Module for fitting a linear model to the data

This module defines the LinRegPlotter and the DensityRegPlotter plotter classes that can be used to fit a linear model to the data and visualize it.

Classes:

`Ci`(key[, plotter, index_in_list, ...])	Draw a confidence interval
`DensityRegPlotter`([data, ax, auto_update, ...])	A plotter that visualizes the density of points together with a linear regression
`FitPointDensity`(key[, plotter, ...])
`FixPoint`(key[, plotter, index_in_list, ...])	Force the fit to go through a given point
`IdealLine`(key[, plotter, index_in_list, ...])	Draw an ideal line of the fit
`IdealLineColor`(args, *kwargs)	The colors of the ideal lines
`InitialParameters`(key[, plotter, ...])	Initial parameters for the `scipy.optimize.curve_fit()` function
`LinRegPlotter`([data, ax, auto_update, ...])	A plotter to visualize the fit on the data
`LinRegTranspose`(args, *kwargs)	Switch x- and y-axes
`LinearRegressionFit`(args, *kwargs)	Choose the linear fitting method
`LinearRegressionFitCombined`(args, *kwargs)	Choose the linear fitting method
`NBoot`(key[, plotter, index_in_list, ...])	Set the number of bootstrap resamples for the confidence interval
`ParameterBounds`(key[, plotter, ...])	Parameter bounds for the function parameters
`XFitRange`(args, *kwargs)	Specify the range for the fit to use for the x-dimension
`XLineRange`(args, *kwargs)	Specify how wide the range for the plot should be
`YFitRange`(args, *kwargs)	Specify the range for the fit to use for the y-dimension

Functions:

`bootstrap`(x, y, func, n_boot[, random_seed])	Simple bootstrap algorithm used to estimate the confidence interval
`calc_ci`(a[, which, axis])	Return a quantile range from an array of values.

class psy_reg.plotters.Ci(key, plotter=None, index_in_list=None, additional_children=[], additional_dependencies=[], **kwargs)

Bases: Formatoption

Draw a confidence interval

Size of the confidence interval for the regression estimate. This will be drawn using translucent bands around the regression line. The confidence interval is estimated using a bootstrap; for large datasets, it may be advisable to avoid that computation by setting this parameter to None.

Possible types

None: Do not draw and calculate a confidence interval
float: A quantile between 0 and 100

Attributes:

`dependencies`	list of str.
`fit`	fit Formatoption instance in the plotter
`fix`	fix Formatoption instance in the plotter
`group`	`str`.
`name`	`str`.
`nboot`	nboot Formatoption instance in the plotter
`priority`	`int`.
`transpose`	transpose Formatoption instance in the plotter

Methods:

`initialize_plot`(args, *kwargs)	Method that is called when the plot is made the first time
`update`(value)	Method that is call to update the formatoption on the axes

dependencies = ['transpose', 'fit', 'nboot', 'fix']: list of str. List of formatoptions that force an update of this formatoption if they are updated.

property fit: fit Formatoption instance in the plotter

property fix: fix Formatoption instance in the plotter

group = 'regression': str. Key of the group name in groups of this formatoption keyword

initialize_plot(*args, **kwargs)

Method that is called when the plot is made the first time

Parameters:: value – The value to use for the initialization

name = 'Draw a confidence interval': str. A bit more verbose name than the formatoption key to be included in the gui. If None, the key is used in the gui

property nboot: nboot Formatoption instance in the plotter

priority = 30

int. Priority value of the the formatoption determining when the formatoption is updated.

10: at the end (for labels, etc.)
20: before the plotting (e.g. for colormaps, etc.)
30: before loading the data (e.g. for lonlatbox)

property transpose: transpose Formatoption instance in the plotter

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

class psy_reg.plotters.DensityRegPlotter(data=None, ax=None, auto_update=None, project=None, draw=False, make_plot=True, clear=False, enable_post=False, **kwargs)

Bases: ScalarCombinedBase, DensityPlotter, LinRegPlotter

A plotter that visualizes the density of points together with a linear regression

Attributes:

`bins`	Specify the bins of the 2D-Histogramm
`cbar`	Specify the position of the colorbars
`ci`	Draw a confidence interval
`clabel`	Show the colorbar label
`color`	Set the color coding
`density`
`error`	Visualize the error range
`erroralpha`	Set the alpha value for the error range
`fit`	Choose the linear fitting method
`fix`	Force the fit to go through a given point
`id_color`	The colors of the ideal lines
`ideal`	Draw an ideal line of the fit
`labelprops`	Set the font properties of both, x- and y-label
`labelsize`	Set the size of both, x- and y-label
`labelweight`	Set the font size of both, x- and y-label
`legend`	Draw a legend
`legendlabels`	Set the labels of the arrays in the legend
`line_xlim`	Specify how wide the range for the plot should be
`lineplot`	Choose the line style of the plot
`nboot`	Set the number of bootstrap resamples for the confidence interval
`normed`	Specify the normalization of the histogram
`p0`	Initial parameters for the `scipy.optimize.curve_fit()` function
`param_bounds`	Parameter bounds for the function parameters
`plot`	Choose how to visualize a 2-dimensional scalar data field
`post`	Apply your own postprocessing script
`post_timing`	Determine when to run the `post` formatoption
`precision`	Set the precision of the data
`ticksize`	Change the ticksize of the ticklabels
`tickweight`	Change the fontweight of the ticks
`title`	Show the title
`transpose`	Switch x- and y-axes
`xlabel`	Set the x-axis label
`xlim`	Set the x-axis limits
`xrange`	Specify the range of the histogram for the x-dimension
`xrotation`	Rotate the x-axis ticks
`xticklabels`	Modify the x-axis ticklabels
`xtickprops`	Specify the x-axis tick parameters
`xticks`	Modify the x-axis ticks
`ylabel`	Set the y-axis label
`ylim`	Set the y-axis limits
`yrange`	Specify the range of the histogram for the x-dimension
`yrotation`	Rotate the y-axis ticks
`yticklabels`	Modify the y-axis ticklabels
`ytickprops`	Specify the y-axis tick parameters
`yticks`	Modify the y-axis ticks

bins

Specify the bins of the 2D-Histogramm

This formatoption can be used to specify, how many bins to use. In other words, it determines the grid size of the resulting histogram or kde plot. If however you also set the precision formatoption keyword then the minimum of precision and the bins specified here will be used.

Possible types

int: If 0, only use the bins specified by the precision keyword (raises an error if the precision is also set to 0), otherwise the number of bins to use
tuple (x, y) of int: The bins for x and y explicitly

cbar

Specify the position of the colorbars

Possible types

bool

True: defaults to ‘b’ False: Don’t draw any colorbar

str

The string can be a combination of one of the following strings: {‘fr’, ‘fb’, ‘fl’, ‘ft’, ‘b’, ‘r’, ‘sv’, ‘sh’}

‘b’, ‘r’ stand for bottom and right of the axes
‘fr’, ‘fb’, ‘fl’, ‘ft’ stand for bottom, right, left and top of the figure
‘sv’ and ‘sh’ stand for a vertical or horizontal colorbar in a separate figure

list

A containing one of the above positions

Examples

Draw a colorbar at the bottom and left of the axes:

>>> plotter.update(cbar='bl')

ci

Draw a confidence interval

Size of the confidence interval for the regression estimate. This will be drawn using translucent bands around the regression line. The confidence interval is estimated using a bootstrap; for large datasets, it may be advisable to avoid that computation by setting this parameter to None.

