Using the Results on the Host
Once the module returns a full distribution, the signaloid package from the
signaloid-python repository turns it into something you can
read, plot, and reuse. Both recipes on this page start from a parsed distribution, as
produced by
Return the Full Distribution as Ux Binary Data.
from signaloid.distributional.distributional import DistributionalValue
dist = DistributionalValue.parse(payload)
Install the package with
python -m pip install git+https://github.com/signaloid/signaloid-python. The same
parse() call also accepts a Ux String, so everything below works equally on distributions
read from a file or captured from other Signaloid tools.
Print the Ux String and the particle value
The Ux String is the portable text form of a distribution, and the particle value is the point result the computation would have produced on conventional hardware. Print both straight from the parsed object.
print(str(dist))
print(dist.particle_value)
Save the printed Ux String anywhere you keep text. Parsing it back with
DistributionalValue.parse() reproduces the distribution, which makes the Ux String a
simple storage and interchange format. See
Uncertainty Hexadecimal (“Ux”) Data Format.
Plot the distribution
The plotting module renders a DistributionalValue as a histogram with variable bin
widths.
from signaloid.distributional_information_plotting.plot_histogram_dirac_deltas import PlotData
from signaloid.distributional_information_plotting.plot_wrapper import plot
plot(PlotData(dist=dist))
By default plot() opens a window with the figure. Pass save=True together with
path="result.png" to write the figure to a file instead, which suits host
applications that run unattended.
For more ways to work with Ux data on the host, including the TypeScript tools, see
Tools and Utilities.
Next steps
- Sending Inputs to the Module, feeding a stored or freshly built distribution back into the next computation.
- Debugging and Logging, watching the device application while it produces these results.
- Examples and Demos, complete applications that print and plot module results.