Stack

zenml.stack

Initialization of the ZenML Stack.

The stack is essentially all the configuration for the infrastructure of your MLOps platform.

A stack is made up of multiple components. Some examples are:

• An Artifact Store
• An Orchestrator
• A Step Operator (Optional)
• A Container Registry (Optional)

Attributes

__all__ = ['Flavor', 'Stack', 'StackComponent', 'StackValidator', 'StackComponentConfig'] module-attribute

Classes

Flavor

Class for ZenML Flavors.

Attributes

config_class: Type[StackComponentConfig] abstractmethod property

Returns StackComponentConfig config class.

Returns:

Type	Description
Type[StackComponentConfig]	The config class.

config_schema: Dict[str, Any] property

The config schema for a flavor.

Returns:

Type	Description
Dict[str, Any]	The config schema.

docs_url: Optional[str] property

A url to point at docs explaining this flavor.

Returns:

Type	Description
Optional[str]	A flavor docs url.

implementation_class: Type[StackComponent] abstractmethod property

Implementation class for this flavor.

Returns:

Type	Description
Type[StackComponent]	The implementation class for this flavor.

logo_url: Optional[str] property

A url to represent the flavor in the dashboard.

Returns:

Type	Description
Optional[str]	The flavor logo.

name: str abstractmethod property

The flavor name.

Returns:

Type	Description
str	The flavor name.

sdk_docs_url: Optional[str] property

A url to point at SDK docs explaining this flavor.

Returns:

Type	Description
Optional[str]	A flavor SDK docs url.

service_connector_requirements: Optional[ServiceConnectorRequirements] property

Service connector resource requirements for service connectors.

Specifies resource requirements that are used to filter the available service connector types that are compatible with this flavor.

Returns:

Type	Description
Optional[ServiceConnectorRequirements]	Requirements for compatible service connectors, if a service
Optional[ServiceConnectorRequirements]	connector is required for this flavor.

type: StackComponentType abstractmethod property

The stack component type.

Returns:

Type	Description
StackComponentType	The stack component type.

Functions

from_model(flavor_model: FlavorResponse) -> Flavor classmethod

Loads a flavor from a model.

Parameters:

Name	Type	Description	Default
flavor_model	FlavorResponse	The model to load from.	required

Raises:

Type	Description
CustomFlavorImportError	If the custom flavor can't be imported.
ImportError	If the flavor can't be imported.

Returns:

Type	Description
Flavor	The loaded flavor.

Source code in src/zenml/stack/flavor.py

@classmethod
def from_model(cls, flavor_model: FlavorResponse) -> "Flavor":
    """Loads a flavor from a model.

    Args:
        flavor_model: The model to load from.

    Raises:
        CustomFlavorImportError: If the custom flavor can't be imported.
        ImportError: If the flavor can't be imported.

    Returns:
        The loaded flavor.
    """
    try:
        flavor = source_utils.load(flavor_model.source)()
    except (ModuleNotFoundError, ImportError, NotImplementedError) as err:
        if flavor_model.is_custom:
            flavor_module, _ = flavor_model.source.rsplit(".", maxsplit=1)
            expected_file_path = os.path.join(
                source_utils.get_source_root(),
                flavor_module.replace(".", os.path.sep),
            )
            raise CustomFlavorImportError(
                f"Couldn't import custom flavor {flavor_model.name}: "
                f"{err}. Make sure the custom flavor class "
                f"`{flavor_model.source}` is importable. If it is part of "
                "a library, make sure it is installed. If "
                "it is a local code file, make sure it exists at "
                f"`{expected_file_path}.py`."
            )
        else:
            raise ImportError(
                f"Couldn't import flavor {flavor_model.name}: {err}"
            )
    return cast(Flavor, flavor)

generate_default_docs_url() -> str

Generate the doc urls for all inbuilt and integration flavors.

Note that this method is not going to be useful for custom flavors, which do not have any docs in the main zenml docs.

Returns:

Type	Description
str	The complete url to the zenml documentation

Source code in src/zenml/stack/flavor.py

def generate_default_docs_url(self) -> str:
    """Generate the doc urls for all inbuilt and integration flavors.

    Note that this method is not going to be useful for custom flavors,
    which do not have any docs in the main zenml docs.

    Returns:
        The complete url to the zenml documentation
    """
    from zenml import __version__

    component_type = self.type.plural.replace("_", "-")
    name = self.name.replace("_", "-")

    try:
        is_latest = is_latest_zenml_version()
    except RuntimeError:
        # We assume in error cases that we are on the latest version
        is_latest = True

    if is_latest:
        base = "https://docs.zenml.io"
    else:
        base = f"https://zenml-io.gitbook.io/zenml-legacy-documentation/v/{__version__}"
    return f"{base}/stack-components/{component_type}/{name}"

generate_default_sdk_docs_url() -> str

Generate SDK docs url for a flavor.

Returns:

Type	Description
str	The complete url to the zenml SDK docs

Source code in src/zenml/stack/flavor.py

def generate_default_sdk_docs_url(self) -> str:
    """Generate SDK docs url for a flavor.

    Returns:
        The complete url to the zenml SDK docs
    """
    from zenml import __version__

    base = f"https://sdkdocs.zenml.io/{__version__}"

    component_type = self.type.plural

    if "zenml.integrations" in self.__module__:
        # Get integration name out of module path which will look something
        #  like this "zenml.integrations.<integration>....
        integration = self.__module__.split(
            "zenml.integrations.", maxsplit=1
        )[1].split(".")[0]

        return (
            f"{base}/integration_code_docs"
            f"/integrations-{integration}/#{self.__module__}"
        )

    else:
        return (
            f"{base}/core_code_docs/core-{component_type}/"
            f"#{self.__module__}"
        )

to_model(integration: Optional[str] = None, is_custom: bool = True) -> FlavorRequest

Converts a flavor to a model.

Parameters:

Name	Type	Description	Default
integration	Optional[str]	The integration to use for the model.	None
is_custom	bool	Whether the flavor is a custom flavor.	True

Returns:

Type	Description
FlavorRequest	The model.

Source code in src/zenml/stack/flavor.py

def to_model(
    self,
    integration: Optional[str] = None,
    is_custom: bool = True,
) -> FlavorRequest:
    """Converts a flavor to a model.

    Args:
        integration: The integration to use for the model.
        is_custom: Whether the flavor is a custom flavor.

    Returns:
        The model.
    """
    connector_requirements = self.service_connector_requirements
    connector_type = (
        connector_requirements.connector_type
        if connector_requirements
        else None
    )
    resource_type = (
        connector_requirements.resource_type
        if connector_requirements
        else None
    )
    resource_id_attr = (
        connector_requirements.resource_id_attr
        if connector_requirements
        else None
    )

    model = FlavorRequest(
        name=self.name,
        type=self.type,
        source=source_utils.resolve(self.__class__).import_path,
        config_schema=self.config_schema,
        connector_type=connector_type,
        connector_resource_type=resource_type,
        connector_resource_id_attr=resource_id_attr,
        integration=integration,
        logo_url=self.logo_url,
        docs_url=self.docs_url,
        sdk_docs_url=self.sdk_docs_url,
        is_custom=is_custom,
    )
    return model

Stack(id: UUID, name: str, *, orchestrator: BaseOrchestrator, artifact_store: BaseArtifactStore, container_registry: Optional[BaseContainerRegistry] = None, step_operator: Optional[BaseStepOperator] = None, feature_store: Optional[BaseFeatureStore] = None, model_deployer: Optional[BaseModelDeployer] = None, experiment_tracker: Optional[BaseExperimentTracker] = None, alerter: Optional[BaseAlerter] = None, annotator: Optional[BaseAnnotator] = None, data_validator: Optional[BaseDataValidator] = None, image_builder: Optional[BaseImageBuilder] = None, model_registry: Optional[BaseModelRegistry] = None)

ZenML stack class.

A ZenML stack is a collection of multiple stack components that are required to run ZenML pipelines. Some of these components (orchestrator, and artifact store) are required to run any kind of pipeline, other components like the container registry are only required if other stack components depend on them.

Initializes and validates a stack instance.

Parameters:

Name	Type	Description	Default
id	UUID	Unique ID of the stack.	required
name	str	Name of the stack.	required
orchestrator	BaseOrchestrator	Orchestrator component of the stack.	required
artifact_store	BaseArtifactStore	Artifact store component of the stack.	required
container_registry	Optional[BaseContainerRegistry]	Container registry component of the stack.	None
step_operator	Optional[BaseStepOperator]	Step operator component of the stack.	None
feature_store	Optional[BaseFeatureStore]	Feature store component of the stack.	None
model_deployer	Optional[BaseModelDeployer]	Model deployer component of the stack.	None
experiment_tracker	Optional[BaseExperimentTracker]	Experiment tracker component of the stack.	None
alerter	Optional[BaseAlerter]	Alerter component of the stack.	None
annotator	Optional[BaseAnnotator]	Annotator component of the stack.	None
data_validator	Optional[BaseDataValidator]	Data validator component of the stack.	None
image_builder	Optional[BaseImageBuilder]	Image builder component of the stack.	None
model_registry	Optional[BaseModelRegistry]	Model registry component of the stack.	None

Source code in src/zenml/stack/stack.py

def __init__(
    self,
    id: UUID,
    name: str,
    *,
    orchestrator: "BaseOrchestrator",
    artifact_store: "BaseArtifactStore",
    container_registry: Optional["BaseContainerRegistry"] = None,
    step_operator: Optional["BaseStepOperator"] = None,
    feature_store: Optional["BaseFeatureStore"] = None,
    model_deployer: Optional["BaseModelDeployer"] = None,
    experiment_tracker: Optional["BaseExperimentTracker"] = None,
    alerter: Optional["BaseAlerter"] = None,
    annotator: Optional["BaseAnnotator"] = None,
    data_validator: Optional["BaseDataValidator"] = None,
    image_builder: Optional["BaseImageBuilder"] = None,
    model_registry: Optional["BaseModelRegistry"] = None,
):
    """Initializes and validates a stack instance.

    Args:
        id: Unique ID of the stack.
        name: Name of the stack.
        orchestrator: Orchestrator component of the stack.
        artifact_store: Artifact store component of the stack.
        container_registry: Container registry component of the stack.
        step_operator: Step operator component of the stack.
        feature_store: Feature store component of the stack.
        model_deployer: Model deployer component of the stack.
        experiment_tracker: Experiment tracker component of the stack.
        alerter: Alerter component of the stack.
        annotator: Annotator component of the stack.
        data_validator: Data validator component of the stack.
        image_builder: Image builder component of the stack.
        model_registry: Model registry component of the stack.
    """
    self._id = id
    self._name = name
    self._orchestrator = orchestrator
    self._artifact_store = artifact_store
    self._container_registry = container_registry
    self._step_operator = step_operator
    self._feature_store = feature_store
    self._model_deployer = model_deployer
    self._experiment_tracker = experiment_tracker
    self._alerter = alerter
    self._annotator = annotator
    self._data_validator = data_validator
    self._model_registry = model_registry
    self._image_builder = image_builder

Attributes

alerter: Optional[BaseAlerter] property

The alerter of the stack.

Returns:

Type	Description
Optional[BaseAlerter]	The alerter of the stack.

annotator: Optional[BaseAnnotator] property

The annotator of the stack.

Returns:

Type	Description
Optional[BaseAnnotator]	The annotator of the stack.

apt_packages: List[str] property

List of APT package requirements for the stack.

Returns:

Type	Description
List[str]	A list of APT package requirements for the stack.

artifact_store: BaseArtifactStore property

The artifact store of the stack.

Returns:

Type	Description
BaseArtifactStore	The artifact store of the stack.

components: Dict[StackComponentType, StackComponent] property

All components of the stack.

Returns:

Type	Description
Dict[StackComponentType, StackComponent]	A dictionary of all components of the stack.

container_registry: Optional[BaseContainerRegistry] property

The container registry of the stack.

Returns:

Type	Description
Optional[BaseContainerRegistry]	The container registry of the stack or None if the stack does not
Optional[BaseContainerRegistry]	have a container registry.

data_validator: Optional[BaseDataValidator] property

The data validator of the stack.

Returns:

Type	Description
Optional[BaseDataValidator]	The data validator of the stack.

experiment_tracker: Optional[BaseExperimentTracker] property

The experiment tracker of the stack.

Returns:

Type	Description
Optional[BaseExperimentTracker]	The experiment tracker of the stack.

feature_store: Optional[BaseFeatureStore] property

The feature store of the stack.

Returns:

Type	Description
Optional[BaseFeatureStore]	The feature store of the stack.

id: UUID property

The ID of the stack.

Returns:

Type	Description
UUID	The ID of the stack.

image_builder: Optional[BaseImageBuilder] property

The image builder of the stack.

Returns:

Type	Description
Optional[BaseImageBuilder]	The image builder of the stack.

model_deployer: Optional[BaseModelDeployer] property

The model deployer of the stack.

Returns:

Type	Description
Optional[BaseModelDeployer]	The model deployer of the stack.

model_registry: Optional[BaseModelRegistry] property

The model registry of the stack.

Returns:

Type	Description
Optional[BaseModelRegistry]	The model registry of the stack.

name: str property

The name of the stack.

Returns:

Name	Type	Description
str	str	The name of the stack.

orchestrator: BaseOrchestrator property

The orchestrator of the stack.

Returns:

Type	Description
BaseOrchestrator	The orchestrator of the stack.

required_secrets: Set[secret_utils.SecretReference] property

All required secrets for this stack.

Returns:

Type	Description
Set[SecretReference]	The required secrets of this stack.

requires_remote_server: bool property

If the stack requires a remote ZenServer to run.

This is the case if any code is getting executed remotely. This is the case for both remote orchestrators as well as remote step operators.

