rest-model-service 0.2.0

RESTful service for hosting machine learning models.

REST Model Service

RESTful service for hosting machine learning models.

Installation

The package can be installed from pypi:

pip install rest_model_service

To use the service you must first have a working model class that uses the MLModel base class from the ml_base package.

You can then set up a configuration file that points at the model class, the configuration file should look like this:

service_title: REST Model Service
models:
  - qualified_name: iris_model
    class_path: tests.mocks.IrisModel
    create_endpoint: true

The config file should be YAML and be in the current working directory.

The qualified name of your model and the class path to your model class should be placed in the correct place in the configuration file. The create_endpoint option is there for cases when you might want to load a model but not create an endpoint for it.

The rest_model_service package also supports decorators as defined in the ml_base package and explained here. A decorator can be added to a model by adding the "decorators" key to the model's configuration:

service_title: REST Model Service With Decorators
models:
  - qualified_name: iris_model
    class_path: tests.mocks.IrisModel
    create_endpoint: true
    decorators:
      - class_path: tests.mocks.PredictionIDDecorator

The PredictionIDDecorator will be instantiated and added to the IrisModel instance at application startup. Parameters can also be provided to the decorator's __init__() method like this:

service_title: REST Model Service With Decorators
models:
  - qualified_name: iris_model
    class_path: tests.mocks.IrisModel
    create_endpoint: true
    decorators:
      - class_path: tests.mocks.PredictionIDDecorator
        configuration:
          parameter1: "asdf"
          parameter2: "zxcv"

The configuration dictionary will be passed to the decorator as keyword arguments.

Creating an OpenAPI Contract

An OpenAPI contract can be generated dynamically for your models as hosted within the REST model service. To create the contract and save it execute this command:

generate_openapi --output_file=example.yaml

The script should be able to find your configuration file, but if you did not place it in the current working directory you can point the script to the right path by setting an environment variable like this:

export REST_CONFIG=examples/rest_config.yaml
generate_openapi --output_file=example.yaml

An example rest_config.yaml file is provided in the examples of the project. It points at a model class in the tests package.

The OpenAPI contract should be in your current working directory.

If you get an error that says something about not being able to find a module, you might need to update your PYTHONPATH environment variable:

export PYTHONPATH=./

The service relies on being able to find the model class in the python environment to load it and instantiate it. If your python interpreter is not able to find the model class, then the script won't be able to create an OpenAPI contract for it.

Running the Service

To start the service in development mode, execute this command:

uvicorn rest_model_service.main:app --reload

The service should be able to find your configuration file, but if you did not place it in the current working directory you can point the service to the right path like this:

export REST_CONFIG='examples/rest_config.yaml'
uvicorn rest_model_service.main:app --reload

Downloading Code and Setting Up for Development

To download the code and set up a development environment use these instructions.

To download the source code execute this command:

git clone https://github.com/schmidtbri/rest-model-service

Then create a virtual environment and activate it:

cd rest-model-service

make venv

# on Macs
source venv/bin/activate

Install the dependencies:

make dependencies

Running the Unit Tests

To run the unit test suite execute these commands:

# first install the test dependencies
make test-dependencies

# run the test suite
make test

# clean up the unit tests
make clean-test