Creating an API process for inference

The code required for this section is in the pystock-inference-api folder. The MLflow infrastructure is provided in the Docker image accompanying the code as shown in the following figure:

Figure 9.2 – The structure of the API job

Setting up an API system is quite easy by relying on the MLflow built-in REST API environment. We will rely on the artifact store on the local filesystem to test the APIs.

With the following set of commands, which at its core consists of using the models serve command in the CLI, we can serve our models:

cd /gradflow/
export MLFLOW_TRACKING_URI=http://localhost:5000
mlflow models serve -m "models:/training-model-psystock/Production" -p 6000

We next will package the preceding commands in a Docker image so it can be used on any environment for deployment. The steps to achieve this are the following:

1. Generate a Docker image specifying the work directory and the...