Returns:

Type	Description
bool	If the stack requires a remote ZenServer to run.

setting_classes: Dict[str, Type[BaseSettings]] property

Setting classes of all components of this stack.

Returns:

Type	Description
Dict[str, Type[BaseSettings]]	All setting classes and their respective keys.

step_operator: Optional[BaseStepOperator] property

The step operator of the stack.

Returns:

Type	Description
Optional[BaseStepOperator]	The step operator of the stack.

Functions

check_local_paths() -> bool

Checks if the stack has local paths.

Returns:

Type	Description
bool	True if the stack has local paths, False otherwise.

Raises:

Type	Description
ValueError	If the stack has local paths that do not conform to the convention that all local path must be relative to the local stores directory.

Source code in src/zenml/stack/stack.py

def check_local_paths(self) -> bool:
    """Checks if the stack has local paths.

    Returns:
        True if the stack has local paths, False otherwise.

    Raises:
        ValueError: If the stack has local paths that do not conform to
            the convention that all local path must be relative to the
            local stores directory.
    """
    from zenml.config.global_config import GlobalConfiguration

    local_stores_path = GlobalConfiguration().local_stores_path

    # go through all stack components and identify those that advertise
    # a local path where they persist information that they need to be
    # available when running pipelines.
    has_local_paths = False
    for stack_comp in self.components.values():
        local_path = stack_comp.local_path
        if not local_path:
            continue
        # double-check this convention, just in case it wasn't respected
        # as documented in `StackComponent.local_path`
        if not local_path.startswith(local_stores_path):
            raise ValueError(
                f"Local path {local_path} for component "
                f"{stack_comp.name} is not in the local stores "
                f"directory ({local_stores_path})."
            )
        has_local_paths = True

    return has_local_paths

cleanup_step_run(info: StepRunInfo, step_failed: bool) -> None

Cleans up resources after the step run is finished.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that was executed.	required
step_failed	bool	Whether the step failed.	required

Source code in src/zenml/stack/stack.py

def cleanup_step_run(self, info: "StepRunInfo", step_failed: bool) -> None:
    """Cleans up resources after the step run is finished.

    Args:
        info: Info about the step that was executed.
        step_failed: Whether the step failed.
    """
    for component in self._get_active_components_for_step(
        info.config
    ).values():
        component.cleanup_step_run(info=info, step_failed=step_failed)

deploy_pipeline(deployment: PipelineDeploymentResponse, placeholder_run: Optional[PipelineRunResponse] = None) -> None

Deploys a pipeline on this stack.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentResponse	The pipeline deployment.	required
placeholder_run	Optional[PipelineRunResponse]	An optional placeholder run for the deployment.	None

Source code in src/zenml/stack/stack.py

def deploy_pipeline(
    self,
    deployment: "PipelineDeploymentResponse",
    placeholder_run: Optional["PipelineRunResponse"] = None,
) -> None:
    """Deploys a pipeline on this stack.

    Args:
        deployment: The pipeline deployment.
        placeholder_run: An optional placeholder run for the deployment.
    """
    self.orchestrator.run(
        deployment=deployment, stack=self, placeholder_run=placeholder_run
    )

dict() -> Dict[str, str]

Converts the stack into a dictionary.

Returns:

Type	Description
Dict[str, str]	A dictionary containing the stack components.

Source code in src/zenml/stack/stack.py

def dict(self) -> Dict[str, str]:
    """Converts the stack into a dictionary.

    Returns:
        A dictionary containing the stack components.
    """
    component_dict = {
        component_type.value: json.dumps(
            component.config.model_dump(mode="json"), sort_keys=True
        )
        for component_type, component in self.components.items()
    }
    component_dict.update({"name": self.name})
    return component_dict

from_components(id: UUID, name: str, components: Dict[StackComponentType, StackComponent]) -> Stack classmethod

Creates a stack instance from a dict of stack components.

noqa: DAR402

Parameters:

Name	Type	Description	Default
id	UUID	Unique ID of the stack.	required
name	str	The name of the stack.	required
components	Dict[StackComponentType, StackComponent]	The components of the stack.	required

Returns:

Type	Description
Stack	A stack instance consisting of the given components.

Raises:

Type	Description
TypeError	If a required component is missing or a component doesn't inherit from the expected base class.

Source code in src/zenml/stack/stack.py

@classmethod
def from_components(
    cls,
    id: UUID,
    name: str,
    components: Dict[StackComponentType, "StackComponent"],
) -> "Stack":
    """Creates a stack instance from a dict of stack components.

    # noqa: DAR402

    Args:
        id: Unique ID of the stack.
        name: The name of the stack.
        components: The components of the stack.

    Returns:
        A stack instance consisting of the given components.

    Raises:
        TypeError: If a required component is missing or a component
            doesn't inherit from the expected base class.
    """
    from zenml.alerter import BaseAlerter
    from zenml.annotators import BaseAnnotator
    from zenml.artifact_stores import BaseArtifactStore
    from zenml.container_registries import BaseContainerRegistry
    from zenml.data_validators import BaseDataValidator
    from zenml.experiment_trackers import BaseExperimentTracker
    from zenml.feature_stores import BaseFeatureStore
    from zenml.image_builders import BaseImageBuilder
    from zenml.model_deployers import BaseModelDeployer
    from zenml.model_registries import BaseModelRegistry
    from zenml.orchestrators import BaseOrchestrator
    from zenml.step_operators import BaseStepOperator

    def _raise_type_error(
        component: Optional["StackComponent"], expected_class: Type[Any]
    ) -> NoReturn:
        """Raises a TypeError that the component has an unexpected type.

        Args:
            component: The component that has an unexpected type.
            expected_class: The expected type of the component.

        Raises:
            TypeError: If the component has an unexpected type.
        """
        raise TypeError(
            f"Unable to create stack: Wrong stack component type "
            f"`{component.__class__.__name__}` (expected: subclass "
            f"of `{expected_class.__name__}`)"
        )

    orchestrator = components.get(StackComponentType.ORCHESTRATOR)
    if not isinstance(orchestrator, BaseOrchestrator):
        _raise_type_error(orchestrator, BaseOrchestrator)

    artifact_store = components.get(StackComponentType.ARTIFACT_STORE)
    if not isinstance(artifact_store, BaseArtifactStore):
        _raise_type_error(artifact_store, BaseArtifactStore)

    container_registry = components.get(
        StackComponentType.CONTAINER_REGISTRY
    )
    if container_registry is not None and not isinstance(
        container_registry, BaseContainerRegistry
    ):
        _raise_type_error(container_registry, BaseContainerRegistry)

    step_operator = components.get(StackComponentType.STEP_OPERATOR)
    if step_operator is not None and not isinstance(
        step_operator, BaseStepOperator
    ):
        _raise_type_error(step_operator, BaseStepOperator)

    feature_store = components.get(StackComponentType.FEATURE_STORE)
    if feature_store is not None and not isinstance(
        feature_store, BaseFeatureStore
    ):
        _raise_type_error(feature_store, BaseFeatureStore)

    model_deployer = components.get(StackComponentType.MODEL_DEPLOYER)
    if model_deployer is not None and not isinstance(
        model_deployer, BaseModelDeployer
    ):
        _raise_type_error(model_deployer, BaseModelDeployer)

    experiment_tracker = components.get(
        StackComponentType.EXPERIMENT_TRACKER
    )
    if experiment_tracker is not None and not isinstance(
        experiment_tracker, BaseExperimentTracker
    ):
        _raise_type_error(experiment_tracker, BaseExperimentTracker)

    alerter = components.get(StackComponentType.ALERTER)
    if alerter is not None and not isinstance(alerter, BaseAlerter):
        _raise_type_error(alerter, BaseAlerter)

    annotator = components.get(StackComponentType.ANNOTATOR)
    if annotator is not None and not isinstance(annotator, BaseAnnotator):
        _raise_type_error(annotator, BaseAnnotator)

    data_validator = components.get(StackComponentType.DATA_VALIDATOR)
    if data_validator is not None and not isinstance(
        data_validator, BaseDataValidator
    ):
        _raise_type_error(data_validator, BaseDataValidator)

    image_builder = components.get(StackComponentType.IMAGE_BUILDER)
    if image_builder is not None and not isinstance(
        image_builder, BaseImageBuilder
    ):
        _raise_type_error(image_builder, BaseImageBuilder)

    model_registry = components.get(StackComponentType.MODEL_REGISTRY)
    if model_registry is not None and not isinstance(
        model_registry, BaseModelRegistry
    ):
        _raise_type_error(model_registry, BaseModelRegistry)

    return Stack(
        id=id,
        name=name,
        orchestrator=orchestrator,
        artifact_store=artifact_store,
        container_registry=container_registry,
        step_operator=step_operator,
        feature_store=feature_store,
        model_deployer=model_deployer,
        experiment_tracker=experiment_tracker,
        alerter=alerter,
        annotator=annotator,
        data_validator=data_validator,
        image_builder=image_builder,
        model_registry=model_registry,
    )

from_model(stack_model: StackResponse) -> Stack classmethod

Creates a Stack instance from a StackModel.

Parameters:

Name	Type	Description	Default
stack_model	StackResponse	The StackModel to create the Stack from.	required

Returns:

Type	Description
Stack	The created Stack instance.

Source code in src/zenml/stack/stack.py

@classmethod
def from_model(cls, stack_model: "StackResponse") -> "Stack":
    """Creates a Stack instance from a StackModel.

    Args:
        stack_model: The StackModel to create the Stack from.

    Returns:
        The created Stack instance.
    """
    global _STACK_CACHE
    key = (stack_model.id, stack_model.updated)
    if key in _STACK_CACHE:
        return _STACK_CACHE[key]

    from zenml.stack import StackComponent

    # Run a hydrated list call once to avoid one request per component
    component_models = pagination_utils.depaginate(
        Client().list_stack_components,
        stack_id=stack_model.id,
        hydrate=True,
    )

    stack_components = {
        model.type: StackComponent.from_model(model)
        for model in component_models
    }
    stack = Stack.from_components(
        id=stack_model.id,
        name=stack_model.name,
        components=stack_components,
    )
    _STACK_CACHE[key] = stack

    client = Client()
    if stack_model.id == client.active_stack_model.id:
        if stack_model.updated > client.active_stack_model.updated:
            if client._config:
                client._config.set_active_stack(stack_model)
            else:
                GlobalConfiguration().set_active_stack(stack_model)

    return stack

get_docker_builds(deployment: PipelineDeploymentBase) -> List[BuildConfiguration]

Gets the Docker builds required for the stack.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentBase	The pipeline deployment for which to get the builds.	required

Returns:

Type	Description
List[BuildConfiguration]	The required Docker builds.

Source code in src/zenml/stack/stack.py

def get_docker_builds(
    self, deployment: "PipelineDeploymentBase"
) -> List["BuildConfiguration"]:
    """Gets the Docker builds required for the stack.

    Args:
        deployment: The pipeline deployment for which to get the builds.

    Returns:
        The required Docker builds.
    """
    return list(
        itertools.chain.from_iterable(
            component.get_docker_builds(deployment=deployment)
            for component in self.components.values()
        )
    )

get_pipeline_run_metadata(run_id: UUID) -> Dict[UUID, Dict[str, MetadataType]]

Get general component-specific metadata for a pipeline run.

Parameters:

Name	Type	Description	Default
run_id	UUID	ID of the pipeline run.	required

Returns:

Type	Description
Dict[UUID, Dict[str, MetadataType]]	A dictionary mapping component IDs to the metadata they created.

Source code in src/zenml/stack/stack.py

def get_pipeline_run_metadata(
    self, run_id: UUID
) -> Dict[UUID, Dict[str, MetadataType]]:
    """Get general component-specific metadata for a pipeline run.

    Args:
        run_id: ID of the pipeline run.

    Returns:
        A dictionary mapping component IDs to the metadata they created.
    """
    pipeline_run_metadata: Dict[UUID, Dict[str, MetadataType]] = {}
    for component in self.components.values():
        try:
            component_metadata = component.get_pipeline_run_metadata(
                run_id=run_id
            )
            if component_metadata:
                pipeline_run_metadata[component.id] = component_metadata
        except Exception as e:
            logger.warning(
                f"Extracting pipeline run metadata failed for component "
                f"'{component.name}' of type '{component.type}': {e}"
            )
    return pipeline_run_metadata

get_step_run_metadata(info: StepRunInfo) -> Dict[UUID, Dict[str, MetadataType]]

Get component-specific metadata for a step run.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that was executed.	required

Returns:

Type	Description
Dict[UUID, Dict[str, MetadataType]]	A dictionary mapping component IDs to the metadata they created.

Source code in src/zenml/stack/stack.py

def get_step_run_metadata(
    self, info: "StepRunInfo"
) -> Dict[UUID, Dict[str, MetadataType]]:
    """Get component-specific metadata for a step run.

    Args:
        info: Info about the step that was executed.

    Returns:
        A dictionary mapping component IDs to the metadata they created.
    """
    step_run_metadata: Dict[UUID, Dict[str, MetadataType]] = {}
    for component in self._get_active_components_for_step(
        info.config
    ).values():
        try:
            component_metadata = component.get_step_run_metadata(info=info)
            if component_metadata:
                step_run_metadata[component.id] = component_metadata
        except Exception as e:
            logger.warning(
                f"Extracting step run metadata failed for component "
                f"'{component.name}' of type '{component.type}': {e}"
            )
    return step_run_metadata

prepare_pipeline_deployment(deployment: PipelineDeploymentResponse) -> None

Prepares the stack for a pipeline deployment.

This method is called before a pipeline is deployed.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentResponse	The pipeline deployment	required

Raises:

Type	Description
RuntimeError	If trying to deploy a pipeline that requires a remote ZenML server with a local one.

