nessai.plot =========== .. py:module:: nessai.plot .. autoapi-nested-parse:: Plotting utilities. .. !! processed by numpydoc !! Functions --------- .. autoapisummary:: nessai.plot.nessai_style nessai.plot.sanitise_array nessai.plot.plot_live_points nessai.plot.plot_1d_comparison nessai.plot.plot_indices nessai.plot.plot_loss nessai.plot.plot_trace nessai.plot.plot_histogram nessai.plot.corner_plot Module Contents --------------- .. py:function:: nessai_style(line_styles=True) Decorator for plotting function that sets the style. Functions as both standard decorator :code:`@nessai_style` or as a callable decorator :code:`@nessai_style()`. Style can be disabled by setting :py:data:`nessai.config.DISABLE_STYLE` to :code:`True`. :Parameters: **line_styles** : boolean Use custom line styles. .. !! processed by numpydoc !! .. py:function:: sanitise_array(a: numpy.ndarray, /, a_min: Optional[float] = None, a_max: Optional[float] = None) Sanitise an array for plotting. If :code:`x_min` is not specified, it is set to the value in :code:`nessai.config.plotting.clip_min`. :Parameters: **x** : array_like Array to sanitise. **x_min** : float, optional Minimum value to clip the data to. **xmax** : float, optional Maximum value to clip the data to. :Returns: np.ndarray Sanitised array. .. !! processed by numpydoc !! .. py:function:: plot_live_points(live_points, filename=None, bounds=None, c=None, **kwargs) Plot a set of live points in a corner-like plot. Will drop columns where all elements are NaNs :Parameters: **live_points** : ndarray Structured array of live points to plot. **filename** : str Filename for resulting figure **bounds** : dict: Dictionary of lower and upper bounds to plot **c** : str, optional Name of field in the structured array to use as the hue when plotting the samples. If not specified, no hue is used. **kwargs** Keyword arguments used to update the pairplot kwargs. Diagonal and off- diagonal plots can be configured with ``diag_kws`` and ``plot_kws``. .. !! processed by numpydoc !! .. py:function:: plot_1d_comparison(*live_points, parameters=None, labels=None, colours=None, bounds=None, hist_kwargs={}, filename=None, convert_to_live_points=False) Plot 1d histograms comparing different sets of live points :Parameters: **\*live_points** : iterable of ndarrays Variable length argument of live points in structured arrays with fields. Also see ``parameters`` argument. **parameters** : array_like, optional Array of parameters (field names) to plot. Default None implies all fields are plotted. **labels** : list, optional List of labels for each structured array being plotted (default None). If None each set of live points is labelled numerically **colours** : list, optional List of colours to use for each set of live points. **bounds** : dict, optional Dictionary of upper and lowers bounds to plot. Each key must match a field and each value must be an interable of length 2 in order lower then upper bound. If None (default), no bounds plotted. **hist_kwargs** : dict, optional Dictionary of keyword arguments passed to matplotlib.pyplot.hist. **filename** : str, optional Name of file for saving figure. (Default None implies figure is not saved). **convert_to_live_points** : bool, optional Set to true if inputs are not structured arrays of live points .. !! processed by numpydoc !! .. py:function:: plot_indices(indices, nlive=None, filename=None, ks_test_mode='D+', confidence_intervals=(0.68, 0.95, 0.997), plot_breakdown=True, n_breakdown=8, cmap='viridis') Histogram indices for index insertion tests, also includes the cmf. :Parameters: **indices** : array_like List of insertion indices to plot **nlive** : int Number of live points used in the nested sampling run. If nlive is not specified, it is set to the maximum value in indices. **filename** : str Filename used to save the figure. **plot_breakdown** : bool, optional If true, then the CDF for every nlive points is also plotted as grey lines. **ks_test_mode** : Literal["D+", "D-"] Mode for computing the KS test. See :py:func:`nessai.utils.indices.compute_indices_ks_test` for details. **confidence_intervals** : tuple Confidence intervals to plot as shaded regions on the cmf plot. **plot_breakdown** : bool If true, plots the cmf for batches of samples over the course of the run. The number of batches in controlled by :code:`n_breakdown`. **n_breakdown** : int The number of batches to plot. Also see :code:`plot_breakdown`. **cmap** : str Colourmap to use when :code:`plot_breakdown=True`. .. !! processed by numpydoc !! .. py:function:: plot_loss(epoch, history, filename=None) Plot the loss function per epoch. :Parameters: **epoch** : int Final training epoch **history** : dict Dictionary with keys ``'loss'`` and ``'val_loss'`` **filename** : str, optional Path for saving the figure. If not specified figure is returned instead. .. !! processed by numpydoc !! .. py:function:: plot_trace(log_x: numpy.ndarray, nested_samples: numpy.ndarray, parameters: Optional[List[str]] = None, live_points: Optional[numpy.ndarray] = None, log_x_live_points: Optional[numpy.ndarray] = None, labels: Optional[List[str]] = None, filename: Optional[str] = None, **kwargs) Produce trace plot for the nested samples. By default this includes all parameters in the samples, not just those included in the model being sampled. :Parameters: **log_x** : array_like Array of log prior volumes **nested_samples** : ndarray Array of nested samples to plot **parameters** : list, optional List of parameters to include the trace plot. If not specified, all of the parameters in the nested samples are included. **live_points** : ndarray Optional array of live points to include in the plot. See also :code:`log_x_live_points`. **log_x_live_points** : ndarray Optional array of log-prior volumes for the live points. Required if :code:`live_points` is specified. **labels** : list, optional List of labels to use instead of the names of parameters **filename** : str, optional Filename for saving the plot, if none plot is not saved and figure is returned instead. **kwargs** Keyword arguments passed to :code:`matplotlib.pyplot.plot`. .. !! processed by numpydoc !! .. py:function:: plot_histogram(samples, label=None, filename=None, **kwargs) Plot a histogram of samples. :Parameters: **samples** : array_like Samples to plot. **label** : str, optional Label to the x axis. **filename** : str, optional Filename for saving the plot. If not specified, figure is returned. **kwargs** Keyword arguments passed to `matplotlib.pyplot.hist`. .. !! processed by numpydoc !! .. py:function:: corner_plot(array, include=None, exclude=None, labels=None, truths=None, filename=None, **kwargs) Produce a corner plot for a structured array. Removes any fields with no dynamic range. :Parameters: **array** : numpy.ndarray Structured array **include** : Optional[list] List of parameters to plot. **exclude** : Optional[list] List of parameters to exclude. **labels** : Optional[Iterable] Labels for each parameter that is to be plotted. **truths** : Optional[Union[Iterable, Dict]] Truth values for each parameters, parameters can be skipped by setting the value to None. **filename** : Optional[str] Filename for saving the plot. If not specified, figure is returned. **kwargs** : Dict[Any] Dictionary of keyword arguments passed to :code:`corner.corner`. .. !! processed by numpydoc !!