Possible types

None: Do not draw and calculate a confidence interval
float: A quantile between 0 and 100

clabel

Show the colorbar label

Set the label of the colorbar. You can insert any meta key from the xarray.DataArray.attrs via a string like '%(key)s'. Furthermore there are some special cases:

Strings like '%Y', '%b', etc. will be replaced using the datetime.datetime.strftime() method as long as the data has a time coordinate and this can be converted to a datetime object.
'%(x)s', '%(y)s', '%(z)s', '%(t)s' will be replaced by the value of the x-, y-, z- or time coordinate (as long as this coordinate is one-dimensional in the data)
any attribute of one of the above coordinates is inserted via axis + key (e.g. the name of the x-coordinate can be inserted via '%(xname)s').
Labels defined in the psyplot.rcParams 'texts.labels' key are also replaced when enclosed by ‘{}’. The standard labels are
- tinfo: %H:%M
- dtinfo: %B %d, %Y. %H:%M
- dinfo: %B %d, %Y
- desc: %(long_name)s [%(units)s]
- sdesc: %(name)s [%(units)s]

Possible types

str: The title for the set_label() method.

See also

clabelsize, clabelweight, clabelprops

color

Set the color coding

This formatoptions sets the color of the lines, bars, etc.

Possible types

None: to use the axes color_cycle
iterable: (e.g. list) to specify the colors manually

str

Strings may be any valid colormap name suitable for the matplotlib.cm.get_cmap() function or one of the color lists defined in the ‘colors.cmaps’ key of the psyplot.rcParams dictionary (including their reversed color maps given via the ‘_r’ extension).

matplotlib.colors.ColorMap: to automatically choose the colors according to the number of lines, etc. from the given colormap

density

error

Visualize the error range

This formatoption visualizes the error range. For this, you must provide a two-dimensional data array as input. The first dimension might be either of length

2 to provide the deviation from minimum and maximum error range from the data
3 to provide the minimum and maximum error range explicitly

Possible types

None: No errors are visualized
‘fill’: The area between min- and max-error is filled with the same color as the line and the alpha is determined by the fillalpha attribute

Examples

Assume you have the standard deviation stored in the 'std'-variable and the data in the 'data' variable. Then you can visualize the standard deviation simply via:

>>> psy.plot.lineplot(input_ds, name=[['data', 'std']])

On the other hand, assume you want to visualize the area between the 25th and 75th percentile (stored in the variables 'p25' and 'p75'):

>>> psy.plot.lineplot(input_ds, name=[['data', 'p25', 'p75']])

See also

erroralpha

erroralpha

Set the alpha value for the error range

This formatoption can be used to set the alpha value (opacity) for the error formatoption

Possible types

float: A float between 0 and 1

See also

error

fit

Choose the linear fitting method

This formatoption consists makes a linear fit of the data

Possible types

‘fit’ or ‘linear’: make a linear fit
‘robust’: make a robust linear fit
‘poly<deg>’: Make a polynomial fit of the order '<deg>'
function: A callable function that takes an x-array and a y-array as input and can be used for the scipy.optimize.curve_fit() function
any object with a fit and predict method: A model that with a fit signature such as model.fit(x, y).predict(x)
None: make no fit

Notes

You can access the intercept, slope and rsquared by the correponding attribute. E.g.:

>>> plotter.update(
...     legendlabels="%(intercept)s + %(slope)s * x, "
...     "$R^2$=%(rsquared)s"
... )

See also

fix

fix

Force the fit to go through a given point

Possible types

None: do not force the fit at all
float f: make a linear fit forced through (x, y) = (0, f)
tuple (x’, y’): make a linear fit forced through (x, y) = (x', y')

See also

fit

id_color

The colors of the ideal lines

Possible types

None: Let it be determined by the color cycle of the color formatoption
iterable: (e.g. list) to specify the colors manually

str

Strings may be any valid colormap name suitable for the matplotlib.cm.get_cmap() function or one of the color lists defined in the ‘colors.cmaps’ key of the psyplot.rcParams dictionary (including their reversed color maps given via the ‘_r’ extension).

matplotlib.colors.ColorMap: to automatically choose the colors according to the number of lines, etc. from the given colormap

See also

ideal

ideal

Draw an ideal line of the fit

Possible types

None: Don’t draw an ideal line
list of floats: The parameters for the line. If the fit formatoption is in 'robust' or 'fit', then the first value corresponds to the interception, the second to the slope. Otherwise the list corrensponds to the parameters as used in the fit function of the lines
list of list of floats: The same as above but with the specification for each array

See also

id_color

labelprops

Set the font properties of both, x- and y-label

Possible types

dict: A dictionary with the keys 'x' and (or) 'y' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is used for the x- and y-axis. The values in the dictionary can be one types below.
dict: Items may be any valid text property

labelsize

Set the size of both, x- and y-label

Possible types

dict: A dictionary with the keys 'x' and (or) 'y' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is used for the x- and y-axis. The values in the dictionary can be one types below.
float: The absolute font size in points (e.g., 12)
string: Strings might be ‘xx-small’, ‘x-small’, ‘small’, ‘medium’, ‘large’, ‘x-large’, ‘xx-large’.

labelweight

Set the font size of both, x- and y-label

Possible types

dict: A dictionary with the keys 'x' and (or) 'y' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is used for the x- and y-axis. The values in the dictionary can be one types below.
float: a float between 0 and 1000
string: Possible strings are one of ‘ultralight’, ‘light’, ‘normal’, ‘regular’, ‘book’, ‘medium’, ‘roman’, ‘semibold’, ‘demibold’, ‘demi’, ‘bold’, ‘heavy’, ‘extra bold’, ‘black’.

legend

Draw a legend

This formatoption determines where and if to draw the legend. It uses the labels formatoption to determine the labels.