Source code in src/zenml/stack/stack.py

def prepare_pipeline_deployment(
    self, deployment: "PipelineDeploymentResponse"
) -> None:
    """Prepares the stack for a pipeline deployment.

    This method is called before a pipeline is deployed.

    Args:
        deployment: The pipeline deployment

    Raises:
        RuntimeError: If trying to deploy a pipeline that requires a remote
            ZenML server with a local one.
    """
    self.validate(fail_if_secrets_missing=True)

    if self.requires_remote_server and Client().zen_store.is_local_store():
        raise RuntimeError(
            "Stacks with remote components such as remote orchestrators "
            "and step operators require a remote "
            "ZenML server. To run a pipeline with this stack you need to "
            "connect to a remote ZenML server first. Check out "
            "https://docs.zenml.io/getting-started/deploying-zenml "
            "for more information on how to deploy ZenML."
        )

    for component in self.components.values():
        component.prepare_pipeline_deployment(
            deployment=deployment, stack=self
        )

prepare_step_run(info: StepRunInfo) -> None

Prepares running a step.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that will be executed.	required

Source code in src/zenml/stack/stack.py

def prepare_step_run(self, info: "StepRunInfo") -> None:
    """Prepares running a step.

    Args:
        info: Info about the step that will be executed.
    """
    for component in self._get_active_components_for_step(
        info.config
    ).values():
        component.prepare_step_run(info=info)

requirements(exclude_components: Optional[AbstractSet[StackComponentType]] = None) -> Set[str]

Set of PyPI requirements for the stack.

This method combines the requirements of all stack components (except the ones specified in exclude_components).

Parameters:

Name	Type	Description	Default
exclude_components	Optional[AbstractSet[StackComponentType]]	Set of component types for which the requirements should not be included in the output.	None

Returns:

Type	Description
Set[str]	Set of PyPI requirements.

Source code in src/zenml/stack/stack.py

def requirements(
    self,
    exclude_components: Optional[AbstractSet[StackComponentType]] = None,
) -> Set[str]:
    """Set of PyPI requirements for the stack.

    This method combines the requirements of all stack components (except
    the ones specified in `exclude_components`).

    Args:
        exclude_components: Set of component types for which the
            requirements should not be included in the output.

    Returns:
        Set of PyPI requirements.
    """
    exclude_components = exclude_components or set()
    requirements = [
        component.requirements
        for component in self.components.values()
        if component.type not in exclude_components
    ]
    return set.union(*requirements) if requirements else set()

validate(fail_if_secrets_missing: bool = False) -> None

Checks whether the stack configuration is valid.

To check if a stack configuration is valid, the following criteria must be met: - the stack must have an image builder if other components require it - the StackValidator of each stack component has to validate the stack to make sure all the components are compatible with each other - the required secrets of all components need to exist

Parameters:

Name	Type	Description	Default
fail_if_secrets_missing	bool	If this is True, an error will be raised if a secret for a component is missing. Otherwise, only a warning will be logged.	False

Source code in src/zenml/stack/stack.py

def validate(
    self,
    fail_if_secrets_missing: bool = False,
) -> None:
    """Checks whether the stack configuration is valid.

    To check if a stack configuration is valid, the following criteria must
    be met:
    - the stack must have an image builder if other components require it
    - the `StackValidator` of each stack component has to validate the
        stack to make sure all the components are compatible with each other
    - the required secrets of all components need to exist

    Args:
        fail_if_secrets_missing: If this is `True`, an error will be raised
            if a secret for a component is missing. Otherwise, only a
            warning will be logged.
    """
    if handle_bool_env_var(ENV_ZENML_SKIP_STACK_VALIDATION, default=False):
        logger.debug("Skipping stack validation.")
        return

    self.validate_image_builder()
    for component in self.components.values():
        if component.validator:
            component.validator.validate(stack=self)

    self._validate_secrets(raise_exception=fail_if_secrets_missing)

validate_image_builder() -> None

Validates that the stack has an image builder if required.

If the stack requires an image builder, but none is specified, a local image builder will be created and assigned to the stack to ensure backwards compatibility.

Source code in src/zenml/stack/stack.py

def validate_image_builder(self) -> None:
    """Validates that the stack has an image builder if required.

    If the stack requires an image builder, but none is specified, a
    local image builder will be created and assigned to the stack to
    ensure backwards compatibility.
    """
    requires_image_builder = (
        self.orchestrator.flavor != "local"
        or self.step_operator
        or (self.model_deployer and self.model_deployer.flavor != "mlflow")
    )
    skip_default_image_builder = handle_bool_env_var(
        ENV_ZENML_SKIP_IMAGE_BUILDER_DEFAULT, default=False
    )
    if (
        requires_image_builder
        and not skip_default_image_builder
        and not self.image_builder
    ):
        from uuid import uuid4

        from zenml.image_builders import (
            LocalImageBuilder,
            LocalImageBuilderConfig,
            LocalImageBuilderFlavor,
        )

        flavor = LocalImageBuilderFlavor()

        now = utc_now()
        image_builder = LocalImageBuilder(
            id=uuid4(),
            name="temporary_default",
            flavor=flavor.name,
            type=flavor.type,
            config=LocalImageBuilderConfig(),
            user=Client().active_user.id,
            created=now,
            updated=now,
        )

        self._image_builder = image_builder

StackComponent(name: str, id: UUID, config: StackComponentConfig, flavor: str, type: StackComponentType, user: Optional[UUID], created: datetime, updated: datetime, labels: Optional[Dict[str, Any]] = None, connector_requirements: Optional[ServiceConnectorRequirements] = None, connector: Optional[UUID] = None, connector_resource_id: Optional[str] = None, *args: Any, **kwargs: Any)

Abstract StackComponent class for all components of a ZenML stack.

Initializes a StackComponent.

Parameters:

Name	Type	Description	Default
name	str	The name of the component.	required
id	UUID	The unique ID of the component.	required
config	StackComponentConfig	The config of the component.	required
flavor	str	The flavor of the component.	required
type	StackComponentType	The type of the component.	required
user	Optional[UUID]	The ID of the user who created the component.	required
created	datetime	The creation time of the component.	required
updated	datetime	The last update time of the component.	required
labels	Optional[Dict[str, Any]]	The labels of the component.	None
connector_requirements	Optional[ServiceConnectorRequirements]	The requirements for the connector.	None
connector	Optional[UUID]	The ID of a connector linked to the component.	None
connector_resource_id	Optional[str]	The custom resource ID to access through the connector.	None
*args	Any	Additional positional arguments.	()
**kwargs	Any	Additional keyword arguments.	{}

Raises:

Type	Description
ValueError	If a secret reference is passed as name.

Source code in src/zenml/stack/stack_component.py

def __init__(
    self,
    name: str,
    id: UUID,
    config: StackComponentConfig,
    flavor: str,
    type: StackComponentType,
    user: Optional[UUID],
    created: datetime,
    updated: datetime,
    labels: Optional[Dict[str, Any]] = None,
    connector_requirements: Optional[ServiceConnectorRequirements] = None,
    connector: Optional[UUID] = None,
    connector_resource_id: Optional[str] = None,
    *args: Any,
    **kwargs: Any,
):
    """Initializes a StackComponent.

    Args:
        name: The name of the component.
        id: The unique ID of the component.
        config: The config of the component.
        flavor: The flavor of the component.
        type: The type of the component.
        user: The ID of the user who created the component.
        created: The creation time of the component.
        updated: The last update time of the component.
        labels: The labels of the component.
        connector_requirements: The requirements for the connector.
        connector: The ID of a connector linked to the component.
        connector_resource_id: The custom resource ID to access through
            the connector.
        *args: Additional positional arguments.
        **kwargs: Additional keyword arguments.

    Raises:
        ValueError: If a secret reference is passed as name.
    """
    if secret_utils.is_secret_reference(name):
        raise ValueError(
            "Passing the `name` attribute of a stack component as a "
            "secret reference is not allowed."
        )

    self.id = id
    self.name = name
    self._config = config
    self.flavor = flavor
    self.type = type
    self.user = user
    self.created = created
    self.updated = updated
    self.labels = labels
    self.connector_requirements = connector_requirements
    self.connector = connector
    self.connector_resource_id = connector_resource_id
    self._connector_instance: Optional[ServiceConnector] = None

Attributes

apt_packages: List[str] property

List of APT package requirements for the component.

Returns:

Type	Description
List[str]	A list of APT package requirements for the component.

config: StackComponentConfig property

Returns the configuration of the stack component.

This should be overwritten by any subclasses that define custom configs to return the correct config class.

Returns:

Type	Description
StackComponentConfig	The configuration of the stack component.

local_path: Optional[str] property

Path to a local directory to store persistent information.

This property should only be implemented by components that need to store persistent information in a directory on the local machine and also need that information to be available during pipeline runs.

IMPORTANT: the path returned by this property must always be a path that is relative to the ZenML local store's directory. The local orchestrators rely on this convention to correctly mount the local folders in the containers. This is an example of a valid path:

from zenml.config.global_config import GlobalConfiguration

...

@property
def local_path(self) -> Optional[str]:

    return os.path.join(
        GlobalConfiguration().local_stores_path,
        str(self.uuid),
    )

Returns:

Type	Description
Optional[str]	A path to a local directory used by the component to store
Optional[str]	persistent information.

log_file: Optional[str] property

Optional path to a log file for the stack component.

Returns:

Type	Description
Optional[str]	Optional path to a log file for the stack component.

post_registration_message: Optional[str] property

Optional message printed after the stack component is registered.

Returns:

Type	Description
Optional[str]	An optional message.

requirements: Set[str] property

Set of PyPI requirements for the component.

Returns:

Type	Description
Set[str]	A set of PyPI requirements for the component.

settings_class: Optional[Type[BaseSettings]] property

Class specifying available settings for this component.

Returns:

Type	Description
Optional[Type[BaseSettings]]	Optional settings class.

validator: Optional[StackValidator] property

The optional validator of the stack component.

This validator will be called each time a stack with the stack component is initialized. Subclasses should override this property and return a StackValidator that makes sure they're not included in any stack that they're not compatible with.

Returns:

Type	Description
Optional[StackValidator]	An optional StackValidator instance.

Functions

cleanup() -> None

Cleans up the component after it has been used.

Source code in src/zenml/stack/stack_component.py

def cleanup(self) -> None:
    """Cleans up the component after it has been used."""
    pass

cleanup_step_run(info: StepRunInfo, step_failed: bool) -> None

Cleans up resources after the step run is finished.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that was executed.	required
step_failed	bool	Whether the step failed.	required

Source code in src/zenml/stack/stack_component.py

def cleanup_step_run(self, info: "StepRunInfo", step_failed: bool) -> None:
    """Cleans up resources after the step run is finished.

    Args:
        info: Info about the step that was executed.
        step_failed: Whether the step failed.
    """

connector_has_expired() -> bool

Checks whether the connector linked to this stack component has expired.

Returns:

Type	Description
bool	Whether the connector linked to this stack component has expired, or isn't linked to a connector.

Source code in src/zenml/stack/stack_component.py

def connector_has_expired(self) -> bool:
    """Checks whether the connector linked to this stack component has expired.

    Returns:
        Whether the connector linked to this stack component has expired, or isn't linked to a connector.
    """
    if self.connector is None:
        # The stack component isn't linked to a connector
        return False

    if self._connector_instance is None:
        return True

    return self._connector_instance.has_expired()

from_model(component_model: ComponentResponse) -> StackComponent classmethod

Creates a StackComponent from a ComponentModel.

Parameters:

Name	Type	Description	Default
component_model	ComponentResponse	The ComponentModel to create the StackComponent	required

Returns:

Type	Description
StackComponent	The created StackComponent.

Raises:

Type	Description
ImportError	If the flavor can't be imported.

Source code in src/zenml/stack/stack_component.py

@classmethod
def from_model(
    cls, component_model: "ComponentResponse"
) -> "StackComponent":
    """Creates a StackComponent from a ComponentModel.

    Args:
        component_model: The ComponentModel to create the StackComponent

    Returns:
        The created StackComponent.

    Raises:
        ImportError: If the flavor can't be imported.
    """
    from zenml.stack import Flavor

    flavor_model = component_model.flavor
    flavor = Flavor.from_model(flavor_model)

    configuration = flavor.config_class(**component_model.configuration)

    if component_model.user is not None:
        user_id = component_model.user.id
    else:
        user_id = None

    try:
        return flavor.implementation_class(
            user=user_id,
            name=component_model.name,
            id=component_model.id,
            config=configuration,
            labels=component_model.labels,
            flavor=component_model.flavor_name,
            type=component_model.type,
            created=component_model.created,
            updated=component_model.updated,
            connector_requirements=flavor.service_connector_requirements,
            connector=component_model.connector.id
            if component_model.connector
            else None,
            connector_resource_id=component_model.connector_resource_id,
        )
    except ImportError as e:
        from zenml.integrations.registry import integration_registry

        integration_requirements = " ".join(
            integration_registry.select_integration_requirements(
                flavor_model.integration
            )
        )

        if integration_registry.is_installed(flavor_model.integration):
            raise ImportError(
                f"{e}\n\n"
                f"Something went wrong while trying to import from the "
                f"`{flavor_model.integration}` integration. Please make "
                "sure that all its requirements are installed properly by "
                "reinstalling the integration either through our CLI: "
                f"`zenml integration install {flavor_model.integration} "
                "-y` or by manually installing its requirements: "
                f"`pip install {integration_requirements}`. If the error "
                "persists, please contact the ZenML team."
            ) from e
        else:
            raise ImportError(
                f"{e}\n\n"
                f"The `{flavor_model.integration}` integration that you "
                "are trying to use is not installed in your current "
                "environment. Please make sure that it is installed by "
                "either using our CLI: `zenml integration install "
                f"{flavor_model.integration}` or by manually installing "
                f"its requirements: `pip install "
                f"{integration_requirements}`"
            ) from e

get_connector() -> Optional[ServiceConnector]

Returns the connector linked to this stack component.

