Dataset handling

Measurement datasets are a central part in model validation and therefore we designed the Dataset class that offer a number of useful methods to deal with them.

A typical workflow consists in casting your log-data into Signal objects and then use the created Signals to instantiate a Dataset object.

Signals

Signal are used to represent real-world signals.

Dymoval Signals are Typeddict with the following keys

Keys

Signal.name	Signal name.
Signal.values()
Signal.signal_unit	Signal unit.
Signal.sampling_period	Signal sampling period.
Signal.time_unit	Signal sampling period.

Functions

Dymoval offers few function for dealing with Signals. Such functions are the following

validate_signals(*signals)	Perform a number of checks to verify that the passed list of Signals can be used to create a Dataset.
plot_signals(*signals)	Plot Signals.

Dataset class

The Dataset class is used to store and manipulate datasets.

Since to validate a model you need a datasets, objects of this class are used also to instantiate ValidationSession objects, and the passed Dataset object becomes an attribute of the newly created ValidationSession object.

A Dataset object can be instantiated in two ways

1. Through a list of dymoval Signals (see validate_signals() )

2. Through a pandas DataFrame with a specific structure (see validate_dataframe())

Constructor

Dataset(name, signal_list, u_names, y_names)

The Dataset class stores the candidate signals to be used as a dataset and it provides methods for analyzing and manipulating them.

Attributes

Dataset.name
Dataset.dataset
Dataset.coverage
Dataset.information_level

Manipulation methods

Dataset.add_input(*signals)	Add input signals to the Dataset object.
Dataset.add_output(*signals)	Add output signals to the Dataset object.
Dataset.remove_signals(*signals)	Remove signals from the Dataset.
Dataset.remove_means(*signals)	Remove the mean value to the specified signals.
Dataset.remove_offset(*signals_values)	Remove specified offsets to the specified signals.
Dataset.remove_NaNs(**kwargs)	Replace NaN:s values in the Dataset.
Dataset.apply(*signal_function, **kwargs)	Apply a function to specified signals and change their unit.
Dataset.low_pass_filter(*signals_values)	Low-pass filter a list of specified signals.
Dataset.fft(*signals)	Return the FFT of the dataset as pandas DataFrame.
Dataset.trim(*signals[, tin, tout, verbosity])	Trim the Dataset Dataset object.

Plotting methods

Dataset.plot(*signals[, overlap, ...])	Plot the Dataset.
Dataset.plotxy(*signal_pairs[, layout, ...])	Plot a signal against another signal in a plane (XY-plot).
Dataset.plot_coverage(*signals[, nbins, ...])	Plot the dataset Dataset coverage in histograms.
Dataset.plot_spectrum(*signals[, kind, ...])	Plot the spectrum of the specified signals in the dataset in different format.
change_axes_layout(fig, nrows, ncols)	Change Axes layout of an existing Matplotlib Figure.

Other methods

Dataset.dump_to_signals()	Dump a Dataset object into a list of Signals objects.
Dataset.dataset_values()	Return the dataset values as a tuple (t,u,y) of numpy ndarrays.
Dataset.export_to_mat(filename)	Write the dataset in a .mat file.
Dataset.dataset_values()	Return the dataset values as a tuple (t,u,y) of numpy ndarrays.
Dataset.signal_list()	Return the list of signals in form (["INPUT" | "OUTPUT"], name, unit)
validate_dataframe(df)	Check if a pandas DataFrame is suitable for instantiating a Dataset object.
compare_datasets(*datasets[, kind, layout, ...])	Compare different Datasets graphically by overlapping them.