Possible types

bool: Draw a legend or not
str or int: Specifies where to plot the legend (i.e. the location)
dict: Give the keywords for the matplotlib.pyplot.legend() function

See also

labels

legendlabels

Set the labels of the arrays in the legend

This formatoption specifies the labels for each array in the legend. You can insert any meta key from the xarray.DataArray.attrs via a string like '%(key)s'. Furthermore there are some special cases:

Strings like '%Y', '%b', etc. will be replaced using the datetime.datetime.strftime() method as long as the data has a time coordinate and this can be converted to a datetime object.
'%(x)s', '%(y)s', '%(z)s', '%(t)s' will be replaced by the value of the x-, y-, z- or time coordinate (as long as this coordinate is one-dimensional in the data)
any attribute of one of the above coordinates is inserted via axis + key (e.g. the name of the x-coordinate can be inserted via '%(xname)s').
Labels defined in the psyplot.rcParams 'texts.labels' key are also replaced when enclosed by ‘{}’. The standard labels are
- tinfo: %H:%M
- dtinfo: %B %d, %Y. %H:%M
- dinfo: %B %d, %Y
- desc: %(long_name)s [%(units)s]
- sdesc: %(name)s [%(units)s]

Possible types

str:: A single string that shall be used for all arrays.
list of str:: Same as a single string but specified for each array

See also

legend

line_xlim

Specify how wide the range for the plot should be

This formatoption specifies the range of the line to use

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

See also

xrange

lineplot

Choose the line style of the plot

Possible types

None: Don’t make any plotting
'area': To make an area plot (filled between y=0 and y), see matplotlib.pyplot.fill_between()
'areax': To make a transposed area plot (filled between x=0 and x), see matplotlib.pyplot.fill_betweenx()
'stacked': Make a stacked plot
str or list of str: The line style string to use ([‘solid’ | ‘dashed’, ‘dashdot’, ‘dotted’ | (offset, on-off-dash-seq) | ‘-’ | ‘–’ | ‘-.’ | ‘:’ | ‘None’ | ‘ ‘ | ‘’]).

nboot

Set the number of bootstrap resamples for the confidence interval

Parameters:: int – Number of bootstrap resamples used to estimate the ci. The default value attempts to balance time and stability; you may want to increase this value for “final” versions of plots.

See also

ci

normed

Specify the normalization of the histogram

This formatoption can be used to normalize the histogram. It has no effect if the density formatoption is set to 'kde'

Possible types

None

Do not make any normalization

str

One of

counts: To make the normalization based on the total number counts
area: To make the normalization basen on the total number of counts and area (the default behaviour of numpy.histogram2d())
x, col, column or columns: To normalize every column
y, row or rows: To normalize every row

See also

density

p0

Initial parameters for the scipy.optimize.curve_fit() function

This formatoptions can be used to set the initial parameters if the value of the fit formatoption is a callable function.

Note that the automatic estimation uses the boundaries of the param_bounds formatoption. This only works if the boundaries are given for each parameter and finite.

Possible types

‘auto’: The initial parameters are estimated automatically using the from scipy.optimize.differential_evolution() function
list of floats: The initial parameters
list of list of floats or ‘auto’: A combination of the above types where each corresponds to one data array

param_bounds

Parameter bounds for the function parameters

This formatoption can be used to specify the boundaries for the parameters. It only has an effect if the value of the fit formatoption is a callable function.

These bounds will also be used by the p0 formatoption to estimate the initial parameters.

Possible types

None: Use open boundaries
list of tuples with length 2: The boundaries for each of the parameters
list of tuples or None: A combination of the above types where each corresponds to one data array

plot

Choose how to visualize a 2-dimensional scalar data field

Possible types

None: Don’t make any plotting
‘mesh’: Use the matplotlib.pyplot.pcolormesh() function to make the plot or the matplotlib.pyplot.tripcolor() for an unstructered grid
‘poly’: Draw each polygon indivually. This method is used by default for unstructured grids. If there are no grid cell boundaries in the dataset, we will interpolate them
‘contourf’: Make a filled contour plot using the matplotlib.pyplot.contourf() function. The levels for the contour plot are controlled by the levels formatoption
‘contour’: Same a 'contourf', but does not make a filled contour plot, only lines.

post

Apply your own postprocessing script

This formatoption let’s you apply your own post processing script. Just enter the script as a string and it will be executed. The formatoption will be made available via the self variable

Possible types

None: Don’t do anything
str: The post processing script as string

Note

This formatoption uses the built-in exec() function to compile the script. Since this poses a security risk when loading psyplot projects, it is by default disabled through the Plotter.enable_post attribute. If you are sure that you can trust the script in this formatoption, set this attribute of the corresponding Plotter to True

Examples

Assume, you want to manually add the mean of the data to the title of the matplotlib axes. You can simply do this via

from psyplot.plotter import Plotter
from xarray import DataArray

plotter = Plotter(DataArray([1, 2, 3]))
# enable the post formatoption
plotter.enable_post = True
plotter.update(post="self.ax.set_title(str(self.data.mean()))")
plotter.ax.get_title()
"2.0"

By default, the post formatoption is only ran, when it is explicitly updated. However, you can use the post_timing formatoption, to run it automatically. E.g. for running it after every update of the plotter, you can set

plotter.update(post_timing="always")

See also

post_timing: Determine the timing of this formatoption

post_timing

Determine when to run the post formatoption

This formatoption determines, whether the post formatoption should be run never, after replot or after every update.

Possible types

‘never’: Never run post processing scripts
‘always’: Always run post processing scripts
‘replot’: Only run post processing scripts when the data changes or a replot is necessary

See also

post: The post processing formatoption

precision

Set the precision of the data

This formatoption can be used to specify the precision of the data which then will be the minimal bin width of the 2D histogram or the bandwith of the kernel size (if the density formatoption is set to 'kde')

Possible types

float: If 0, this formatoption has no effect at all. Otherwise it is assumed to be the precision of the data
str: One of {'scott' | 'silverman'}. This uses the statsmodels package to estimate the bandwidth of the data that is then used in the histogram or KDE plot

ticksize

Change the ticksize of the ticklabels

Possible types

dict: A dictionary with the keys 'minor' and (or) 'major' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is put into a dictionary with the key determined by the rcParams 'ticks.which' key (usually 'major'). The values in the dictionary can be one types below.
float: The absolute font size in points (e.g., 12)
string: Strings might be ‘xx-small’, ‘x-small’, ‘small’, ‘medium’, ‘large’, ‘x-large’, ‘xx-large’.

tickweight

Change the fontweight of the ticks

Possible types

dict: A dictionary with the keys 'minor' and (or) 'major' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is put into a dictionary with the key determined by the rcParams 'ticks.which' key (usually 'major'). The values in the dictionary can be one types below.
float: a float between 0 and 1000
string: Possible strings are one of ‘ultralight’, ‘light’, ‘normal’, ‘regular’, ‘book’, ‘medium’, ‘roman’, ‘semibold’, ‘demibold’, ‘demi’, ‘bold’, ‘heavy’, ‘extra bold’, ‘black’.

See also

ticksize, xtickprops, ytickprops

title

Show the title

Set the title of the plot. You can insert any meta key from the xarray.DataArray.attrs via a string like '%(key)s'. Furthermore there are some special cases:

Strings like '%Y', '%b', etc. will be replaced using the datetime.datetime.strftime() method as long as the data has a time coordinate and this can be converted to a datetime object.
'%(x)s', '%(y)s', '%(z)s', '%(t)s' will be replaced by the value of the x-, y-, z- or time coordinate (as long as this coordinate is one-dimensional in the data)
any attribute of one of the above coordinates is inserted via axis + key (e.g. the name of the x-coordinate can be inserted via '%(xname)s').
Labels defined in the psyplot.rcParams 'texts.labels' key are also replaced when enclosed by ‘{}’. The standard labels are
- tinfo: %H:%M
- dtinfo: %B %d, %Y. %H:%M
- dinfo: %B %d, %Y
- desc: %(long_name)s [%(units)s]
- sdesc: %(name)s [%(units)s]

Possible types

str: The title for the title() function.