Returns:

Type	Description
Optional[ServiceConnector]	The connector linked to this stack component.

Raises:

Type	Description
RuntimeError	If the stack component does not specify connector requirements or if the connector linked to the component is not compatible or not found.

Source code in src/zenml/stack/stack_component.py

def get_connector(self) -> Optional["ServiceConnector"]:
    """Returns the connector linked to this stack component.

    Returns:
        The connector linked to this stack component.

    Raises:
        RuntimeError: If the stack component does not specify connector
            requirements or if the connector linked to the component is not
            compatible or not found.
    """
    from zenml.client import Client

    if self.connector is None:
        return None

    if self._connector_instance is not None:
        # If the connector instance is still valid, return it. Otherwise,
        # we'll try to get a new one.
        if not self._connector_instance.has_expired():
            return self._connector_instance

    if self.connector_requirements is None:
        raise RuntimeError(
            f"Unable to get connector for component {self} because this "
            "component does not declare any connector requirements in its. "
            "flavor specification. Override the "
            "`service_connector_requirements` method in its flavor class "
            "to return a connector requirements specification and try "
            "again."
        )

    if self.connector_requirements.resource_id_attr is not None:
        # Check if an attribute is set in the component configuration
        resource_id = getattr(
            self.config, self.connector_requirements.resource_id_attr
        )
    else:
        # Otherwise, use the resource ID configured in the component
        resource_id = self.connector_resource_id

    client = Client()
    try:
        self._connector_instance = client.get_service_connector_client(
            name_id_or_prefix=self.connector,
            resource_type=self.connector_requirements.resource_type,
            resource_id=resource_id,
        )
    except KeyError:
        raise RuntimeError(
            f"The connector with ID {self.connector} linked "
            f"to the '{self.name}' {self.type} stack component could not "
            f"be found or is not accessible. Please verify that the "
            f"connector exists and that you have access to it."
        )
    except ValueError as e:
        raise RuntimeError(
            f"The connector with ID {self.connector} linked "
            f"to the '{self.name}' {self.type} stack component could not "
            f"be correctly configured: {e}."
        )
    except AuthorizationException as e:
        raise RuntimeError(
            f"The connector with ID {self.connector} linked "
            f"to the '{self.name}' {self.type} stack component could not "
            f"be accessed due to an authorization error: {e}. Please "
            f"verify that you have access to the connector and try again."
        )

    return self._connector_instance

get_docker_builds(deployment: PipelineDeploymentBase) -> List[BuildConfiguration]

Gets the Docker builds required for the component.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentBase	The pipeline deployment for which to get the builds.	required

Returns:

Type	Description
List[BuildConfiguration]	The required Docker builds.

Source code in src/zenml/stack/stack_component.py

def get_docker_builds(
    self, deployment: "PipelineDeploymentBase"
) -> List["BuildConfiguration"]:
    """Gets the Docker builds required for the component.

    Args:
        deployment: The pipeline deployment for which to get the builds.

    Returns:
        The required Docker builds.
    """
    return []

get_pipeline_run_metadata(run_id: UUID) -> Dict[str, MetadataType]

Get general component-specific metadata for a pipeline run.

Parameters:

Name	Type	Description	Default
run_id	UUID	The ID of the pipeline run.	required

Returns:

Type	Description
Dict[str, MetadataType]	A dictionary of metadata.

Source code in src/zenml/stack/stack_component.py

def get_pipeline_run_metadata(
    self, run_id: UUID
) -> Dict[str, "MetadataType"]:
    """Get general component-specific metadata for a pipeline run.

    Args:
        run_id: The ID of the pipeline run.

    Returns:
        A dictionary of metadata.
    """
    return {}

get_settings(container: Union[Step, StepRunResponse, StepRunInfo, PipelineDeploymentBase, PipelineDeploymentResponse, PipelineRunResponse]) -> BaseSettings

Gets settings for this stack component.

This will return None if the stack component doesn't specify a settings class or the container doesn't contain runtime options for this component.

Parameters:

Name	Type	Description	Default
container	Union[Step, StepRunResponse, StepRunInfo, PipelineDeploymentBase, PipelineDeploymentResponse, PipelineRunResponse]	The Step, StepRunInfo or PipelineDeployment from which to get the settings.	required

Returns:

Type	Description
BaseSettings	Settings for this stack component.

Raises:

Type	Description
RuntimeError	If the stack component does not specify a settings class.

Source code in src/zenml/stack/stack_component.py

def get_settings(
    self,
    container: Union[
        "Step",
        "StepRunResponse",
        "StepRunInfo",
        "PipelineDeploymentBase",
        "PipelineDeploymentResponse",
        "PipelineRunResponse",
    ],
) -> "BaseSettings":
    """Gets settings for this stack component.

    This will return `None` if the stack component doesn't specify a
    settings class or the container doesn't contain runtime
    options for this component.

    Args:
        container: The `Step`, `StepRunInfo` or `PipelineDeployment` from
            which to get the settings.

    Returns:
        Settings for this stack component.

    Raises:
        RuntimeError: If the stack component does not specify a settings
            class.
    """
    if not self.settings_class:
        raise RuntimeError(
            f"Unable to get settings for component {self} because this "
            "component does not have an associated settings class. "
            "Return a settings class from the `@settings_class` property "
            "and try again."
        )

    key = settings_utils.get_stack_component_setting_key(self)

    all_settings = (
        container.config.settings
        if isinstance(
            container,
            (Step, StepRunResponse, StepRunInfo, PipelineRunResponse),
        )
        else container.pipeline_configuration.settings
    )

    if key in all_settings:
        return self.settings_class.model_validate(dict(all_settings[key]))
    else:
        return self.settings_class()

get_step_run_metadata(info: StepRunInfo) -> Dict[str, MetadataType]

Get component- and step-specific metadata after a step ran.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that was executed.	required

Returns:

Type	Description
Dict[str, MetadataType]	A dictionary of metadata.

Source code in src/zenml/stack/stack_component.py

def get_step_run_metadata(
    self, info: "StepRunInfo"
) -> Dict[str, "MetadataType"]:
    """Get component- and step-specific metadata after a step ran.

    Args:
        info: Info about the step that was executed.

    Returns:
        A dictionary of metadata.
    """
    return {}

prepare_pipeline_deployment(deployment: PipelineDeploymentResponse, stack: Stack) -> None

Prepares deploying the pipeline.

This method gets called immediately before a pipeline is deployed. Subclasses should override it if they require runtime configuration options or if they need to run code before the pipeline deployment.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentResponse	The pipeline deployment configuration.	required
stack	Stack	The stack on which the pipeline will be deployed.	required

Source code in src/zenml/stack/stack_component.py

def prepare_pipeline_deployment(
    self,
    deployment: "PipelineDeploymentResponse",
    stack: "Stack",
) -> None:
    """Prepares deploying the pipeline.

    This method gets called immediately before a pipeline is deployed.
    Subclasses should override it if they require runtime configuration
    options or if they need to run code before the pipeline deployment.

    Args:
        deployment: The pipeline deployment configuration.
        stack: The stack on which the pipeline will be deployed.
    """

prepare_step_run(info: StepRunInfo) -> None

Prepares running a step.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that will be executed.	required

Source code in src/zenml/stack/stack_component.py

def prepare_step_run(self, info: "StepRunInfo") -> None:
    """Prepares running a step.

    Args:
        info: Info about the step that will be executed.
    """

StackComponentConfig(warn_about_plain_text_secrets: bool = False, **kwargs: Any)

Bases: BaseModel, ABC

Base class for all ZenML stack component configs.

Ensures that secret references don't clash with pydantic validation.

StackComponents allow the specification of all their string attributes using secret references of the form {{secret_name.key}}. This however is only possible when the stack component does not perform any explicit validation of this attribute using pydantic validators. If this were the case, the validation would run on the secret reference and would fail or in the worst case, modify the secret reference and lead to unexpected behavior. This method ensures that no attributes that require custom pydantic validation are set as secret references.

Parameters:

Name	Type	Description	Default
warn_about_plain_text_secrets	bool	If true, then warns about using plain-text secrets.	False
**kwargs	Any	Arguments to initialize this stack component.	{}

Raises:

Type	Description
ValueError	If an attribute that requires custom pydantic validation is passed as a secret reference, or if the name attribute was passed as a secret reference.

Source code in src/zenml/stack/stack_component.py

def __init__(
    self, warn_about_plain_text_secrets: bool = False, **kwargs: Any
) -> None:
    """Ensures that secret references don't clash with pydantic validation.

    StackComponents allow the specification of all their string attributes
    using secret references of the form `{{secret_name.key}}`. This however
    is only possible when the stack component does not perform any explicit
    validation of this attribute using pydantic validators. If this were
    the case, the validation would run on the secret reference and would
    fail or in the worst case, modify the secret reference and lead to
    unexpected behavior. This method ensures that no attributes that require
    custom pydantic validation are set as secret references.

    Args:
        warn_about_plain_text_secrets: If true, then warns about using
            plain-text secrets.
        **kwargs: Arguments to initialize this stack component.

    Raises:
        ValueError: If an attribute that requires custom pydantic validation
            is passed as a secret reference, or if the `name` attribute
            was passed as a secret reference.
    """
    for key, value in kwargs.items():
        try:
            field = self.__class__.model_fields[key]
        except KeyError:
            # Value for a private attribute or non-existing field, this
            # will fail during the upcoming pydantic validation
            continue

        if value is None:
            continue

        if not secret_utils.is_secret_reference(value):
            if (
                secret_utils.is_secret_field(field)
                and warn_about_plain_text_secrets
            ):
                logger.warning(
                    "You specified a plain-text value for the sensitive "
                    f"attribute `{key}` for a `{self.__class__.__name__}` "
                    "stack component. This is currently only a warning, "
                    "but future versions of ZenML will require you to pass "
                    "in sensitive information as secrets. Check out the "
                    "documentation on how to configure your stack "
                    "components with secrets here: "
                    "https://docs.zenml.io/getting-started/deploying-zenml/secret-management"
                )
            continue

        if pydantic_utils.has_validators(
            pydantic_class=self.__class__, field_name=key
        ):
            raise ValueError(
                f"Passing the stack component attribute `{key}` as a "
                "secret reference is not allowed as additional validation "
                "is required for this attribute."
            )

    super().__init__(**kwargs)

Attributes

is_local: bool property

Checks if this stack component is running locally.

Concrete stack component configuration classes should override this method to return True if the stack component is relying on local resources or capabilities (e.g. local filesystem, local database or other services).

Examples:

• Artifact Stores that store artifacts in the local filesystem
• Orchestrators that are connected to local orchestration runtime services (e.g. local Kubernetes clusters, Docker containers etc).

Returns:

Type	Description
bool	True if this config is for a local component, False otherwise.

is_remote: bool property

Checks if this stack component is running remotely.

Concrete stack component configuration classes should override this method to return True if the stack component is running in a remote location, and it needs to access the ZenML database.

This designation is used to determine if the stack component can be used with a local ZenML database or if it requires a remote ZenML server.

Examples:

• Orchestrators that are running pipelines in the cloud or in a location other than the local host
• Step Operators that are running steps in the cloud or in a location other than the local host

Returns:

Type	Description
bool	True if this config is for a remote component, False otherwise.

is_valid: bool property

Checks if the stack component configurations are valid.

Concrete stack component configuration classes should override this method to return False if the stack component configurations are invalid.

Returns:

Type	Description
bool	True if the stack component config is valid, False otherwise.

required_secrets: Set[secret_utils.SecretReference] property

All required secrets for this stack component.

Returns:

Type	Description
Set[SecretReference]	The required secrets of this stack component.

Functions

StackValidator(required_components: Optional[AbstractSet[StackComponentType]] = None, custom_validation_function: Optional[Callable[[Stack], Tuple[bool, str]]] = None)

A StackValidator is used to validate a stack configuration.

Each StackComponent can provide a StackValidator to make sure it is compatible with all components of the stack. The KubeflowOrchestrator for example will always require the stack to have a container registry in order to push the docker images that are required to run a pipeline in Kubeflow Pipelines.

Initializes a StackValidator instance.

Parameters:

Name	Type	Description	Default
required_components	Optional[AbstractSet[StackComponentType]]	Optional set of stack components that must exist in the stack.	None
custom_validation_function	Optional[Callable[[Stack], Tuple[bool, str]]]	Optional function that returns whether a stack is valid and an error message to show if not valid.	None

Source code in src/zenml/stack/stack_validator.py

def __init__(
    self,
    required_components: Optional[AbstractSet[StackComponentType]] = None,
    custom_validation_function: Optional[
        Callable[["Stack"], Tuple[bool, str]]
    ] = None,
):
    """Initializes a `StackValidator` instance.

    Args:
        required_components: Optional set of stack components that must
            exist in the stack.
        custom_validation_function: Optional function that returns whether
            a stack is valid and an error message to show if not valid.
    """
    self._required_components = required_components or set()
    self._custom_validation_function = custom_validation_function

Functions

validate(stack: Stack) -> None

Validates the given stack.

Checks if the stack contains all the required components and passes the custom validation function of the validator.

Parameters:

Name	Type	Description	Default
stack	Stack	The stack to validate.	required

Raises:

Type	Description
StackValidationError	If the stack does not meet all the validation criteria.

Source code in src/zenml/stack/stack_validator.py

def validate(self, stack: "Stack") -> None:
    """Validates the given stack.

    Checks if the stack contains all the required components and passes
    the custom validation function of the validator.

    Args:
        stack: The stack to validate.

