`find_at_percentile`(percentiles)	Return the values at the given percentiles.
`find_percentile`(values)	Return the percentiles of given values from the data distribution.
`fit_pdf`([model, verbose, fit_type, floc, ...])	Fit a model to the PDF.
`input_data_header`(data, header[, need_copy])	Check if the header is given separately from the data type.
`load_beam`([beam])	Try loading the beam from the header or a given object.
`load_results`(pickle_file)	Load in a saved pickle file.
`make_ecdf`()	Create the ECDF.
`make_pdf`([bins])	Create the PDF.
`plot_distrib`([save_name, color, fit_color, ...])	Plot the PDF distribution and (if fitted) the best fit model.
`run`([verbose, save_name, bins, do_fit, ...])	Compute the PDF and ECDF.
`save_results`(output_name[, keep_data])	Save the results of the SCF to avoid re-computing.
`trace_plot`(**kwargs)	Create a trace plot from the MCMC.

Attributes Documentation

bins¶: Bin centers.

data¶

distance¶

ecdf¶: ECDF values in bins.

header¶

model_params¶: Parameters of the fitted model.

model_stderrs¶: Standard errors of the fitted model. If using an MCMC, the 15th and 85th percentiles are returned.

need_header_flag = True¶

no_data_flag = False¶

normalization_type¶

pdf¶: PDF values in bins.

Methods Documentation

corner_plot(**kwargs)[source]¶

Create a corner plot from the MCMC. Requires the ‘corner’ package.

Parameters:

kwargsPassed to corner.

find_at_percentile(percentiles)[source]¶

Return the values at the given percentiles.

Parameters:

percentilesfloat or np.ndarray: Percentile or array of percentiles. Must be between 0 and 100.

find_percentile(values)[source]¶

Return the percentiles of given values from the data distribution.

Parameters:

valuesfloat or np.ndarray: Value or array of values.

fit_pdf(model=<scipy.stats._continuous_distns.lognorm_gen object>, verbose=False, fit_type='mle', floc=True, loc=0.0, fscale=False, scale=1.0, **kwargs)[source]¶

Fit a model to the PDF. Use statsmodel’s generalized likelihood setup to get uncertainty estimates and such.

Parameters:

modelscipy.stats distribution, optional

Pass any scipy distribution. NOTE: All fits assume loc can be fixed to 0. This is reasonable for all realistic PDF forms in the ISM.

verbosebool, optional

Enable printing of the fit results.

fit_type{‘mle’, ‘mcmc’}, optional

Type of fitting to use. By default Maximum Likelihood Estimation: (‘mle’) is used. An MCMC approach (‘mcmc’) may also be used. This requires the optional emcee to be installed. kwargs can be passed to adjust various properties of the MCMC chain.

flocbool, optional

Fix the loc parameter when fitting.

locfloat, optional

Value to set loc to when fixed.

fscalebool, optional

Fix the scale parameter when fitting.

scalefloat, optional

Value to set scale to when fixed.

kwargsPassed to EnsembleSampler.

input_data_header(data, header, need_copy=False)¶: Check if the header is given separately from the data type.

load_beam(beam=None)¶

Try loading the beam from the header or a given object.

Parameters:

beamBeam, optional: The beam.

static load_results(pickle_file)¶

Load in a saved pickle file.

Parameters:

pickle_filestr: Name of filename to load in.

Returns:

selfSave statistic class: Statistic instance with saved results.

Examples

Load saved results. >>> stat = Statistic.load_results(“stat_saved.pkl”) # doctest: +SKIP

make_ecdf()[source]¶: Create the ECDF.

make_pdf(bins=None)[source]¶

Create the PDF.

Parameters:

binslist or numpy.ndarray or int, optional: Bins to compute the PDF from. Overrides initial bin input.

plot_distrib(save_name=None, color='r', fit_color='k', show_ecdf=True)[source]¶

Plot the PDF distribution and (if fitted) the best fit model. Optionally show the ECDF and fit ECDF, too.

Parameters:

save_namestr,optional: Save the figure when a file name is given.
color{str, RGB tuple}, optional: Color to show the Genus curves in.
fit_color{str, RGB tuple}, optional: Color of the fitted line. Defaults to color when no input is given.
show_ecdfbool, optional: Plot the ECDF when enabled.

run(verbose=False, save_name=None, bins=None, do_fit=True, model=<scipy.stats._continuous_distns.lognorm_gen object>, color=None, **kwargs)[source]¶

Compute the PDF and ECDF. Enabling verbose provides a summary plot.

Parameters:

verbosebool, optional: Enables plotting of the results.
save_namestr,optional: Save the figure when a file name is given.
binslist or numpy.ndarray or int, optional: Bins to compute the PDF from. Overrides initial bin input.
do_fitbool, optional: Enables (by default) fitting a given model.
modelscipy.stats distribution, optional: Pass any scipy distribution. See fit_pdf.
color{str, RGB tuple}, optional: Color to show the Genus curves in when verbose=True.
kwargsPassed to fit_pdf.

save_results(output_name, keep_data=False)¶

Save the results of the SCF to avoid re-computing. The pickled file will not include the data cube by default.

Parameters:

output_namestr: Name of the outputted pickle file.
keep_databool, optional: Save the data cube in the pickle file when enabled.

trace_plot(**kwargs)[source]¶

Create a trace plot from the MCMC.

Parameters:

kwargsPassed to plot.