Notes

This is the title of this specific subplot! For the title of the whole figure, see the figtitle formatoption.

See also

figtitle, titlesize, titleweight, titleprops

transpose

Switch x- and y-axes

By default, one-dimensional arrays have the dimension on the x-axis and two dimensional arrays have the first dimension on the y and the second on the x-axis. You can set this formatoption to True to change this behaviour

Possible types

bool: If True, axes are switched

xlabel

Set the x-axis label

Set the label for the x-axis. You can insert any meta key from the xarray.DataArray.attrs via a string like '%(key)s'. Furthermore there are some special cases:

Strings like '%Y', '%b', etc. will be replaced using the datetime.datetime.strftime() method as long as the data has a time coordinate and this can be converted to a datetime object.
'%(x)s', '%(y)s', '%(z)s', '%(t)s' will be replaced by the value of the x-, y-, z- or time coordinate (as long as this coordinate is one-dimensional in the data)
any attribute of one of the above coordinates is inserted via axis + key (e.g. the name of the x-coordinate can be inserted via '%(xname)s').
Labels defined in the psyplot.rcParams 'texts.labels' key are also replaced when enclosed by ‘{}’. The standard labels are
- tinfo: %H:%M
- dtinfo: %B %d, %Y. %H:%M
- dinfo: %B %d, %Y
- desc: %(long_name)s [%(units)s]
- sdesc: %(name)s [%(units)s]

Possible types

str: The text for the xlabel() function.

See also

xlabelsize, xlabelweight, xlabelprops

xlim

Set the x-axis limits

Possible types

None

To not change the current limits

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

See also

ylim

xrange

Specify the range of the histogram for the x-dimension

This formatoption specifies the minimum and maximum of the histogram in the x-dimension

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

Notes

This formatoption always acts on the coordinate, no matter what the value of the transpose formatoption is

See also

yrange

xrotation

Rotate the x-axis ticks

Possible types

float: The rotation angle in degrees

See also

yrotation

xticklabels

Modify the x-axis ticklabels

Possible types

dict: A dictionary with the keys 'minor' and (or) 'major' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is put into a dictionary with the key determined by the rcParams 'ticks.which' key (usually 'major'). The values in the dictionary can be one types below.
str: A formatstring like '%Y' for plotting the year (in the case that time is shown on the axis) or ‘%i’ for integers
array: An array of strings to use for the ticklabels

xtickprops

Specify the x-axis tick parameters

This formatoption can be used to make a detailed change of the ticks parameters on the x-axis.

Possible types

dict: A dictionary with the keys 'minor' and (or) 'major' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is put into a dictionary with the key determined by the rcParams 'ticks.which' key (usually 'major'). The values in the dictionary can be one types below.
dict: Items may be anything of the matplotlib.pyplot.tick_params() function

xticks

Modify the x-axis ticks

Possible types

dict

A dictionary with the keys 'minor' and (or) 'major' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is put into a dictionary with the key determined by the rcParams 'ticks.which' key (usually 'major'). The values in the dictionary can be one types below.

None

use the default ticks

int

for an integer i, only every i-th tick of the default ticks are used

numeric array

specifies the ticks manually

str or list [str, …]

A list of the below mentioned values of the mapping like [method, N, percmin, percmax, vmin, vmax], where only the first one is absolutely necessary

dict

Automatically determine the ticks corresponding to the data. The mapping can have the following keys, but only method is not optional.

N

An integer describing the number of boundaries (or ticks per power of ten, see log and symlog above)

percmin

The percentile to use for the minimum (by default, 0, i.e. the minimum of the array)

percmax

The percentile to use for the maximum (by default, 100, i.e. the maximum of the array)

vmin

The minimum to use (in which case it is not calculated from the specified method)

vmax

The maximum to use (in which case it is not calculated from the specified method)

method

A string that defines how minimum and maximum shall be set. This argument is not optional and can be one of the following:

data: plot the ticks exactly where the data is.
mid: plot the ticks in the middle of the data.
rounded: Sets the minimum and maximum of the ticks to the rounded data minimum or maximum. Ticks are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimal tick will always be lower or equal than the data minimum, the maximal tick will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the ticks are chose such that they are symmetric around zero
minmax: Uses the minimum as minimal tick and maximum as maximal tick
sym: Same as minmax but symmetric around zero
log: Use logarithmic bounds. In this case, the given number N determines the number of bounds per power of tenth (i.e. N == 2 results in something like 1.0, 5.0, 10.0, 50.0, etc., If this second number is None, then it will be chosen such that we have around 11 boundaries but at least one per power of ten.
symlog: The same as log but symmetric around 0. If the number N is None, then we have around 12 boundaries but at least one per power of ten
hour: draw ticks every hour
day: draw ticks every day
week: draw ticks every week
month, monthend, monthbegin: draw ticks in the middle, at the end or at the beginning of each month
year, yearend, yearbegin: draw ticks in the middle, at the end or at the beginning of each year

For data, mid, hour, day, week, month, etc., the optional second value can be an integer i determining that every i-th data point shall be used (by default, it is set to 1). For rounded, roundedsym, minmax and sym, the second value determines the total number of ticks (defaults to 11).

Examples

Plot 11 ticks over the whole data range:

>>> plotter.update(xticks='rounded')

Plot 7 ticks over the whole data range where the maximal and minimal tick matches the data maximum and minimum:

>>> plotter.update(xticks=['minmax', 7])

Plot ticks every year and minor ticks every month:

>>> plotter.update(xticks={'major': 'year', 'minor': 'month'})

ylabel

Set the y-axis label

Set the label for the y-axis. You can insert any meta key from the xarray.DataArray.attrs via a string like '%(key)s'. Furthermore there are some special cases:

Strings like '%Y', '%b', etc. will be replaced using the datetime.datetime.strftime() method as long as the data has a time coordinate and this can be converted to a datetime object.
'%(x)s', '%(y)s', '%(z)s', '%(t)s' will be replaced by the value of the x-, y-, z- or time coordinate (as long as this coordinate is one-dimensional in the data)
any attribute of one of the above coordinates is inserted via axis + key (e.g. the name of the x-coordinate can be inserted via '%(xname)s').
Labels defined in the psyplot.rcParams 'texts.labels' key are also replaced when enclosed by ‘{}’. The standard labels are
- tinfo: %H:%M
- dtinfo: %B %d, %Y. %H:%M
- dinfo: %B %d, %Y
- desc: %(long_name)s [%(units)s]
- sdesc: %(name)s [%(units)s]

Possible types

str: The text for the ylabel() function.

