API Reference#

class bsavi.Observable(name, data, plot_type=None, plot_opts=None, latex_labels=None)#

Annotate your data with names and plotting instructions to easily create interactive plots. Observable accepts data in the form of tables where each row contains a set of datapoints to plot.

name#

Specifies the display name of the observable for things like plot titles.

Type:

str or list[str]

data#

The data associated with that observable. Can be python dict (or pandas DataFrame) whose keys (or column names) will be used for things like plot axis labels.

Type:

dict-like or list[dict-like]

Value:

None

plot_type#

Specifies how the data should be visualized. Currently can pick either 'Curve' that connects data points together, 'Bars' for a series of columns with their heights determined by the y-axis value at each point, or 'Scatter' for a simple scatter plot. The default value is 'Curve'

Pass a single value to set it for all datasets in the Observable, or a list of values to be set for each dataset.

Type:

str or list[str]

plot_opts#

Customization options for the observable plot. For more information see HoloViews documentation.

Type:

HoloViews Options object

Value:

None

latex_labels#

A dictionary that has parameter labels as keys and their corresponding LaTeX format as values.

Type:

dict

Value:

None

properties()#

Prints information about the Observable.

generate_plot(index)#

Generates plots of the data at the given indexes. Will call Observable.myfunc with Observable.myfunc_args on the data if given. The plots are returned as a dictionary of plot objects which can be manipulated as you wish.

Parameters:

index (list) – A list of indexes

Returns:

A dictionary of Holoviews Elements

draw_plot(index)#

Displays an interactive plot of the data at the given index. Whereas Observable.generate_plot() returns a dict of plot objects but does not display them, this method will display plots arranged in a layout when evaluated in a Jupyter Notebook cell.

Parameters:

index (list) – A list of indexes

Returns:

A layout of Holoviews Elements

class bsavi.LiveObservable(name, myfunc, myfunc_args, plot_type=None, plot_opts=None, latex_labels=None)#

Annotate a function with names and plotting instructions to easily create interactive plots. Live Observable will call the function to get a set of datapoints to plot.

name#

Specifies the display name of the observable for things like plot titles.

Type:

str or list[str]

myfunc#

A user-provided function that returns data in the same data format that Observable.data accepts. Can return more than one set of data.

Type:

Callable

Value:

None

myfunc_args#

Arguments for the user-provided function Observable.myfunc.

Type:

tuple

Value:

None

plot_type#

Specifies how the data should be visualized. Currently can pick either 'Curve' that connects data points together, 'Bars' for a series of columns with their heights determined by the y-axis value at each point, or 'Scatter' for a simple scatter plot. The default value is 'Curve'

Pass a single value to set it for all datasets in the Observable, or a list of values to be set for each dataset.

Type:

str or list[str]

plot_opts#

Customization options for the observable plot. For more information see HoloViews documentation.

Type:

HoloViews Options object

Value:

None

latex_labels#

A dictionary that has parameter labels as keys and their corresponding LaTeX format as values.

Type:

dict

Value:

None

properties()#

Prints information about the Observable.

generate_plot(index)#

Generates plots of the data at the given indexes. Will call Observable.myfunc with Observable.myfunc_args on the data if given. The plots are returned as a dictionary of plot objects which can be manipulated as you wish.

Parameters:

index (list) – A list of indexes

Returns:

A dictionary of Holoviews Elements

draw_plot(index)#

Displays an interactive plot of the data at the given index. Whereas Observable.generate_plot() returns a dict of plot objects but does not display them, this method will display plots arranged in a layout when evaluated in a Jupyter Notebook cell.

Parameters:

index (list) – A list of indexes

Returns:

A layout of Holoviews Elements

bsavi.viz(data, observables=None, show_observables=False, latex_dict=None)#

Displays an interactive dashboard that links data to observables.

Parameters:

  • data (dict-like) – The data or distribution to be visualized as a scatterplot

  • observables (list[bsavi.Observable]) – A list of the observables to be visualized

  • show_observables (bool) – Whether to display the observable plots or not. Default behavior is: True if observables are given, False if not.

  • latex_dict (dict) – A dictionary containing the LaTeX formatting for the scatterplot axis labels

Returns:

A collection of Panel components

bsavi.loaders.load_params(filename)#

Reads in a .paramnames file and returns a dict of each parameter’s plain text and LaTeX name. Assumes that the file is in the proper format: each line should contain one plain text param name and its LaTeX counterpart separated by a tab character (\t). Any amount of whitespace on either side of the tab character is acceptable.

Parameters:

filename (str) – path to the .paramnames file or glob pattern. If glob returns multiple paths, load_params will only used the first one

Returns:

a dict of parameter names and LaTeX code

bsavi.loaders.load_chains(path, params, params_only=True)#

Reads in a chain file and converts it to a DataFrame. Assumes that the file is a .txt file with the following columns: weight, -LogLkl, param1, param2, ….

Weight is the number of iterations the MCMC sampler stayed at that parameter set (the sample weight) and -LogLkl is the negative log of the likelihood. This is the standard format of both CosmoMC and Monte-Python chain files.

Parameters:

  • path (str, list['str']) – name of the chain file, list of names, or glob pattern

  • params (list['str']) – list of parameter names which will be used as column names for the DataFrame.

  • params_only (bool) – whether to ignore the first two columns of the chain file (weight and -LogLKL). Default is True, which will disregard those columns when reading in the file.

Returns:

Pandas DataFrame

bsavi.cosmo.run_class(index, sample)#

Calls the CLASS code on a given index of the sample data to calculate the matter power spectrum \(P(k)\), the lensed power spectrum of the CMB temperature \(C_{l}^{TT}\), and the lensed power spectrum of the CMB polarization \(C_{l}^{EE}\).

Uses the following settings:

{'output':'mPk, tCl, pCl, lCl','P_k_max_1/Mpc':3.0, 'lensing':'yes'}
Parameters:

  • index (int) – index location of the sample to be run through CLASS

  • sample (Pandas DataFrame) – a DataFrame where each row contains samples of cosmological parameters which CLASS accepts as inputs

Returns:

the three power spectra in the form of a dictionary where each key contains an array of wave numbers \(k\) or multipole moments \(\ell\) and each value contains an array of the calculated values for each \(k\) or \(\ell\).

bsavi.cosmo.compute_residuals(index, sample, sample_CDM)#

Useful for exploring beyond-CDM cosmologies. Calls the CLASS code on two sets of sample data (one with beyond-CDM parameters, and one with CDM parameters), at the specified index. Computes the percent difference in the three observables (\(P(k)\), \(C_{l}^{TT}\), \(C_{l}^{EE}\)) for each value of \(k\) or \(\ell\).

Parameters:

  • index (int) – index location of the sample to be run through CLASS

  • sample (Pandas DataFrame) – a DataFrame where each row contains samples of beyond-CDM cosmological parameters

  • sample_CDM (Pandas DataFrame) – a DataFrame where each row contains samples of LCDM cosmological parameters

Returns:

the power spectrum residuals in the same format as run_class()