    Raises:
        StackValidationError: If the stack does not meet all the
            validation criteria.
    """
    missing_components = self._required_components - set(stack.components)
    if missing_components:
        raise StackValidationError(
            f"Missing stack components {missing_components} for "
            f"stack: {stack.name}"
        )

    if self._custom_validation_function:
        valid, err_msg = self._custom_validation_function(stack)
        if not valid:
            raise StackValidationError(
                f"Custom validation function failed to validate "
                f"stack '{stack.name}': {err_msg}"
            )

Modules

authentication_mixin

Stack component mixin for authentication.

Classes

AuthenticationConfigMixin(warn_about_plain_text_secrets: bool = False, **kwargs: Any)

Bases: StackComponentConfig

Base config for authentication mixins.

Any stack component that implements AuthenticationMixin should have a config that inherits from this class.

Attributes:

Name	Type	Description
authentication_secret	Optional[str]	Name of the secret that stores the authentication credentials.

Source code in src/zenml/stack/stack_component.py

def __init__(
    self, warn_about_plain_text_secrets: bool = False, **kwargs: Any
) -> None:
    """Ensures that secret references don't clash with pydantic validation.

    StackComponents allow the specification of all their string attributes
    using secret references of the form `{{secret_name.key}}`. This however
    is only possible when the stack component does not perform any explicit
    validation of this attribute using pydantic validators. If this were
    the case, the validation would run on the secret reference and would
    fail or in the worst case, modify the secret reference and lead to
    unexpected behavior. This method ensures that no attributes that require
    custom pydantic validation are set as secret references.

    Args:
        warn_about_plain_text_secrets: If true, then warns about using
            plain-text secrets.
        **kwargs: Arguments to initialize this stack component.

    Raises:
        ValueError: If an attribute that requires custom pydantic validation
            is passed as a secret reference, or if the `name` attribute
            was passed as a secret reference.
    """
    for key, value in kwargs.items():
        try:
            field = self.__class__.model_fields[key]
        except KeyError:
            # Value for a private attribute or non-existing field, this
            # will fail during the upcoming pydantic validation
            continue

        if value is None:
            continue

        if not secret_utils.is_secret_reference(value):
            if (
                secret_utils.is_secret_field(field)
                and warn_about_plain_text_secrets
            ):
                logger.warning(
                    "You specified a plain-text value for the sensitive "
                    f"attribute `{key}` for a `{self.__class__.__name__}` "
                    "stack component. This is currently only a warning, "
                    "but future versions of ZenML will require you to pass "
                    "in sensitive information as secrets. Check out the "
                    "documentation on how to configure your stack "
                    "components with secrets here: "
                    "https://docs.zenml.io/getting-started/deploying-zenml/secret-management"
                )
            continue

        if pydantic_utils.has_validators(
            pydantic_class=self.__class__, field_name=key
        ):
            raise ValueError(
                f"Passing the stack component attribute `{key}` as a "
                "secret reference is not allowed as additional validation "
                "is required for this attribute."
            )

    super().__init__(**kwargs)

AuthenticationMixin(name: str, id: UUID, config: StackComponentConfig, flavor: str, type: StackComponentType, user: Optional[UUID], created: datetime, updated: datetime, labels: Optional[Dict[str, Any]] = None, connector_requirements: Optional[ServiceConnectorRequirements] = None, connector: Optional[UUID] = None, connector_resource_id: Optional[str] = None, *args: Any, **kwargs: Any)

Bases: StackComponent

Stack component mixin for authentication.

Any stack component that implements this mixin should have a config that inherits from AuthenticationConfigMixin.

Source code in src/zenml/stack/stack_component.py

def __init__(
    self,
    name: str,
    id: UUID,
    config: StackComponentConfig,
    flavor: str,
    type: StackComponentType,
    user: Optional[UUID],
    created: datetime,
    updated: datetime,
    labels: Optional[Dict[str, Any]] = None,
    connector_requirements: Optional[ServiceConnectorRequirements] = None,
    connector: Optional[UUID] = None,
    connector_resource_id: Optional[str] = None,
    *args: Any,
    **kwargs: Any,
):
    """Initializes a StackComponent.

    Args:
        name: The name of the component.
        id: The unique ID of the component.
        config: The config of the component.
        flavor: The flavor of the component.
        type: The type of the component.
        user: The ID of the user who created the component.
        created: The creation time of the component.
        updated: The last update time of the component.
        labels: The labels of the component.
        connector_requirements: The requirements for the connector.
        connector: The ID of a connector linked to the component.
        connector_resource_id: The custom resource ID to access through
            the connector.
        *args: Additional positional arguments.
        **kwargs: Additional keyword arguments.

    Raises:
        ValueError: If a secret reference is passed as name.
    """
    if secret_utils.is_secret_reference(name):
        raise ValueError(
            "Passing the `name` attribute of a stack component as a "
            "secret reference is not allowed."
        )

    self.id = id
    self.name = name
    self._config = config
    self.flavor = flavor
    self.type = type
    self.user = user
    self.created = created
    self.updated = updated
    self.labels = labels
    self.connector_requirements = connector_requirements
    self.connector = connector
    self.connector_resource_id = connector_resource_id
    self._connector_instance: Optional[ServiceConnector] = None

Attributes

config: AuthenticationConfigMixin property

Returns the AuthenticationConfigMixin config.

Returns:

Type	Description
AuthenticationConfigMixin	The configuration.

Functions

get_authentication_secret() -> Optional[SecretResponse]

Gets the secret referred to by the authentication secret attribute.

Returns:

Type	Description
Optional[SecretResponse]	The secret if the authentication_secret attribute is set,
Optional[SecretResponse]	None otherwise.

Raises:

Type	Description
KeyError	If the secret does not exist.

Source code in src/zenml/stack/authentication_mixin.py

def get_authentication_secret(
    self,
) -> Optional[SecretResponse]:
    """Gets the secret referred to by the authentication secret attribute.

    Returns:
        The secret if the `authentication_secret` attribute is set,
        `None` otherwise.

    Raises:
        KeyError: If the secret does not exist.
    """
    if not self.config.authentication_secret:
        return None

    # Try to resolve the secret using the secret store
    try:
        return Client().get_secret_by_name_and_private_status(
            name=self.config.authentication_secret,
        )
    except (KeyError, NotImplementedError):
        raise KeyError(
            f"The authentication secret {self.config.authentication_secret} "
            f"referenced by the `{self.name}` `{self.type}` stack "
            "component does not exist."
        )

get_typed_authentication_secret(expected_schema_type: Type[T]) -> Optional[T]

Gets a typed secret referred to by the authentication secret attribute.

Parameters:

Name	Type	Description	Default
expected_schema_type	Type[T]	A Pydantic model class that represents the expected schema type of the secret.	required

Returns:

Type	Description
Optional[T]	The secret values extracted from the secret and converted into the
Optional[T]	indicated Pydantic type, if the authentication_secret attribute is
Optional[T]	set, None otherwise.

Raises:

Type	Description
TypeError	If the secret cannot be converted into the indicated Pydantic type.

Source code in src/zenml/stack/authentication_mixin.py

def get_typed_authentication_secret(
    self, expected_schema_type: Type[T]
) -> Optional[T]:
    """Gets a typed secret referred to by the authentication secret attribute.

    Args:
        expected_schema_type: A Pydantic model class that represents the
            expected schema type of the secret.

    Returns:
        The secret values extracted from the secret and converted into the
        indicated Pydantic type, if the `authentication_secret` attribute is
        set, `None` otherwise.

    Raises:
        TypeError: If the secret cannot be converted into the indicated
            Pydantic type.
    """
    secret = self.get_authentication_secret()

    if not secret:
        return None

    try:
        typed_secret = expected_schema_type(
            **secret.secret_values,
        )
    except (TypeError, ValueError) as e:
        raise TypeError(
            f"Authentication secret `{self.config.authentication_secret}` "
            f"referenced by the `{self.name}` `{self.type}` stack component"
            f"could not be converted to {expected_schema_type}: {e}"
        )

    return typed_secret

flavor

Base ZenML Flavor implementation.

Classes

Flavor

Class for ZenML Flavors.

Attributes

config_class: Type[StackComponentConfig] abstractmethod property

Returns StackComponentConfig config class.

Returns:

Type	Description
Type[StackComponentConfig]	The config class.

config_schema: Dict[str, Any] property

The config schema for a flavor.

Returns:

Type	Description
Dict[str, Any]	The config schema.

docs_url: Optional[str] property

A url to point at docs explaining this flavor.

Returns:

Type	Description
Optional[str]	A flavor docs url.

implementation_class: Type[StackComponent] abstractmethod property

Implementation class for this flavor.

Returns:

Type	Description
Type[StackComponent]	The implementation class for this flavor.

logo_url: Optional[str] property

A url to represent the flavor in the dashboard.

Returns:

Type	Description
Optional[str]	The flavor logo.

name: str abstractmethod property

The flavor name.

Returns:

Type	Description
str	The flavor name.

sdk_docs_url: Optional[str] property

A url to point at SDK docs explaining this flavor.

Returns:

Type	Description
Optional[str]	A flavor SDK docs url.

service_connector_requirements: Optional[ServiceConnectorRequirements] property

Service connector resource requirements for service connectors.

Specifies resource requirements that are used to filter the available service connector types that are compatible with this flavor.

Returns:

Type	Description
Optional[ServiceConnectorRequirements]	Requirements for compatible service connectors, if a service
Optional[ServiceConnectorRequirements]	connector is required for this flavor.

type: StackComponentType abstractmethod property

The stack component type.

Returns:

Type	Description
StackComponentType	The stack component type.

Functions

from_model(flavor_model: FlavorResponse) -> Flavor classmethod

Loads a flavor from a model.

Parameters:

Name	Type	Description	Default
flavor_model	FlavorResponse	The model to load from.	required

Raises:

Type	Description
CustomFlavorImportError	If the custom flavor can't be imported.
ImportError	If the flavor can't be imported.

Returns:

Type	Description
Flavor	The loaded flavor.

Source code in src/zenml/stack/flavor.py

@classmethod
def from_model(cls, flavor_model: FlavorResponse) -> "Flavor":
    """Loads a flavor from a model.

    Args:
        flavor_model: The model to load from.

    Raises:
        CustomFlavorImportError: If the custom flavor can't be imported.
        ImportError: If the flavor can't be imported.

    Returns:
        The loaded flavor.
    """
    try:
        flavor = source_utils.load(flavor_model.source)()
    except (ModuleNotFoundError, ImportError, NotImplementedError) as err:
        if flavor_model.is_custom:
            flavor_module, _ = flavor_model.source.rsplit(".", maxsplit=1)
            expected_file_path = os.path.join(
                source_utils.get_source_root(),
                flavor_module.replace(".", os.path.sep),
            )
            raise CustomFlavorImportError(
                f"Couldn't import custom flavor {flavor_model.name}: "
                f"{err}. Make sure the custom flavor class "
                f"`{flavor_model.source}` is importable. If it is part of "
                "a library, make sure it is installed. If "
                "it is a local code file, make sure it exists at "
                f"`{expected_file_path}.py`."
            )
        else:
            raise ImportError(
                f"Couldn't import flavor {flavor_model.name}: {err}"
            )
    return cast(Flavor, flavor)

generate_default_docs_url() -> str

Generate the doc urls for all inbuilt and integration flavors.

Note that this method is not going to be useful for custom flavors, which do not have any docs in the main zenml docs.

Returns:

Type	Description
str	The complete url to the zenml documentation

Source code in src/zenml/stack/flavor.py

def generate_default_docs_url(self) -> str:
    """Generate the doc urls for all inbuilt and integration flavors.

    Note that this method is not going to be useful for custom flavors,
    which do not have any docs in the main zenml docs.

    Returns:
        The complete url to the zenml documentation
    """
    from zenml import __version__

    component_type = self.type.plural.replace("_", "-")
    name = self.name.replace("_", "-")

    try:
        is_latest = is_latest_zenml_version()
    except RuntimeError:
        # We assume in error cases that we are on the latest version
        is_latest = True

    if is_latest:
        base = "https://docs.zenml.io"
    else:
        base = f"https://zenml-io.gitbook.io/zenml-legacy-documentation/v/{__version__}"
    return f"{base}/stack-components/{component_type}/{name}"

generate_default_sdk_docs_url() -> str

Generate SDK docs url for a flavor.

Returns:

Type	Description
str	The complete url to the zenml SDK docs

Source code in src/zenml/stack/flavor.py

def generate_default_sdk_docs_url(self) -> str:
    """Generate SDK docs url for a flavor.

    Returns:
        The complete url to the zenml SDK docs
    """
    from zenml import __version__

    base = f"https://sdkdocs.zenml.io/{__version__}"

    component_type = self.type.plural

    if "zenml.integrations" in self.__module__:
        # Get integration name out of module path which will look something
        #  like this "zenml.integrations.<integration>....
        integration = self.__module__.split(
            "zenml.integrations.", maxsplit=1
        )[1].split(".")[0]

        return (
            f"{base}/integration_code_docs"
            f"/integrations-{integration}/#{self.__module__}"
        )

    else:
        return (
            f"{base}/core_code_docs/core-{component_type}/"
            f"#{self.__module__}"
        )

to_model(integration: Optional[str] = None, is_custom: bool = True) -> FlavorRequest

Converts a flavor to a model.

Parameters:

Name	Type	Description	Default
integration	Optional[str]	The integration to use for the model.	None
is_custom	bool	Whether the flavor is a custom flavor.	True

Returns:

Type	Description
FlavorRequest	The model.

Source code in src/zenml/stack/flavor.py

def to_model(
    self,
    integration: Optional[str] = None,
    is_custom: bool = True,
) -> FlavorRequest:
    """Converts a flavor to a model.

    Args:
        integration: The integration to use for the model.
        is_custom: Whether the flavor is a custom flavor.

