follows_reference¶

typhoon.test.signals.follows_reference(signal, ref_signal, tol, during=None, strictness=1, time_tol=0, report_plot=None)¶

Checks if signal follows a reference signal within given tolerances.

Note

The reference signal needs not to be on the same sampling rate of the signal to be compared, as they will be normalized in a single dataframe using pandas interpolate https://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.interpolate.html function.

Parameters:

signal (pandas.Series) – Signal to be tested for.
ref_signal (pandas.Series) – Reference signal the tested signal should follow.
tol (float) – Tolerance around which the signal can stay with respect to reference signal when determining if analyis result is True or False.
during (tuple) – Time period (as a range) to be considered for analysis.
strictness (float) – Number between 0.0 and 1.0 that determines percentage of time signal should be inside the defined range for test to pass.
time_tol (float or timedelta) – Time tolerance - argument which allows that signal is leading or lagging up to specified time in seconds, compared to created reference.
report_plot (dict) –
Dictionary which overrides default allure report plot attachment behaviour. It also overrides behaviour specified for whole test run with command line arguments --analysis-plot-type and --analysis-plot-on-fail-only. Dictionary has two keys to be specified:
1. type: specifies which kind of allure plot should be used. Valid values:
  static - only matplotlib plot attached as .png picture is attached
  
  interactive - only interactive html plot created with bokeh library is attached. Advantage of this plot it has options to zoom in/zoom out. Disadvantage is that it consumes significantly more memory.
  
  none - none of the plots will be attached
  
  all - all plots will be added. Currently supported ones are matplotlib plot(static) and bokeh plot(interactive)
2. when: specifies when to add plots that are specified with previous key to report. Available options:
  always - always adds specified plots
  
  on-fail - adds plots only if comparison of reference and measured signal fails. This is good way to decrease size of allure plots.
Note

If report_plot argument is not provided, and command line arguments --analysis-plot-type and analysis-plot-on-fail-only are not specified, default behaviour is static plot, attached always. If command line arguments are specified, they define new default behaviour for whole test run.

Returns:

result – Result of the analysis.

Return type:

AnalysisResult

Notes

The reference signal needs not to be on the same sampling rate of the signal to be compared, as they will be normalized in a single dataframe using pandas’ interpolate <https://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.interpolate.html> function.

Examples

Considering signal as part of a captured dataframe

>>> signal = capture["channel"]

Implementing a typhoon.test.signals.is_constant function, comparing if the signal is always around 10 ± 0.5

>>> import pandas as pd
>>> from typhoon.test.signals import follows_reference
>>>
>>> ref = 10
>>> tol = .5
>>>
>>> ref_signal = pd.Series({
>>>     signal.index[0]: ref,
>>>     signal.index[-1]: ref
>>> })
>>>
>>> result = follows_reference(signal, ref_signal, tol)

Implementing a typhoon.test.signals.is_ramp function with a slope of one unit per second (slope = 1)

>>> from typhoon.types.timedelta import Timedelta as td
>>>
>>> dt = signal.index[-1] - signal.index[0]
>>>
>>> initial_value = signal.values[0]
>>> # Equivalent to dt * (slope/dtbase) e.g. slope with 100 units per microsecond y = 10 + (1us * 100/1us) = 110
>>> vf = initial_value + dt/td("1s")*slope
>>>
>>> # Build signal with two points, providing ramp
>>> ref_signal = pd.Series({
>>>     signal.index[0]:initial_value,
>>>     signal.index[-1]:vf
>>> })
>>> result = follows_reference(signal, ref_signal, tol=0.5)

follows_reference¶

Previous topic

Next topic