ebonite.build.provider.ml_model module

ebonite.build.provider.ml_model.read(path)

class ebonite.build.provider.ml_model.MLModelProvider(model: ebonite.core.objects.core.Model, server: ebonite.runtime.server.base.Server, debug: bool = False)

Bases: ebonite.build.provider.base.PythonProvider

Provider to build service from Model object

Parameters:

• model – Model instance to build from
• server – Server instance to build with
• debug – Whether to run image in debug mode

get_requirements() → ebonite.core.objects.requirements.Requirements

Returns union of model, server and loader requirements

get_sources()

Returns model metadata file and sources of custom modules from requirements

get_artifacts() → ebonite.core.objects.artifacts.ArtifactCollection

Return model binaries

get_python_version()

Returns:version of python that produced model

class ebonite.build.provider.ml_model.ModelBuildable(model_id: Union[int, ebonite.core.objects.core.Model], server_type: str, debug: bool = False)

Bases: ebonite.build.provider.utils.BuildableWithServer, ebonite.core.objects.core.WithMetadataRepository

task

property to get task (can be None, whick forces to provide task manually)

model

get_provider() → ebonite.build.provider.ml_model.MLModelProvider

Abstract method to get a provider for this Buildable

type = 'ebonite.build.provider.ml_model.ModelBuildable'

class ebonite.build.provider.ml_model.BuildableModelHook

Bases: ebonite.core.analyzer.buildable.BuildableHook, ebonite.core.analyzer.base.TypeHookMixin

valid_types = [<class 'ebonite.core.objects.core.Model'>]

process(obj, **kwargs)

Analyzes obj and returns result. Result type is determined by specific Hook class sub-hierarchy

Parameters:

• obj – object to analyze
• kwargs – additional information to be used for analysis

Returns:

analysis result