AI Overview

The AI system consists of an AI class that looks for new requests and results, and uses these to update an internal recommendation ssytem. It is built to be run from the command line as a module. The run command is python -m ai.ai from the root directory. Type python -m ai.ai -h to see options.

While the AI is running, it checks for new requests and results every 5 seconds. When the user clicks on the AI button, the AI receives a request and returns n_recs recommendations. Each recommendation consists of an ML method with associated parameter values, and a score estimate. (Read more in the recommender folder README.)

Within the AI class is self.rec, which is the heart of the recommender system. Various recommenders are available in the recommender/ folder. Each recommender implements an update() method that is called when new results are found, and a recommend() method that is called when new AI requests are receieved.

The LabAPI class is a helper class that encapsulates all calls to the server.