    Returns:
        The model.
    """
    connector_requirements = self.service_connector_requirements
    connector_type = (
        connector_requirements.connector_type
        if connector_requirements
        else None
    )
    resource_type = (
        connector_requirements.resource_type
        if connector_requirements
        else None
    )
    resource_id_attr = (
        connector_requirements.resource_id_attr
        if connector_requirements
        else None
    )

    model = FlavorRequest(
        name=self.name,
        type=self.type,
        source=source_utils.resolve(self.__class__).import_path,
        config_schema=self.config_schema,
        connector_type=connector_type,
        connector_resource_type=resource_type,
        connector_resource_id_attr=resource_id_attr,
        integration=integration,
        logo_url=self.logo_url,
        docs_url=self.docs_url,
        sdk_docs_url=self.sdk_docs_url,
        is_custom=is_custom,
    )
    return model

Functions

validate_flavor_source(source: str, component_type: StackComponentType) -> Type[Flavor]

Import a StackComponent class from a given source and validate its type.

Parameters:

Name	Type	Description	Default
source	str	source path of the implementation	required
component_type	StackComponentType	the type of the stack component	required

Returns:

Type	Description
Type[Flavor]	the imported class

Raises:

Type	Description
ValueError	If ZenML cannot find the given module path
TypeError	If the given module path does not point to a subclass of a StackComponent which has the right component type.

Source code in src/zenml/stack/flavor.py

def validate_flavor_source(
    source: str, component_type: StackComponentType
) -> Type["Flavor"]:
    """Import a StackComponent class from a given source and validate its type.

    Args:
        source: source path of the implementation
        component_type: the type of the stack component

    Returns:
        the imported class

    Raises:
        ValueError: If ZenML cannot find the given module path
        TypeError: If the given module path does not point to a subclass of a
            StackComponent which has the right component type.
    """
    from zenml.stack.stack_component import (
        StackComponent,
        StackComponentConfig,
    )
    from zenml.utils import source_utils

    try:
        flavor_class = source_utils.load(source)
    except (ValueError, AttributeError, ImportError) as e:
        raise ValueError(
            f"ZenML can not import the flavor class '{source}': {e}"
        )

    if not (
        isinstance(flavor_class, type) and issubclass(flavor_class, Flavor)
    ):
        raise TypeError(
            f"The source '{source}' does not point to a subclass of the ZenML"
            f"Flavor."
        )

    flavor = flavor_class()
    try:
        impl_class = flavor.implementation_class
    except (ModuleNotFoundError, ImportError, NotImplementedError) as e:
        raise ValueError(
            f"The implementation class defined within the "
            f"'{flavor_class.__name__}' can not be imported: {e}"
        )

    if not issubclass(impl_class, StackComponent):
        raise TypeError(
            f"The implementation class '{impl_class.__name__}' of a flavor "
            f"needs to be a subclass of the ZenML StackComponent."
        )

    if flavor.type != component_type:  # noqa
        raise TypeError(
            f"The source points to a {impl_class.type}, not a "  # noqa
            f"{component_type}."
        )

    try:
        conf_class = flavor.config_class
    except (ModuleNotFoundError, ImportError, NotImplementedError) as e:
        raise ValueError(
            f"The config class defined within the "
            f"'{flavor_class.__name__}' can not be imported: {e}"
        )

    if not issubclass(conf_class, StackComponentConfig):
        raise TypeError(
            f"The config class '{conf_class.__name__}' of a flavor "
            f"needs to be a subclass of the ZenML StackComponentConfig."
        )

    return flavor_class

Modules

flavor_registry

Implementation of the ZenML flavor registry.

Classes

FlavorRegistry()

Registry for stack component flavors.

The flavors defined by ZenML must be registered here.

Initialization of the flavors.

Source code in src/zenml/stack/flavor_registry.py

def __init__(self) -> None:
    """Initialization of the flavors."""
    self._flavors: DefaultDict[
        StackComponentType, Dict[str, FlavorResponse]
    ] = defaultdict(dict)

Attributes

builtin_flavors: List[Type[Flavor]] property

A list of all default in-built flavors.

Returns:

Type	Description
List[Type[Flavor]]	A list of builtin flavors.

integration_flavors: List[Type[Flavor]] property

A list of all default integration flavors.

Returns:

Type	Description
List[Type[Flavor]]	A list of integration flavors.

Functions

register_builtin_flavors(store: BaseZenStore) -> None

Registers the default built-in flavors.

Parameters:

Name	Type	Description	Default
store	BaseZenStore	The instance of the zen_store to use	required

Source code in src/zenml/stack/flavor_registry.py

def register_builtin_flavors(self, store: BaseZenStore) -> None:
    """Registers the default built-in flavors.

    Args:
        store: The instance of the zen_store to use
    """
    with analytics_disabler():
        for flavor in self.builtin_flavors:
            flavor_request_model = flavor().to_model(
                integration="built-in",
                is_custom=False,
            )
            existing_flavor = store.list_flavors(
                FlavorFilter(
                    name=flavor_request_model.name,
                    type=flavor_request_model.type,
                )
            )

            if len(existing_flavor) == 0:
                store.create_flavor(flavor_request_model)
            else:
                flavor_update_model = FlavorUpdate.model_validate(
                    dict(flavor_request_model)
                )
                store.update_flavor(
                    existing_flavor[0].id, flavor_update_model
                )

register_flavors(store: BaseZenStore) -> None

Register all flavors to the DB.

Parameters:

Name	Type	Description	Default
store	BaseZenStore	The instance of a store to use for persistence	required

Source code in src/zenml/stack/flavor_registry.py

def register_flavors(self, store: BaseZenStore) -> None:
    """Register all flavors to the DB.

    Args:
        store: The instance of a store to use for persistence
    """
    self.register_builtin_flavors(store=store)
    self.register_integration_flavors(store=store)

register_integration_flavors(store: BaseZenStore) -> None staticmethod

Registers the flavors implemented by integrations.

Parameters:

Name	Type	Description	Default
store	BaseZenStore	The instance of the zen_store to use	required

Source code in src/zenml/stack/flavor_registry.py

@staticmethod
def register_integration_flavors(store: BaseZenStore) -> None:
    """Registers the flavors implemented by integrations.

    Args:
        store: The instance of the zen_store to use
    """
    with analytics_disabler():
        for name, integration in integration_registry.integrations.items():
            try:
                integrated_flavors = integration.flavors()
                for flavor in integrated_flavors:
                    flavor_request_model = flavor().to_model(
                        integration=name,
                        is_custom=False,
                    )
                    existing_flavor = store.list_flavors(
                        FlavorFilter(
                            name=flavor_request_model.name,
                            type=flavor_request_model.type,
                        )
                    )
                    if len(existing_flavor) == 0:
                        store.create_flavor(flavor_request_model)
                    else:
                        flavor_update_model = FlavorUpdate.model_validate(
                            dict(flavor_request_model)
                        )
                        store.update_flavor(
                            existing_flavor[0].id, flavor_update_model
                        )
            except Exception as e:
                logger.warning(
                    f"Integration {name} failed to register flavors. "
                    f"Error: {e}"
                )

Functions

stack

Implementation of the ZenML Stack class.

Classes

Stack(id: UUID, name: str, *, orchestrator: BaseOrchestrator, artifact_store: BaseArtifactStore, container_registry: Optional[BaseContainerRegistry] = None, step_operator: Optional[BaseStepOperator] = None, feature_store: Optional[BaseFeatureStore] = None, model_deployer: Optional[BaseModelDeployer] = None, experiment_tracker: Optional[BaseExperimentTracker] = None, alerter: Optional[BaseAlerter] = None, annotator: Optional[BaseAnnotator] = None, data_validator: Optional[BaseDataValidator] = None, image_builder: Optional[BaseImageBuilder] = None, model_registry: Optional[BaseModelRegistry] = None)

ZenML stack class.

A ZenML stack is a collection of multiple stack components that are required to run ZenML pipelines. Some of these components (orchestrator, and artifact store) are required to run any kind of pipeline, other components like the container registry are only required if other stack components depend on them.

Initializes and validates a stack instance.

Parameters:

Name	Type	Description	Default
id	UUID	Unique ID of the stack.	required
name	str	Name of the stack.	required
orchestrator	BaseOrchestrator	Orchestrator component of the stack.	required
artifact_store	BaseArtifactStore	Artifact store component of the stack.	required
container_registry	Optional[BaseContainerRegistry]	Container registry component of the stack.	None
step_operator	Optional[BaseStepOperator]	Step operator component of the stack.	None
feature_store	Optional[BaseFeatureStore]	Feature store component of the stack.	None
model_deployer	Optional[BaseModelDeployer]	Model deployer component of the stack.	None
experiment_tracker	Optional[BaseExperimentTracker]	Experiment tracker component of the stack.	None
alerter	Optional[BaseAlerter]	Alerter component of the stack.	None
annotator	Optional[BaseAnnotator]	Annotator component of the stack.	None
data_validator	Optional[BaseDataValidator]	Data validator component of the stack.	None
image_builder	Optional[BaseImageBuilder]	Image builder component of the stack.	None
model_registry	Optional[BaseModelRegistry]	Model registry component of the stack.	None

Source code in src/zenml/stack/stack.py

def __init__(
    self,
    id: UUID,
    name: str,
    *,
    orchestrator: "BaseOrchestrator",
    artifact_store: "BaseArtifactStore",
    container_registry: Optional["BaseContainerRegistry"] = None,
    step_operator: Optional["BaseStepOperator"] = None,
    feature_store: Optional["BaseFeatureStore"] = None,
    model_deployer: Optional["BaseModelDeployer"] = None,
    experiment_tracker: Optional["BaseExperimentTracker"] = None,
    alerter: Optional["BaseAlerter"] = None,
    annotator: Optional["BaseAnnotator"] = None,
    data_validator: Optional["BaseDataValidator"] = None,
    image_builder: Optional["BaseImageBuilder"] = None,
    model_registry: Optional["BaseModelRegistry"] = None,
):
    """Initializes and validates a stack instance.

    Args:
        id: Unique ID of the stack.
        name: Name of the stack.
        orchestrator: Orchestrator component of the stack.
        artifact_store: Artifact store component of the stack.
        container_registry: Container registry component of the stack.
        step_operator: Step operator component of the stack.
        feature_store: Feature store component of the stack.
        model_deployer: Model deployer component of the stack.
        experiment_tracker: Experiment tracker component of the stack.
        alerter: Alerter component of the stack.
        annotator: Annotator component of the stack.
        data_validator: Data validator component of the stack.
        image_builder: Image builder component of the stack.
        model_registry: Model registry component of the stack.
    """
    self._id = id
    self._name = name
    self._orchestrator = orchestrator
    self._artifact_store = artifact_store
    self._container_registry = container_registry
    self._step_operator = step_operator
    self._feature_store = feature_store
    self._model_deployer = model_deployer
    self._experiment_tracker = experiment_tracker
    self._alerter = alerter
    self._annotator = annotator
    self._data_validator = data_validator
    self._model_registry = model_registry
    self._image_builder = image_builder

Attributes

alerter: Optional[BaseAlerter] property

The alerter of the stack.

Returns:

Type	Description
Optional[BaseAlerter]	The alerter of the stack.

annotator: Optional[BaseAnnotator] property

The annotator of the stack.

Returns:

Type	Description
Optional[BaseAnnotator]	The annotator of the stack.

apt_packages: List[str] property

List of APT package requirements for the stack.

Returns:

Type	Description
List[str]	A list of APT package requirements for the stack.

artifact_store: BaseArtifactStore property

The artifact store of the stack.

Returns:

Type	Description
BaseArtifactStore	The artifact store of the stack.

components: Dict[StackComponentType, StackComponent] property

All components of the stack.

Returns:

Type	Description
Dict[StackComponentType, StackComponent]	A dictionary of all components of the stack.

container_registry: Optional[BaseContainerRegistry] property

The container registry of the stack.

Returns:

Type	Description
Optional[BaseContainerRegistry]	The container registry of the stack or None if the stack does not
Optional[BaseContainerRegistry]	have a container registry.

data_validator: Optional[BaseDataValidator] property

The data validator of the stack.

Returns:

Type	Description
Optional[BaseDataValidator]	The data validator of the stack.

experiment_tracker: Optional[BaseExperimentTracker] property

The experiment tracker of the stack.

Returns:

Type	Description
Optional[BaseExperimentTracker]	The experiment tracker of the stack.

feature_store: Optional[BaseFeatureStore] property

The feature store of the stack.

Returns:

Type	Description
Optional[BaseFeatureStore]	The feature store of the stack.

id: UUID property

The ID of the stack.

Returns:

Type	Description
UUID	The ID of the stack.

image_builder: Optional[BaseImageBuilder] property

The image builder of the stack.

Returns:

Type	Description
Optional[BaseImageBuilder]	The image builder of the stack.

model_deployer: Optional[BaseModelDeployer] property

The model deployer of the stack.

Returns:

Type	Description
Optional[BaseModelDeployer]	The model deployer of the stack.

model_registry: Optional[BaseModelRegistry] property

The model registry of the stack.

Returns:

Type	Description
Optional[BaseModelRegistry]	The model registry of the stack.

name: str property

The name of the stack.

Returns:

Name	Type	Description
str	str	The name of the stack.

orchestrator: BaseOrchestrator property

The orchestrator of the stack.

Returns:

Type	Description
BaseOrchestrator	The orchestrator of the stack.

required_secrets: Set[secret_utils.SecretReference] property

All required secrets for this stack.

Returns:

Type	Description
Set[SecretReference]	The required secrets of this stack.

requires_remote_server: bool property

If the stack requires a remote ZenServer to run.

This is the case if any code is getting executed remotely. This is the case for both remote orchestrators as well as remote step operators.