See also

ylabelsize, ylabelweight, ylabelprops

ylim

Set the y-axis limits

Possible types

None

To not change the current limits

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

See also

xlim

yrange

Specify the range of the histogram for the x-dimension

This formatoption specifies the minimum and maximum of the histogram in the x-dimension

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

Notes

This formatoption always acts on the DataArray, no matter what the value of the transpose formatoption is

See also

xrange

yrotation

Rotate the y-axis ticks

Possible types

float: The rotation angle in degrees

See also

xrotation

yticklabels

Modify the y-axis ticklabels

Possible types

dict: A dictionary with the keys 'minor' and (or) 'major' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is put into a dictionary with the key determined by the rcParams 'ticks.which' key (usually 'major'). The values in the dictionary can be one types below.
str: A formatstring like '%Y' for plotting the year (in the case that time is shown on the axis) or ‘%i’ for integers
array: An array of strings to use for the ticklabels

ytickprops

Specify the y-axis tick parameters

This formatoption can be used to make a detailed change of the ticks parameters of the y-axis.

Possible types

dict: A dictionary with the keys 'minor' and (or) 'major' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is put into a dictionary with the key determined by the rcParams 'ticks.which' key (usually 'major'). The values in the dictionary can be one types below.
dict: Items may be anything of the matplotlib.pyplot.tick_params() function

yticks

Modify the y-axis ticks

Possible types

dict

A dictionary with the keys 'minor' and (or) 'major' to specify which ticks are managed. If the given value is not a dictionary with those keys, it is put into a dictionary with the key determined by the rcParams 'ticks.which' key (usually 'major'). The values in the dictionary can be one types below.

None

use the default ticks

int

for an integer i, only every i-th tick of the default ticks are used

numeric array

specifies the ticks manually

str or list [str, …]

A list of the below mentioned values of the mapping like [method, N, percmin, percmax, vmin, vmax], where only the first one is absolutely necessary

dict

Automatically determine the ticks corresponding to the data. The mapping can have the following keys, but only method is not optional.

N

An integer describing the number of boundaries (or ticks per power of ten, see log and symlog above)

percmin

The percentile to use for the minimum (by default, 0, i.e. the minimum of the array)

percmax

The percentile to use for the maximum (by default, 100, i.e. the maximum of the array)

vmin

The minimum to use (in which case it is not calculated from the specified method)

vmax

The maximum to use (in which case it is not calculated from the specified method)

method

A string that defines how minimum and maximum shall be set. This argument is not optional and can be one of the following:

data: plot the ticks exactly where the data is.
mid: plot the ticks in the middle of the data.
rounded: Sets the minimum and maximum of the ticks to the rounded data minimum or maximum. Ticks are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimal tick will always be lower or equal than the data minimum, the maximal tick will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the ticks are chose such that they are symmetric around zero
minmax: Uses the minimum as minimal tick and maximum as maximal tick
sym: Same as minmax but symmetric around zero
log: Use logarithmic bounds. In this case, the given number N determines the number of bounds per power of tenth (i.e. N == 2 results in something like 1.0, 5.0, 10.0, 50.0, etc., If this second number is None, then it will be chosen such that we have around 11 boundaries but at least one per power of ten.
symlog: The same as log but symmetric around 0. If the number N is None, then we have around 12 boundaries but at least one per power of ten
hour: draw ticks every hour
day: draw ticks every day
week: draw ticks every week
month, monthend, monthbegin: draw ticks in the middle, at the end or at the beginning of each month
year, yearend, yearbegin: draw ticks in the middle, at the end or at the beginning of each year

For data, mid, hour, day, week, month, etc., the optional second value can be an integer i determining that every i-th data point shall be used (by default, it is set to 1). For rounded, roundedsym, minmax and sym, the second value determines the total number of ticks (defaults to 11).

xticks: for possible examples

class psy_reg.plotters.FitPointDensity(key, plotter=None, index_in_list=None, additional_children=[], additional_dependencies=[], **kwargs)

Bases: PointDensity

Attributes:

`bins`	bins Formatoption instance in the plotter
`children`	list of str.
`coord`	coord Formatoption instance in the plotter
`line_xlim`	line_xlim Formatoption instance in the plotter
`normed`	normed Formatoption instance in the plotter
`precision`	precision Formatoption instance in the plotter
`xrange`	xrange Formatoption instance in the plotter
`yrange`	yrange Formatoption instance in the plotter

property bins: bins Formatoption instance in the plotter

children = ['line_xlim']: list of str. List of formatoptions that have to be updated before this one is updated. Those formatoptions are only updated if they exist in the update parameters.

property coord: coord Formatoption instance in the plotter

property line_xlim: line_xlim Formatoption instance in the plotter

property normed: normed Formatoption instance in the plotter

property precision: precision Formatoption instance in the plotter

property xrange: xrange Formatoption instance in the plotter

property yrange: yrange Formatoption instance in the plotter

class psy_reg.plotters.FixPoint(key, plotter=None, index_in_list=None, additional_children=[], additional_dependencies=[], **kwargs)

Bases: Formatoption

Force the fit to go through a given point

Possible types

None: do not force the fit at all
float f: make a linear fit forced through (x, y) = (0, f)
tuple (x’, y’): make a linear fit forced through (x, y) = (x', y')

See also

fit

Attributes:

`connections`	list of str.
`fit`	fit Formatoption instance in the plotter
`group`	`str`.
`name`	`str`.
`priority`	`int`.

Methods:

update(value)

Method that is call to update the formatoption on the axes

connections = ['fit']: list of str. Connections to other formatoptions that are (different from dependencies and children) not important for the update process

property fit: fit Formatoption instance in the plotter

group = 'regression': str. Key of the group name in groups of this formatoption keyword

name = 'Force the fit to go through a given point': str. A bit more verbose name than the formatoption key to be included in the gui. If None, the key is used in the gui

priority = 30

int. Priority value of the the formatoption determining when the formatoption is updated.

10: at the end (for labels, etc.)
20: before the plotting (e.g. for colormaps, etc.)
30: before loading the data (e.g. for lonlatbox)

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

class psy_reg.plotters.IdealLine(key, plotter=None, index_in_list=None, additional_children=[], additional_dependencies=[], **kwargs)

Bases: Formatoption

Draw an ideal line of the fit

Possible types

None: Don’t draw an ideal line
list of floats: The parameters for the line. If the fit formatoption is in 'robust' or 'fit', then the first value corresponds to the interception, the second to the slope. Otherwise the list corrensponds to the parameters as used in the fit function of the lines
list of list of floats: The same as above but with the specification for each array

See also

id_color

Attributes:

`dependencies`	list of str.
`fit`	fit Formatoption instance in the plotter
`group`	`str`.
`id_color`	id_color Formatoption instance in the plotter
`plot`	plot Formatoption instance in the plotter

Methods:

`initialize_plot`(args, *kwargs)	Method that is called when the plot is made the first time
`remove`()	Method to remove the effects of this formatoption
`update`(value)	Method that is call to update the formatoption on the axes

dependencies = ['fit', 'id_color', 'plot']: list of str. List of formatoptions that force an update of this formatoption if they are updated.

property fit: fit Formatoption instance in the plotter

group = 'regression': str. Key of the group name in groups of this formatoption keyword

property id_color: id_color Formatoption instance in the plotter

initialize_plot(*args, **kwargs)

Method that is called when the plot is made the first time

Parameters:: value – The value to use for the initialization

property plot: plot Formatoption instance in the plotter

remove()

Method to remove the effects of this formatoption

This method is called when the axes is cleared due to a formatoption with requires_clearing set to True. You don’t necessarily have to implement this formatoption if your plot results are removed by the usual matplotlib.axes.Axes.clear() method.

