optbayesexpt API

  • OptBayesExpt is the core class that performs Bayesian inference and selects measurement settings. Typically, it is the only class that a user will need to interact with directly.

  • OptBayesExptNoiseParameter is similar to OptBayesExpt but is designed for cases where the measurement uncertainty is a parameter to be estimated.

  • ParticlePDF is inherited by OptBayesExpt to handle the duties of a probability distribution functions.

  • OBE_Server class provides communication with other processes through a mini-language of label-value commands.

  • Socket class is inherited by Server to handle TCP connections and message encoding/decoding.

  • The obe_utils.py file provides

    • A MeasurementSimulator class that uses “true value” parameters and added noise to simulate experimental outputs.

    • For post-processing, a trace_sort() function sorts measurement data by measurement setting and combines all measurements with settings in common.

    • A differential_entropy() function to calculate information entropy from samples of a distribution.