Returns:

Type	Description
bool	If the stack requires a remote ZenServer to run.

setting_classes: Dict[str, Type[BaseSettings]] property

Setting classes of all components of this stack.

Returns:

Type	Description
Dict[str, Type[BaseSettings]]	All setting classes and their respective keys.

step_operator: Optional[BaseStepOperator] property

The step operator of the stack.

Returns:

Type	Description
Optional[BaseStepOperator]	The step operator of the stack.

Functions

check_local_paths() -> bool

Checks if the stack has local paths.

Returns:

Type	Description
bool	True if the stack has local paths, False otherwise.

Raises:

Type	Description
ValueError	If the stack has local paths that do not conform to the convention that all local path must be relative to the local stores directory.

Source code in src/zenml/stack/stack.py

def check_local_paths(self) -> bool:
    """Checks if the stack has local paths.

    Returns:
        True if the stack has local paths, False otherwise.

    Raises:
        ValueError: If the stack has local paths that do not conform to
            the convention that all local path must be relative to the
            local stores directory.
    """
    from zenml.config.global_config import GlobalConfiguration

    local_stores_path = GlobalConfiguration().local_stores_path

    # go through all stack components and identify those that advertise
    # a local path where they persist information that they need to be
    # available when running pipelines.
    has_local_paths = False
    for stack_comp in self.components.values():
        local_path = stack_comp.local_path
        if not local_path:
            continue
        # double-check this convention, just in case it wasn't respected
        # as documented in `StackComponent.local_path`
        if not local_path.startswith(local_stores_path):
            raise ValueError(
                f"Local path {local_path} for component "
                f"{stack_comp.name} is not in the local stores "
                f"directory ({local_stores_path})."
            )
        has_local_paths = True

    return has_local_paths

cleanup_step_run(info: StepRunInfo, step_failed: bool) -> None

Cleans up resources after the step run is finished.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that was executed.	required
step_failed	bool	Whether the step failed.	required

Source code in src/zenml/stack/stack.py

def cleanup_step_run(self, info: "StepRunInfo", step_failed: bool) -> None:
    """Cleans up resources after the step run is finished.

    Args:
        info: Info about the step that was executed.
        step_failed: Whether the step failed.
    """
    for component in self._get_active_components_for_step(
        info.config
    ).values():
        component.cleanup_step_run(info=info, step_failed=step_failed)

deploy_pipeline(deployment: PipelineDeploymentResponse, placeholder_run: Optional[PipelineRunResponse] = None) -> None

Deploys a pipeline on this stack.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentResponse	The pipeline deployment.	required
placeholder_run	Optional[PipelineRunResponse]	An optional placeholder run for the deployment.	None

Source code in src/zenml/stack/stack.py

def deploy_pipeline(
    self,
    deployment: "PipelineDeploymentResponse",
    placeholder_run: Optional["PipelineRunResponse"] = None,
) -> None:
    """Deploys a pipeline on this stack.

    Args:
        deployment: The pipeline deployment.
        placeholder_run: An optional placeholder run for the deployment.
    """
    self.orchestrator.run(
        deployment=deployment, stack=self, placeholder_run=placeholder_run
    )

dict() -> Dict[str, str]

Converts the stack into a dictionary.

Returns:

Type	Description
Dict[str, str]	A dictionary containing the stack components.

Source code in src/zenml/stack/stack.py

def dict(self) -> Dict[str, str]:
    """Converts the stack into a dictionary.

    Returns:
        A dictionary containing the stack components.
    """
    component_dict = {
        component_type.value: json.dumps(
            component.config.model_dump(mode="json"), sort_keys=True
        )
        for component_type, component in self.components.items()
    }
    component_dict.update({"name": self.name})
    return component_dict

from_components(id: UUID, name: str, components: Dict[StackComponentType, StackComponent]) -> Stack classmethod

Creates a stack instance from a dict of stack components.

noqa: DAR402

Parameters:

Name	Type	Description	Default
id	UUID	Unique ID of the stack.	required
name	str	The name of the stack.	required
components	Dict[StackComponentType, StackComponent]	The components of the stack.	required

Returns:

Type	Description
Stack	A stack instance consisting of the given components.

Raises:

Type	Description
TypeError	If a required component is missing or a component doesn't inherit from the expected base class.

Source code in src/zenml/stack/stack.py

@classmethod
def from_components(
    cls,
    id: UUID,
    name: str,
    components: Dict[StackComponentType, "StackComponent"],
) -> "Stack":
    """Creates a stack instance from a dict of stack components.

    # noqa: DAR402

    Args:
        id: Unique ID of the stack.
        name: The name of the stack.
        components: The components of the stack.

    Returns:
        A stack instance consisting of the given components.

    Raises:
        TypeError: If a required component is missing or a component
            doesn't inherit from the expected base class.
    """
    from zenml.alerter import BaseAlerter
    from zenml.annotators import BaseAnnotator
    from zenml.artifact_stores import BaseArtifactStore
    from zenml.container_registries import BaseContainerRegistry
    from zenml.data_validators import BaseDataValidator
    from zenml.experiment_trackers import BaseExperimentTracker
    from zenml.feature_stores import BaseFeatureStore
    from zenml.image_builders import BaseImageBuilder
    from zenml.model_deployers import BaseModelDeployer
    from zenml.model_registries import BaseModelRegistry
    from zenml.orchestrators import BaseOrchestrator
    from zenml.step_operators import BaseStepOperator

    def _raise_type_error(
        component: Optional["StackComponent"], expected_class: Type[Any]
    ) -> NoReturn:
        """Raises a TypeError that the component has an unexpected type.

        Args:
            component: The component that has an unexpected type.
            expected_class: The expected type of the component.

        Raises:
            TypeError: If the component has an unexpected type.
        """
        raise TypeError(
            f"Unable to create stack: Wrong stack component type "
            f"`{component.__class__.__name__}` (expected: subclass "
            f"of `{expected_class.__name__}`)"
        )

    orchestrator = components.get(StackComponentType.ORCHESTRATOR)
    if not isinstance(orchestrator, BaseOrchestrator):
        _raise_type_error(orchestrator, BaseOrchestrator)

    artifact_store = components.get(StackComponentType.ARTIFACT_STORE)
    if not isinstance(artifact_store, BaseArtifactStore):
        _raise_type_error(artifact_store, BaseArtifactStore)

    container_registry = components.get(
        StackComponentType.CONTAINER_REGISTRY
    )
    if container_registry is not None and not isinstance(
        container_registry, BaseContainerRegistry
    ):
        _raise_type_error(container_registry, BaseContainerRegistry)

    step_operator = components.get(StackComponentType.STEP_OPERATOR)
    if step_operator is not None and not isinstance(
        step_operator, BaseStepOperator
    ):
        _raise_type_error(step_operator, BaseStepOperator)

    feature_store = components.get(StackComponentType.FEATURE_STORE)
    if feature_store is not None and not isinstance(
        feature_store, BaseFeatureStore
    ):
        _raise_type_error(feature_store, BaseFeatureStore)

    model_deployer = components.get(StackComponentType.MODEL_DEPLOYER)
    if model_deployer is not None and not isinstance(
        model_deployer, BaseModelDeployer
    ):
        _raise_type_error(model_deployer, BaseModelDeployer)

    experiment_tracker = components.get(
        StackComponentType.EXPERIMENT_TRACKER
    )
    if experiment_tracker is not None and not isinstance(
        experiment_tracker, BaseExperimentTracker
    ):
        _raise_type_error(experiment_tracker, BaseExperimentTracker)

    alerter = components.get(StackComponentType.ALERTER)
    if alerter is not None and not isinstance(alerter, BaseAlerter):
        _raise_type_error(alerter, BaseAlerter)

    annotator = components.get(StackComponentType.ANNOTATOR)
    if annotator is not None and not isinstance(annotator, BaseAnnotator):
        _raise_type_error(annotator, BaseAnnotator)

    data_validator = components.get(StackComponentType.DATA_VALIDATOR)
    if data_validator is not None and not isinstance(
        data_validator, BaseDataValidator
    ):
        _raise_type_error(data_validator, BaseDataValidator)

    image_builder = components.get(StackComponentType.IMAGE_BUILDER)
    if image_builder is not None and not isinstance(
        image_builder, BaseImageBuilder
    ):
        _raise_type_error(image_builder, BaseImageBuilder)

    model_registry = components.get(StackComponentType.MODEL_REGISTRY)
    if model_registry is not None and not isinstance(
        model_registry, BaseModelRegistry
    ):
        _raise_type_error(model_registry, BaseModelRegistry)

    return Stack(
        id=id,
        name=name,
        orchestrator=orchestrator,
        artifact_store=artifact_store,
        container_registry=container_registry,
        step_operator=step_operator,
        feature_store=feature_store,
        model_deployer=model_deployer,
        experiment_tracker=experiment_tracker,
        alerter=alerter,
        annotator=annotator,
        data_validator=data_validator,
        image_builder=image_builder,
        model_registry=model_registry,
    )

from_model(stack_model: StackResponse) -> Stack classmethod

Creates a Stack instance from a StackModel.

Parameters:

Name	Type	Description	Default
stack_model	StackResponse	The StackModel to create the Stack from.	required

Returns:

Type	Description
Stack	The created Stack instance.

Source code in src/zenml/stack/stack.py

@classmethod
def from_model(cls, stack_model: "StackResponse") -> "Stack":
    """Creates a Stack instance from a StackModel.

    Args:
        stack_model: The StackModel to create the Stack from.

    Returns:
        The created Stack instance.
    """
    global _STACK_CACHE
    key = (stack_model.id, stack_model.updated)
    if key in _STACK_CACHE:
        return _STACK_CACHE[key]

    from zenml.stack import StackComponent

    # Run a hydrated list call once to avoid one request per component
    component_models = pagination_utils.depaginate(
        Client().list_stack_components,
        stack_id=stack_model.id,
        hydrate=True,
    )

    stack_components = {
        model.type: StackComponent.from_model(model)
        for model in component_models
    }
    stack = Stack.from_components(
        id=stack_model.id,
        name=stack_model.name,
        components=stack_components,
    )
    _STACK_CACHE[key] = stack

    client = Client()
    if stack_model.id == client.active_stack_model.id:
        if stack_model.updated > client.active_stack_model.updated:
            if client._config:
                client._config.set_active_stack(stack_model)
            else:
                GlobalConfiguration().set_active_stack(stack_model)

    return stack

get_docker_builds(deployment: PipelineDeploymentBase) -> List[BuildConfiguration]

Gets the Docker builds required for the stack.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentBase	The pipeline deployment for which to get the builds.	required

Returns:

Type	Description
List[BuildConfiguration]	The required Docker builds.

Source code in src/zenml/stack/stack.py

def get_docker_builds(
    self, deployment: "PipelineDeploymentBase"
) -> List["BuildConfiguration"]:
    """Gets the Docker builds required for the stack.

    Args:
        deployment: The pipeline deployment for which to get the builds.

    Returns:
        The required Docker builds.
    """
    return list(
        itertools.chain.from_iterable(
            component.get_docker_builds(deployment=deployment)
            for component in self.components.values()
        )
    )

get_pipeline_run_metadata(run_id: UUID) -> Dict[UUID, Dict[str, MetadataType]]

Get general component-specific metadata for a pipeline run.

Parameters:

Name	Type	Description	Default
run_id	UUID	ID of the pipeline run.	required

Returns:

Type	Description
Dict[UUID, Dict[str, MetadataType]]	A dictionary mapping component IDs to the metadata they created.

Source code in src/zenml/stack/stack.py

def get_pipeline_run_metadata(
    self, run_id: UUID
) -> Dict[UUID, Dict[str, MetadataType]]:
    """Get general component-specific metadata for a pipeline run.

    Args:
        run_id: ID of the pipeline run.

    Returns:
        A dictionary mapping component IDs to the metadata they created.
    """
    pipeline_run_metadata: Dict[UUID, Dict[str, MetadataType]] = {}
    for component in self.components.values():
        try:
            component_metadata = component.get_pipeline_run_metadata(
                run_id=run_id
            )
            if component_metadata:
                pipeline_run_metadata[component.id] = component_metadata
        except Exception as e:
            logger.warning(
                f"Extracting pipeline run metadata failed for component "
                f"'{component.name}' of type '{component.type}': {e}"
            )
    return pipeline_run_metadata

get_step_run_metadata(info: StepRunInfo) -> Dict[UUID, Dict[str, MetadataType]]

Get component-specific metadata for a step run.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that was executed.	required

Returns:

Type	Description
Dict[UUID, Dict[str, MetadataType]]	A dictionary mapping component IDs to the metadata they created.

Source code in src/zenml/stack/stack.py

def get_step_run_metadata(
    self, info: "StepRunInfo"
) -> Dict[UUID, Dict[str, MetadataType]]:
    """Get component-specific metadata for a step run.

    Args:
        info: Info about the step that was executed.

    Returns:
        A dictionary mapping component IDs to the metadata they created.
    """
    step_run_metadata: Dict[UUID, Dict[str, MetadataType]] = {}
    for component in self._get_active_components_for_step(
        info.config
    ).values():
        try:
            component_metadata = component.get_step_run_metadata(info=info)
            if component_metadata:
                step_run_metadata[component.id] = component_metadata
        except Exception as e:
            logger.warning(
                f"Extracting step run metadata failed for component "
                f"'{component.name}' of type '{component.type}': {e}"
            )
    return step_run_metadata

prepare_pipeline_deployment(deployment: PipelineDeploymentResponse) -> None

Prepares the stack for a pipeline deployment.

This method is called before a pipeline is deployed.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentResponse	The pipeline deployment	required

Raises:

Type	Description
RuntimeError	If trying to deploy a pipeline that requires a remote ZenML server with a local one.

