Add logic for running a decision making loop #365

Thomas-Christie · 2023-08-25T13:08:35Z

Type of changes

Checklist

I've formatted the new code by running poetry run pre-commit run --all-files --show-diff-on-failure before committing.
I've added tests for new code.
I've added docstrings for the new code.

Description

Add logic for running a decision making loop

Added an AbstractDecisionMaker abstract base class which can be
implemented in order to define a decision making loop.

A concrete implementation, in the form of a DecisionMaker class,
has been added. At its heart it has two core methods:

ask which is used to get a point to be queried next.
tell which is used to tell the DecisionMaker about new
observations. In a typical decision making setup this will result in
the datsets and posteriors being updated.

In addition to this, in the DecisionMaker a run method is provided,
which will automatically run the decision making loop for n steps. After
the ask step, the functions in the post_ask list will be executed,
taking as arguments the decision maker and the point chosen to be
queried next. Similarly, after the tell step, the functions in
the post_tell list are executed, taking the decision maker as the sole
argument.

Also removed the check for type-checking import warnings (TCH) from Ruff in pyproject.toml.

Issue Number: N/A

Added an `AbstractDecisionMaker` abstract base class which can be implemented in order to define a decision making loop. A concrete implementation, in the form of a `DecisionMaker` class, has been added. At its heart it has two core methods: 1. `ask` which is used to get a point to be queried next. 2. `tell` which is used to tell the `DecisionMaker` about new observations. In a typical decision making setup this will result in the datsets and posteriors being updated. In addition to this, in the `DecisionMaker` a `run` method is provided, which will automatically run the decision making loop for n steps. After the `ask` step, the functions in the `post_ask` list will be executed, taking as arguments the decision maker and the point chosen to be queried next. Similarly, after the `tell` step, the functions in the `post_tell` list are executed, taking the decision maker as the sole argument.

gpjax/decision_making/decision_maker.py

henrymoss

Looks great. Just a couple Qs about organisation here

Implemented minor CR feedback and renamed acquisition functions to utility functions to reflect the fact that the package will be used for general Bayesian decision making and not just Bayesian optimisation.

Thomas-Christie requested review from thomaspinder and henrymoss August 25, 2023 13:23

thomaspinder approved these changes Sep 1, 2023

View reviewed changes