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

class psy_reg.plotters.IdealLineColor(*args, **kwargs)

Bases: LineColors

The colors of the ideal lines

Possible types

None: Let it be determined by the color cycle of the color formatoption
iterable: (e.g. list) to specify the colors manually

str

Strings may be any valid colormap name suitable for the matplotlib.cm.get_cmap() function or one of the color lists defined in the ‘colors.cmaps’ key of the psyplot.rcParams dictionary (including their reversed color maps given via the ‘_r’ extension).

matplotlib.colors.ColorMap: to automatically choose the colors according to the number of lines, etc. from the given colormap

See also

ideal

Attributes:

`color`	color Formatoption instance in the plotter
`dependencies`	list of str.
`ideal`	ideal Formatoption instance in the plotter
`parents`	list of str.
`priority`	`int`.

Methods:

update(value)

Method that is call to update the formatoption on the axes

property color: color Formatoption instance in the plotter

dependencies = ['color']: list of str. List of formatoptions that force an update of this formatoption if they are updated.

property ideal: ideal Formatoption instance in the plotter

parents = ['ideal']: list of str. List of formatoptions that, if included in the update, prevent the update of this formatoption.

priority = 10

int. Priority value of the the formatoption determining when the formatoption is updated.

10: at the end (for labels, etc.)
20: before the plotting (e.g. for colormaps, etc.)
30: before loading the data (e.g. for lonlatbox)

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

class psy_reg.plotters.InitialParameters(key, plotter=None, index_in_list=None, additional_children=[], additional_dependencies=[], **kwargs)

Bases: Formatoption

Initial parameters for the scipy.optimize.curve_fit() function

This formatoptions can be used to set the initial parameters if the value of the fit formatoption is a callable function.

Note that the automatic estimation uses the boundaries of the param_bounds formatoption. This only works if the boundaries are given for each parameter and finite.

Possible types

‘auto’: The initial parameters are estimated automatically using the from scipy.optimize.differential_evolution() function
list of floats: The initial parameters
list of list of floats or ‘auto’: A combination of the above types where each corresponds to one data array

Attributes:

`connections`	list of str.
`data_dependent`	`bool` or a callable.
`dependencies`	list of str.
`fit`	fit Formatoption instance in the plotter
`group`	`str`.
`name`	`str`.
`param_bounds`	param_bounds Formatoption instance in the plotter
`priority`	`int`.

Methods:

`p0`([i])
`update`(value)	Method that is call to update the formatoption on the axes

connections = ['fit']: list of str. Connections to other formatoptions that are (different from dependencies and children) not important for the update process

data_dependent = True: bool or a callable. This attribute indicates whether this Formatoption depends on the data and should be updated if the data changes. If it is a callable, it must accept one argument: the new data. (Note: This is automatically set to True for plot formatoptions)

dependencies = ['param_bounds']: list of str. List of formatoptions that force an update of this formatoption if they are updated.

property fit: fit Formatoption instance in the plotter

group = 'regression': str. Key of the group name in groups of this formatoption keyword

name = 'Initial parameter values for the fit': str. A bit more verbose name than the formatoption key to be included in the gui. If None, the key is used in the gui

p0(i=None)

property param_bounds: param_bounds Formatoption instance in the plotter

priority = 30

int. Priority value of the the formatoption determining when the formatoption is updated.

10: at the end (for labels, etc.)
20: before the plotting (e.g. for colormaps, etc.)
30: before loading the data (e.g. for lonlatbox)

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

class psy_reg.plotters.LinRegPlotter(data=None, ax=None, auto_update=None, project=None, draw=False, make_plot=True, clear=False, enable_post=False, **kwargs)

Bases: LinePlotter

A plotter to visualize the fit on the data

The most important formatoptions are the fit and ci formatoption. Otherwise this plotter behaves like the psyplot.plotter.simple.LinePlotter plotter class

Attributes:

`allowed_vars`	The number variables that one data array visualized by this plotter might have.
`ci`	Draw a confidence interval
`fit`	Choose the linear fitting method
`fix`	Force the fit to go through a given point
`id_color`	The colors of the ideal lines
`ideal`	Draw an ideal line of the fit
`line_xlim`	Specify how wide the range for the plot should be
`nboot`	Set the number of bootstrap resamples for the confidence interval
`p0`	Initial parameters for the `scipy.optimize.curve_fit()` function
`param_bounds`	Parameter bounds for the function parameters
`transpose`	Switch x- and y-axes
`xrange`	Specify the range for the fit to use for the x-dimension
`yrange`	Specify the range for the fit to use for the y-dimension

allowed_vars = 1: The number variables that one data array visualized by this plotter might have. We allow up to 3 variableswhere the second and third variable might be the errors (see the error formatoption)

ci

Draw a confidence interval

Size of the confidence interval for the regression estimate. This will be drawn using translucent bands around the regression line. The confidence interval is estimated using a bootstrap; for large datasets, it may be advisable to avoid that computation by setting this parameter to None.

Possible types

None: Do not draw and calculate a confidence interval
float: A quantile between 0 and 100

fit

Choose the linear fitting method

This formatoption consists makes a linear fit of the data

Possible types

‘fit’ or ‘linear’: make a linear fit
‘robust’: make a robust linear fit
‘poly<deg>’: Make a polynomial fit of the order '<deg>'
function: A callable function that takes an x-array and a y-array as input and can be used for the scipy.optimize.curve_fit() function
any object with a fit and predict method: A model that with a fit signature such as model.fit(x, y).predict(x)
None: make no fit

Notes

You can access the intercept, slope and rsquared by the correponding attribute. E.g.:

>>> plotter.update(
...     legendlabels="%(intercept)s + %(slope)s * x, "
...     "$R^2$=%(rsquared)s"
... )

See also

fix

fix

Force the fit to go through a given point

Possible types

None: do not force the fit at all
float f: make a linear fit forced through (x, y) = (0, f)
tuple (x’, y’): make a linear fit forced through (x, y) = (x', y')

See also

fit

id_color

The colors of the ideal lines

Possible types

None: Let it be determined by the color cycle of the color formatoption
iterable: (e.g. list) to specify the colors manually

str

Strings may be any valid colormap name suitable for the matplotlib.cm.get_cmap() function or one of the color lists defined in the ‘colors.cmaps’ key of the psyplot.rcParams dictionary (including their reversed color maps given via the ‘_r’ extension).

matplotlib.colors.ColorMap: to automatically choose the colors according to the number of lines, etc. from the given colormap

See also

ideal

ideal

Draw an ideal line of the fit

Possible types

None: Don’t draw an ideal line
list of floats: The parameters for the line. If the fit formatoption is in 'robust' or 'fit', then the first value corresponds to the interception, the second to the slope. Otherwise the list corrensponds to the parameters as used in the fit function of the lines
list of list of floats: The same as above but with the specification for each array

See also

id_color

line_xlim

Specify how wide the range for the plot should be

This formatoption specifies the range of the line to use

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

See also

xrange

nboot

Set the number of bootstrap resamples for the confidence interval

Parameters:: int – Number of bootstrap resamples used to estimate the ci. The default value attempts to balance time and stability; you may want to increase this value for “final” versions of plots.

