calibr8.utils
This module implements helper functions for a variety of tasks, including imports, timestamp parsing and plotting.
- exception calibr8.utils.BuildMismatchException
Bases:
CompatibilityException
- exception calibr8.utils.CompatibilityException
Bases:
Exception
- class calibr8.utils.ImportWarner(module_name)
Bases:
objectMock for an uninstalled package, raises ImportError when used.
- exception calibr8.utils.MajorMismatchException
Bases:
CompatibilityException
- exception calibr8.utils.MinorMismatchException
Bases:
CompatibilityException
- exception calibr8.utils.PatchMismatchException
Bases:
CompatibilityException
- calibr8.utils.assert_version_match(vA: str, vB: str)
Compares two version numbers and raises exceptions that indicate where they missmatch.
- Parameters:
- vAstr
first version number
- vBstr
second version number
- Raises:
- MajorMismatchException
difference on the first level
- MinorMismatchException
difference on the second level
- PatchMismatchException
difference on the third level
- BuildMismatchException
difference on the fourth level
- calibr8.utils.check_scale_degree(scale_degree: int) int
Evaluates user input for the scale degree and raises warning/error if the value is unexpected. Returns the user input.
- Parameters:
- scale_degreeint
Degree of scale as set by the user.
- Returns:
- scale_degreeint
Degree of scale as set by the user.
- Raises:
- ValueError
scale_degree is None or negative
- UserWarning
scale_degree is unexpectedly high
- calibr8.utils.format_datetime(dt: datetime | None) str | None
Formats a datetime like 2020-08-05T13:37:00Z.
- Returns:
- resultoptional, str
may be None when the input was None
- calibr8.utils.istensor(input: object) bool
- “Convenience function to test whether an input is a TensorVariable
or if an input array or list contains TensorVariables.
- Parameters:
- inputobject
an object shat shall be analyzed
- calibr8.utils.parse_datetime(s: str | None) datetime | None
Parses a timezone-aware datetime formatted like 2020-08-05T13:37:00Z.
- Returns:
- resultoptional, datetime
may be None when the input was None
- calibr8.utils.plot_continuous_band(ax, independent, model, residual_type: str | None = None)
Helper function for plotting the 68, 90 and 95 % likelihood-bands of a univariate distribution.
- Parameters:
- axmatplotlib.Axes
subplot object to plot into
- independentarray-like
x-values for the plot
- modelCalibrationModel
A fitted calibration model with data. The predict_dependent method should return a tuple where the mean is the first entry.
- residual_typestr, optional
One of { None, “absolute”, “relative” }. Specifies if bands are for no, absolute or relative residuals.
- Returns:
- artistslist of matplotlib.Artist
the created artists (1x Line2D, 6x PolyCollection (alternating plot & legend))
- calibr8.utils.plot_model(model, *, fig: Figure | None = None, axs: Sequence[Axes] | None = None, residual_type='absolute', band_xlim: Tuple[float | None, float | None] = (None, None))
Makes a plot of the model with its data.
- Parameters:
- modelCalibrationModel
A fitted calibration model with data. The predict_dependent method should return a tuple where the mean is the first entry.
- figoptional, matplotlib.figure.Figure
An existing figure (to be used in combination with [axs] argument).
- axsoptional, [matplotlib.axes.Axes]
matplotlib subplots to use instead of creating new ones.
- residual_typeoptional, str
Specifies if residuals are plotted absolutely or relatively.
- band_xlimtuple
Optional overrides for the minimum/maximum x coordinates of the likelihood bands in the left and center plots. Either entry can be None or a float. Defaults to the min/max of the calibration data.
- Returns:
- figmatplotlib.figure.Figure
the (created) figure
- axs[matplotlib.axes.Axes]
subplots of x-linear, x-log and residuals (xlog)
- calibr8.utils.plot_norm_band(ax, independent, mu, scale)
Helper function for plotting the 68, 90 and 95 % likelihood-bands of a Normal distribution.
- Parameters:
- axmatplotlib.Axes
subplot object to plot into
- independentarray-like
x-values for the plot
- muarray-like
mu parameter of the Normal distribution
- scalearray-like
scale parameter of the Normal distribution
- Returns:
- artistslist of matplotlib.Artist
the created artists (1x Line2D, 6x PolyCollection (alternating plot & legend))
- calibr8.utils.plot_t_band(ax, independent, mu, scale, df, *, residual_type: str | None = None)
Helper function for plotting the 68, 90 and 95 % likelihood-bands of a t-distribution.
- Parameters:
- axmatplotlib.Axes
subplot object to plot into
- independentarray-like
x-values for the plot
- muarray-like
mu parameter of the t-distribution
- scalearray-like
scale parameter of the t-distribution
- dfarray-like
density parameter of the t-distribution
- residual_typestr, optional
One of { None, “absolute”, “relative” }. Specifies if bands are for no, absolute or relative residuals.
- Returns:
- artistslist of matplotlib.Artist
the created artists (1x Line2D, 6x PolyCollection (alternating plot & legend))