Fits and Plots

pswalker uses callbacks to monitor live data, create fits and pipe this back into scanning logic in order to reach desired positions. These all build off of LiveBuild in order to be used with the fitwalk(). The most prominent example of this in the repository is walk_to_pixel()

LiveBuild

class pswalker.callbacks.LiveBuild(model, y, independent_vars, init_guess=None, update_every=1, filters=None, drop_missing=True, average=1)

Bases: bluesky.callbacks.fitting.LiveFit

Base class for live model building in Skywalker

Parameters:
  • model (lmfit.Model) –
  • y (string) – Key of dependent variable
  • indpendent_vars (dict) – Map independent variables names to keys in the event document stream
  • init_guess (dict, optional) – Initial guesses for other values if expected
  • update_every (int or None, optional) – Update rate of the model. If set to None, the model will only be computed at the end of the run. By default, this is set to 1 i.e update on every new event
backsolve(target, **kwargs)

Use the most recent fit to find the independent variables that create the requested dependent variable

..note:

For multivariable functions the user may have to specify which
variable to solve for, and which to keep fixed
eval(*args, **kwargs)

Estimate a point based on the current fit of the model. Reimplemented by subclasses

field_names

Name of all the keys associated with the fit

install_filters(filters)

Install additional filters

Parameters:filters (dict) – Filters are provided in a dictionary of key / callable pairs that take a single input from the data stream and return a boolean value.
name

Name of the model

class pswalker.callbacks.LinearFit(y, x, init_guess=None, update_every=1, name=None, average=1)

Bases: pswalker.callbacks.LiveBuild

Model to fit a linear relationship between a single variable axis and a depended variable

Parameters:
  • y (str) – Keyword in the event document that reports the dependent variable
  • x (str) – Keyword in the event document that reports the independent variable
  • init_guess (dict, optional) – Initialization guess for the linear fit, available keys are slope and intercept
  • name (optional , str) – Name for the contained model. When None (default) the name is the same as the model function
  • update_every (int or None, optional) – Update rate of the model. If set to None, the model will only be computed at the end of the run. By default, this is set to 1 i.e update on every new event
backsolve(target, **kwargs)

Find the x position that solves the reaches the given target

Parameters:target (float) – Desired y value
Returns:x – Variable name and floating value
Return type:dict
eval(**kwargs)

Evaluate the predicted outcome based on the most recent fit of the given information.

Parameters:
  • x (float or int, optional) – Independent variable to evaluate linear model
  • kwargs – The value for the indepenedent variable can also be given as the field name in the event document
Returns:

estimate – Y value as determined by current linear fit
Return type:

float

Plots

class pswalker.callbacks.LivePlotWithGoal(y, x=None, *, goal=0.0, tolerance=0.0, averages=None, **kwargs)

Bases: bluesky.callbacks.mpl_plotting.LivePlot

Build a function that updates a plot from a stream of Events.

Parameters:
  • y (str) – the name of a data field in an Event
  • x (str, optional) – the name of a data field in an Event, or ‘seq_num’ or ‘time’ If None, use the Event’s sequence number. Special case: If the Event’s data includes a key named ‘seq_num’ or ‘time’, that takes precedence over the standard ‘seq_num’ and ‘time’ recorded in every Event.
  • goal (float) – the target pixel
  • tolerance (float, optional) – the tolerance for the pixel
  • averages (float, optional) – The number of images to average. If None is specified, every point is rendered as they come, otherwise the graph will update every `averages` points.
  • legend_keys (list, optional) – The list of keys to extract from the RunStart document and format in the legend of the plot. The legend will always show the scan_id followed by a colon (“1: “). Each
  • xlim (tuple, optional) – passed to Axes.set_xlim
  • ylim (tuple, optional) – passed to Axes.set_ylim
  • ax (Axes, optional) – matplotib Axes; if none specified, new figure and axes are made.
  • fig (Figure, optional) – deprecated: use ax instead
  • epoch ({'run', 'unix'}, optional) – If ‘run’ t=0 is the time recorded in the RunStart document. If ‘unix’, t=0 is 1 Jan 1970 (“the UNIX epoch”). Default is ‘run’.
  • kwargs –
  • additional keyword arguments are passed through to Axes.plot. (All) –

Notes

If your figure blocks the main thread when you are trying to scan with this callback, call plt.ion() in your IPython session.

Examples

>>> my_plotter = LivePlotWithGoals('det', 'motor', goal=10.0, tolerance=1.5, averages=None, legend_keys=['sample'])
>>> RE(my_scan, my_plotter)