See also

ci

p0

Initial parameters for the scipy.optimize.curve_fit() function

This formatoptions can be used to set the initial parameters if the value of the fit formatoption is a callable function.

Note that the automatic estimation uses the boundaries of the param_bounds formatoption. This only works if the boundaries are given for each parameter and finite.

Possible types

‘auto’: The initial parameters are estimated automatically using the from scipy.optimize.differential_evolution() function
list of floats: The initial parameters
list of list of floats or ‘auto’: A combination of the above types where each corresponds to one data array

param_bounds

Parameter bounds for the function parameters

This formatoption can be used to specify the boundaries for the parameters. It only has an effect if the value of the fit formatoption is a callable function.

These bounds will also be used by the p0 formatoption to estimate the initial parameters.

Possible types

None: Use open boundaries
list of tuples with length 2: The boundaries for each of the parameters
list of tuples or None: A combination of the above types where each corresponds to one data array

transpose

Switch x- and y-axes

By default, one-dimensional arrays have the dimension on the x-axis and two dimensional arrays have the first dimension on the y and the second on the x-axis. You can set this formatoption to True to change this behaviour

Possible types

bool: If True, axes are switched

xrange

Specify the range for the fit to use for the x-dimension

This formatoption specifies the minimum and maximum of the fit in the x-dimension

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

Notes

This formatoption always acts on the coordinate, no matter what the value of the transpose formatoption is

See also

yrange, line_xlim

yrange

Specify the range for the fit to use for the y-dimension

This formatoption specifies the minimum and maximum of the fit in the y-dimension

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

Notes

This formatoption always acts on the coordinate, no matter what the value of the transpose formatoption is

See also

xrange

class psy_reg.plotters.LinRegTranspose(*args, **kwargs)

Bases: Transpose

Switch x- and y-axes

By default, one-dimensional arrays have the dimension on the x-axis and two dimensional arrays have the first dimension on the y and the second on the x-axis. You can set this formatoption to True to change this behaviour

Possible types

bool: If True, axes are switched

Attributes:

priority

int.

priority = 30

int. Priority value of the the formatoption determining when the formatoption is updated.

10: at the end (for labels, etc.)
20: before the plotting (e.g. for colormaps, etc.)
30: before loading the data (e.g. for lonlatbox)

class psy_reg.plotters.LinearRegressionFit(*args, **kwargs)

Bases: Formatoption

Choose the linear fitting method

This formatoption consists makes a linear fit of the data

Possible types

‘fit’ or ‘linear’: make a linear fit
‘robust’: make a robust linear fit
‘poly<deg>’: Make a polynomial fit of the order '<deg>'
function: A callable function that takes an x-array and a y-array as input and can be used for the scipy.optimize.curve_fit() function
any object with a fit and predict method: A model that with a fit signature such as model.fit(x, y).predict(x)
None: make no fit

Notes

You can access the intercept, slope and rsquared by the correponding attribute. E.g.:

>>> plotter.update(
...     legendlabels="%(intercept)s + %(slope)s * x, "
...     "$R^2$=%(rsquared)s"
... )

See also

fix

Attributes:

`coord`	coord Formatoption instance in the plotter
`data_dependent`	`bool` or a callable.
`dependencies`	list of str.
`fix`	fix Formatoption instance in the plotter
`func_args`	The arguments for the fit function if the `method` is 'curve_fit'
`group`	`str`.
`line_xlim`	line_xlim Formatoption instance in the plotter
`name`	`str`.
`p0`	p0 Formatoption instance in the plotter
`param_bounds`	param_bounds Formatoption instance in the plotter
`priority`	`int`.
`transpose`	transpose Formatoption instance in the plotter
`xrange`	xrange Formatoption instance in the plotter
`yrange`	yrange Formatoption instance in the plotter

Methods:

`get_kwargs`(i)	Get the fitting kwargs for the line at index i
`get_xline`([i])	Get the x-data for the best fit line
`get_xy`(i, da)
`make_fit`(i, x, y[, x_line])
`set_method`(i)
`update`(value)	Method that is call to update the formatoption on the axes

property coord: coord Formatoption instance in the plotter

data_dependent = True: bool or a callable. This attribute indicates whether this Formatoption depends on the data and should be updated if the data changes. If it is a callable, it must accept one argument: the new data. (Note: This is automatically set to True for plot formatoptions)

dependencies = ['transpose', 'fix', 'xrange', 'yrange', 'coord', 'line_xlim', 'p0', 'param_bounds']: list of str. List of formatoptions that force an update of this formatoption if they are updated.

property fix: fix Formatoption instance in the plotter

property func_args: The arguments for the fit function if the method is ‘curve_fit’

get_kwargs(i): Get the fitting kwargs for the line at index i

get_xline(i=0): Get the x-data for the best fit line

get_xy(i, da)

group = 'regression': str. Key of the group name in groups of this formatoption keyword

property line_xlim: line_xlim Formatoption instance in the plotter

make_fit(i, x, y, x_line=None, **kwargs)

name = 'Change the fit method': str. A bit more verbose name than the formatoption key to be included in the gui. If None, the key is used in the gui

property p0: p0 Formatoption instance in the plotter

property param_bounds: param_bounds Formatoption instance in the plotter

priority = 30

int. Priority value of the the formatoption determining when the formatoption is updated.

10: at the end (for labels, etc.)
20: before the plotting (e.g. for colormaps, etc.)
30: before loading the data (e.g. for lonlatbox)

set_method(i)

property transpose: transpose Formatoption instance in the plotter

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

property xrange: xrange Formatoption instance in the plotter

property yrange: yrange Formatoption instance in the plotter

class psy_reg.plotters.LinearRegressionFitCombined(*args, **kwargs)

Bases: LinearRegressionFit

Choose the linear fitting method

This formatoption consists makes a linear fit of the data

Possible types

‘fit’ or ‘linear’: make a linear fit
‘robust’: make a robust linear fit
‘poly<deg>’: Make a polynomial fit of the order '<deg>'
function: A callable function that takes an x-array and a y-array as input and can be used for the scipy.optimize.curve_fit() function
any object with a fit and predict method: A model that with a fit signature such as model.fit(x, y).predict(x)
None: make no fit

Notes

You can access the intercept, slope and rsquared by the correponding attribute. E.g.:

>>> plotter.update(
...     legendlabels="%(intercept)s + %(slope)s * x, "
...     "$R^2$=%(rsquared)s"
... )

See also

fix

Attributes:

`coord`	coord Formatoption instance in the plotter
`fix`	fix Formatoption instance in the plotter
`line_xlim`	line_xlim Formatoption instance in the plotter
`p0`	p0 Formatoption instance in the plotter
`param_bounds`	param_bounds Formatoption instance in the plotter
`transpose`	transpose Formatoption instance in the plotter
`xrange`	xrange Formatoption instance in the plotter
`yrange`	yrange Formatoption instance in the plotter

Methods:

set_data(data[, i])

Reimplemented to change the arr_name attribute of the given array

property coord: coord Formatoption instance in the plotter

property fix: fix Formatoption instance in the plotter

property line_xlim: line_xlim Formatoption instance in the plotter

property p0: p0 Formatoption instance in the plotter

property param_bounds: param_bounds Formatoption instance in the plotter

set_data(data, i=None): Reimplemented to change the arr_name attribute of the given array

property transpose: transpose Formatoption instance in the plotter

property xrange: xrange Formatoption instance in the plotter

property yrange: yrange Formatoption instance in the plotter

class psy_reg.plotters.NBoot(key, plotter=None, index_in_list=None, additional_children=[], additional_dependencies=[], **kwargs)

Bases: Formatoption

Set the number of bootstrap resamples for the confidence interval

Parameters:: int – Number of bootstrap resamples used to estimate the ci. The default value attempts to balance time and stability; you may want to increase this value for “final” versions of plots.

See also

ci

Attributes:

`group`	`str`.
`name`	`str`.
`priority`	`int`.

Methods:

update(value)

Does nothing.

group = 'regression': str. Key of the group name in groups of this formatoption keyword

name = 'Set the bootstrapping number to calculate the confidence interval': str. A bit more verbose name than the formatoption key to be included in the gui. If None, the key is used in the gui

priority = 30

int. Priority value of the the formatoption determining when the formatoption is updated.

10: at the end (for labels, etc.)
20: before the plotting (e.g. for colormaps, etc.)
30: before loading the data (e.g. for lonlatbox)

update(value): Does nothing. The work is done by the Ci formatoption

class psy_reg.plotters.ParameterBounds(key, plotter=None, index_in_list=None, additional_children=[], additional_dependencies=[], **kwargs)

Bases: Formatoption

Parameter bounds for the function parameters

This formatoption can be used to specify the boundaries for the parameters. It only has an effect if the value of the fit formatoption is a callable function.

These bounds will also be used by the p0 formatoption to estimate the initial parameters.

Possible types

None: Use open boundaries
list of tuples with length 2: The boundaries for each of the parameters
list of tuples or None: A combination of the above types where each corresponds to one data array

Methods:

update(value)

Method that is call to update the formatoption on the axes

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

class psy_reg.plotters.XFitRange(*args, **kwargs)

Bases: Hist2DXRange

Specify the range for the fit to use for the x-dimension

This formatoption specifies the minimum and maximum of the fit in the x-dimension

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

Notes

This formatoption always acts on the coordinate, no matter what the value of the transpose formatoption is

See also

yrange, line_xlim

Attributes:

`coord`	coord Formatoption instance in the plotter
`group`	`str`.
`plot`	plot Formatoption instance in the plotter
`range`	The range for each of the curves
`transpose`	transpose Formatoption instance in the plotter

Methods:

`set_limit`(*args)	The method to set the minimum and maximum limit
`update`(value)	Method that is call to update the formatoption on the axes

property coord: coord Formatoption instance in the plotter

group = 'regression': str. Key of the group name in groups of this formatoption keyword

property plot: plot Formatoption instance in the plotter

property range: The range for each of the curves

set_limit(*args)

The method to set the minimum and maximum limit

Parameters:

min_val (float) – The value for the lower limit
max_val (float) – The value for the upper limit

property transpose: transpose Formatoption instance in the plotter

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

class psy_reg.plotters.XLineRange(*args, **kwargs)

Bases: XFitRange

Specify how wide the range for the plot should be

This formatoption specifies the range of the line to use

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

See also

xrange

Attributes:

`coord`	coord Formatoption instance in the plotter
`plot`	plot Formatoption instance in the plotter
`transpose`	transpose Formatoption instance in the plotter

property coord: coord Formatoption instance in the plotter

property plot: plot Formatoption instance in the plotter

property transpose: transpose Formatoption instance in the plotter

class psy_reg.plotters.YFitRange(*args, **kwargs)

Bases: Hist2DYRange

Specify the range for the fit to use for the y-dimension

This formatoption specifies the minimum and maximum of the fit in the y-dimension

Possible types

str or list [str, str] or [[str, float], [str, float]]

Automatically determine the ticks corresponding to the data. The given string determines how the limits are calculated. The float determines the percentile to use A string can be one of the following:

rounded: Sets the minimum and maximum of the limits to the rounded data minimum or maximum. Limits are rounded to the next 0.5 value with to the difference between data max- and minimum. The minimum will always be lower or equal than the data minimum, the maximum will always be higher or equal than the data maximum.
roundedsym: Same as rounded above but the limits are chosen such that they are symmetric around zero
minmax: Uses the minimum and maximum
sym: Same as minmax but symmetric around zero

tuple (xmin, xmax)

xmin is the smaller value, xmax the larger. Any of those values can be None or one of the strings (or lists) above to use the corresponding value here

Notes

This formatoption always acts on the coordinate, no matter what the value of the transpose formatoption is

See also

xrange

Attributes:

`coord`	coord Formatoption instance in the plotter
`group`	`str`.
`plot`	plot Formatoption instance in the plotter
`range`	The range for each of the curves
`transpose`	transpose Formatoption instance in the plotter

Methods:

`set_limit`(*args)	The method to set the minimum and maximum limit
`update`(value)	Method that is call to update the formatoption on the axes

property coord: coord Formatoption instance in the plotter

group = 'regression': str. Key of the group name in groups of this formatoption keyword

property plot: plot Formatoption instance in the plotter

property range: The range for each of the curves

set_limit(*args)

The method to set the minimum and maximum limit

Parameters:

min_val (float) – The value for the lower limit
max_val (float) – The value for the upper limit

property transpose: transpose Formatoption instance in the plotter

update(value)

Method that is call to update the formatoption on the axes

Parameters:: value – Value to update

psy_reg.plotters.bootstrap(x, y, func, n_boot, random_seed=None, **kwargs)

Simple bootstrap algorithm used to estimate the confidence interval

This function is motivated by seaborns bootstrap algorithm seaborn.algorithms.bootstrap()

psy_reg.plotters.calc_ci(a, which=95, axis=None): Return a quantile range from an array of values.