Source code in src/zenml/stack/stack.py

def prepare_pipeline_deployment(
    self, deployment: "PipelineDeploymentResponse"
) -> None:
    """Prepares the stack for a pipeline deployment.

    This method is called before a pipeline is deployed.

    Args:
        deployment: The pipeline deployment

    Raises:
        RuntimeError: If trying to deploy a pipeline that requires a remote
            ZenML server with a local one.
    """
    self.validate(fail_if_secrets_missing=True)

    if self.requires_remote_server and Client().zen_store.is_local_store():
        raise RuntimeError(
            "Stacks with remote components such as remote orchestrators "
            "and step operators require a remote "
            "ZenML server. To run a pipeline with this stack you need to "
            "connect to a remote ZenML server first. Check out "
            "https://docs.zenml.io/getting-started/deploying-zenml "
            "for more information on how to deploy ZenML."
        )

    for component in self.components.values():
        component.prepare_pipeline_deployment(
            deployment=deployment, stack=self
        )

prepare_step_run(info: StepRunInfo) -> None

Prepares running a step.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that will be executed.	required

Source code in src/zenml/stack/stack.py

def prepare_step_run(self, info: "StepRunInfo") -> None:
    """Prepares running a step.

    Args:
        info: Info about the step that will be executed.
    """
    for component in self._get_active_components_for_step(
        info.config
    ).values():
        component.prepare_step_run(info=info)

requirements(exclude_components: Optional[AbstractSet[StackComponentType]] = None) -> Set[str]

Set of PyPI requirements for the stack.

This method combines the requirements of all stack components (except the ones specified in exclude_components).

Parameters:

Name	Type	Description	Default
exclude_components	Optional[AbstractSet[StackComponentType]]	Set of component types for which the requirements should not be included in the output.	None

Returns:

Type	Description
Set[str]	Set of PyPI requirements.

Source code in src/zenml/stack/stack.py

def requirements(
    self,
    exclude_components: Optional[AbstractSet[StackComponentType]] = None,
) -> Set[str]:
    """Set of PyPI requirements for the stack.

    This method combines the requirements of all stack components (except
    the ones specified in `exclude_components`).

    Args:
        exclude_components: Set of component types for which the
            requirements should not be included in the output.

    Returns:
        Set of PyPI requirements.
    """
    exclude_components = exclude_components or set()
    requirements = [
        component.requirements
        for component in self.components.values()
        if component.type not in exclude_components
    ]
    return set.union(*requirements) if requirements else set()

validate(fail_if_secrets_missing: bool = False) -> None

Checks whether the stack configuration is valid.

To check if a stack configuration is valid, the following criteria must be met: - the stack must have an image builder if other components require it - the StackValidator of each stack component has to validate the stack to make sure all the components are compatible with each other - the required secrets of all components need to exist

Parameters:

Name	Type	Description	Default
fail_if_secrets_missing	bool	If this is True, an error will be raised if a secret for a component is missing. Otherwise, only a warning will be logged.	False

Source code in src/zenml/stack/stack.py

def validate(
    self,
    fail_if_secrets_missing: bool = False,
) -> None:
    """Checks whether the stack configuration is valid.

    To check if a stack configuration is valid, the following criteria must
    be met:
    - the stack must have an image builder if other components require it
    - the `StackValidator` of each stack component has to validate the
        stack to make sure all the components are compatible with each other
    - the required secrets of all components need to exist

    Args:
        fail_if_secrets_missing: If this is `True`, an error will be raised
            if a secret for a component is missing. Otherwise, only a
            warning will be logged.
    """
    if handle_bool_env_var(ENV_ZENML_SKIP_STACK_VALIDATION, default=False):
        logger.debug("Skipping stack validation.")
        return

    self.validate_image_builder()
    for component in self.components.values():
        if component.validator:
            component.validator.validate(stack=self)

    self._validate_secrets(raise_exception=fail_if_secrets_missing)

validate_image_builder() -> None

Validates that the stack has an image builder if required.

If the stack requires an image builder, but none is specified, a local image builder will be created and assigned to the stack to ensure backwards compatibility.

Source code in src/zenml/stack/stack.py

def validate_image_builder(self) -> None:
    """Validates that the stack has an image builder if required.

    If the stack requires an image builder, but none is specified, a
    local image builder will be created and assigned to the stack to
    ensure backwards compatibility.
    """
    requires_image_builder = (
        self.orchestrator.flavor != "local"
        or self.step_operator
        or (self.model_deployer and self.model_deployer.flavor != "mlflow")
    )
    skip_default_image_builder = handle_bool_env_var(
        ENV_ZENML_SKIP_IMAGE_BUILDER_DEFAULT, default=False
    )
    if (
        requires_image_builder
        and not skip_default_image_builder
        and not self.image_builder
    ):
        from uuid import uuid4

        from zenml.image_builders import (
            LocalImageBuilder,
            LocalImageBuilderConfig,
            LocalImageBuilderFlavor,
        )

        flavor = LocalImageBuilderFlavor()

        now = utc_now()
        image_builder = LocalImageBuilder(
            id=uuid4(),
            name="temporary_default",
            flavor=flavor.name,
            type=flavor.type,
            config=LocalImageBuilderConfig(),
            user=Client().active_user.id,
            created=now,
            updated=now,
        )

        self._image_builder = image_builder

Functions

Modules

stack_component

Implementation of the ZenML Stack Component class.

Classes

StackComponent(name: str, id: UUID, config: StackComponentConfig, flavor: str, type: StackComponentType, user: Optional[UUID], created: datetime, updated: datetime, labels: Optional[Dict[str, Any]] = None, connector_requirements: Optional[ServiceConnectorRequirements] = None, connector: Optional[UUID] = None, connector_resource_id: Optional[str] = None, *args: Any, **kwargs: Any)

Abstract StackComponent class for all components of a ZenML stack.

Initializes a StackComponent.

Parameters:

Name	Type	Description	Default
name	str	The name of the component.	required
id	UUID	The unique ID of the component.	required
config	StackComponentConfig	The config of the component.	required
flavor	str	The flavor of the component.	required
type	StackComponentType	The type of the component.	required
user	Optional[UUID]	The ID of the user who created the component.	required
created	datetime	The creation time of the component.	required
updated	datetime	The last update time of the component.	required
labels	Optional[Dict[str, Any]]	The labels of the component.	None
connector_requirements	Optional[ServiceConnectorRequirements]	The requirements for the connector.	None
connector	Optional[UUID]	The ID of a connector linked to the component.	None
connector_resource_id	Optional[str]	The custom resource ID to access through the connector.	None
*args	Any	Additional positional arguments.	()
**kwargs	Any	Additional keyword arguments.	{}

Raises:

Type	Description
ValueError	If a secret reference is passed as name.

Source code in src/zenml/stack/stack_component.py

def __init__(
    self,
    name: str,
    id: UUID,
    config: StackComponentConfig,
    flavor: str,
    type: StackComponentType,
    user: Optional[UUID],
    created: datetime,
    updated: datetime,
    labels: Optional[Dict[str, Any]] = None,
    connector_requirements: Optional[ServiceConnectorRequirements] = None,
    connector: Optional[UUID] = None,
    connector_resource_id: Optional[str] = None,
    *args: Any,
    **kwargs: Any,
):
    """Initializes a StackComponent.

    Args:
        name: The name of the component.
        id: The unique ID of the component.
        config: The config of the component.
        flavor: The flavor of the component.
        type: The type of the component.
        user: The ID of the user who created the component.
        created: The creation time of the component.
        updated: The last update time of the component.
        labels: The labels of the component.
        connector_requirements: The requirements for the connector.
        connector: The ID of a connector linked to the component.
        connector_resource_id: The custom resource ID to access through
            the connector.
        *args: Additional positional arguments.
        **kwargs: Additional keyword arguments.

    Raises:
        ValueError: If a secret reference is passed as name.
    """
    if secret_utils.is_secret_reference(name):
        raise ValueError(
            "Passing the `name` attribute of a stack component as a "
            "secret reference is not allowed."
        )

    self.id = id
    self.name = name
    self._config = config
    self.flavor = flavor
    self.type = type
    self.user = user
    self.created = created
    self.updated = updated
    self.labels = labels
    self.connector_requirements = connector_requirements
    self.connector = connector
    self.connector_resource_id = connector_resource_id
    self._connector_instance: Optional[ServiceConnector] = None

Attributes

apt_packages: List[str] property

List of APT package requirements for the component.

Returns:

Type	Description
List[str]	A list of APT package requirements for the component.

config: StackComponentConfig property

Returns the configuration of the stack component.

This should be overwritten by any subclasses that define custom configs to return the correct config class.

Returns:

Type	Description
StackComponentConfig	The configuration of the stack component.

local_path: Optional[str] property

Path to a local directory to store persistent information.

This property should only be implemented by components that need to store persistent information in a directory on the local machine and also need that information to be available during pipeline runs.

IMPORTANT: the path returned by this property must always be a path that is relative to the ZenML local store's directory. The local orchestrators rely on this convention to correctly mount the local folders in the containers. This is an example of a valid path:

from zenml.config.global_config import GlobalConfiguration

...

@property
def local_path(self) -> Optional[str]:

    return os.path.join(
        GlobalConfiguration().local_stores_path,
        str(self.uuid),
    )

Returns:

Type	Description
Optional[str]	A path to a local directory used by the component to store
Optional[str]	persistent information.

log_file: Optional[str] property

Optional path to a log file for the stack component.

Returns:

Type	Description
Optional[str]	Optional path to a log file for the stack component.

post_registration_message: Optional[str] property

Optional message printed after the stack component is registered.

Returns:

Type	Description
Optional[str]	An optional message.

requirements: Set[str] property

Set of PyPI requirements for the component.

Returns:

Type	Description
Set[str]	A set of PyPI requirements for the component.

settings_class: Optional[Type[BaseSettings]] property

Class specifying available settings for this component.

Returns:

Type	Description
Optional[Type[BaseSettings]]	Optional settings class.

validator: Optional[StackValidator] property

The optional validator of the stack component.

This validator will be called each time a stack with the stack component is initialized. Subclasses should override this property and return a StackValidator that makes sure they're not included in any stack that they're not compatible with.

Returns:

Type	Description
Optional[StackValidator]	An optional StackValidator instance.

Functions

cleanup() -> None

Cleans up the component after it has been used.

Source code in src/zenml/stack/stack_component.py

def cleanup(self) -> None:
    """Cleans up the component after it has been used."""
    pass

cleanup_step_run(info: StepRunInfo, step_failed: bool) -> None

Cleans up resources after the step run is finished.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that was executed.	required
step_failed	bool	Whether the step failed.	required

Source code in src/zenml/stack/stack_component.py

def cleanup_step_run(self, info: "StepRunInfo", step_failed: bool) -> None:
    """Cleans up resources after the step run is finished.

    Args:
        info: Info about the step that was executed.
        step_failed: Whether the step failed.
    """

connector_has_expired() -> bool

Checks whether the connector linked to this stack component has expired.

Returns:

Type	Description
bool	Whether the connector linked to this stack component has expired, or isn't linked to a connector.

Source code in src/zenml/stack/stack_component.py

def connector_has_expired(self) -> bool:
    """Checks whether the connector linked to this stack component has expired.

    Returns:
        Whether the connector linked to this stack component has expired, or isn't linked to a connector.
    """
    if self.connector is None:
        # The stack component isn't linked to a connector
        return False

    if self._connector_instance is None:
        return True

    return self._connector_instance.has_expired()

from_model(component_model: ComponentResponse) -> StackComponent classmethod

Creates a StackComponent from a ComponentModel.

Parameters:

Name	Type	Description	Default
component_model	ComponentResponse	The ComponentModel to create the StackComponent	required

Returns:

Type	Description
StackComponent	The created StackComponent.

Raises:

Type	Description
ImportError	If the flavor can't be imported.

Source code in src/zenml/stack/stack_component.py

@classmethod
def from_model(
    cls, component_model: "ComponentResponse"
) -> "StackComponent":
    """Creates a StackComponent from a ComponentModel.

    Args:
        component_model: The ComponentModel to create the StackComponent

    Returns:
        The created StackComponent.

    Raises:
        ImportError: If the flavor can't be imported.
    """
    from zenml.stack import Flavor

    flavor_model = component_model.flavor
    flavor = Flavor.from_model(flavor_model)

    configuration = flavor.config_class(**component_model.configuration)

    if component_model.user is not None:
        user_id = component_model.user.id
    else:
        user_id = None

    try:
        return flavor.implementation_class(
            user=user_id,
            name=component_model.name,
            id=component_model.id,
            config=configuration,
            labels=component_model.labels,
            flavor=component_model.flavor_name,
            type=component_model.type,
            created=component_model.created,
            updated=component_model.updated,
            connector_requirements=flavor.service_connector_requirements,
            connector=component_model.connector.id
            if component_model.connector
            else None,
            connector_resource_id=component_model.connector_resource_id,
        )
    except ImportError as e:
        from zenml.integrations.registry import integration_registry

        integration_requirements = " ".join(
            integration_registry.select_integration_requirements(
                flavor_model.integration
            )
        )

        if integration_registry.is_installed(flavor_model.integration):
            raise ImportError(
                f"{e}\n\n"
                f"Something went wrong while trying to import from the "
                f"`{flavor_model.integration}` integration. Please make "
                "sure that all its requirements are installed properly by "
                "reinstalling the integration either through our CLI: "
                f"`zenml integration install {flavor_model.integration} "
                "-y` or by manually installing its requirements: "
                f"`pip install {integration_requirements}`. If the error "
                "persists, please contact the ZenML team."
            ) from e
        else:
            raise ImportError(
                f"{e}\n\n"
                f"The `{flavor_model.integration}` integration that you "
                "are trying to use is not installed in your current "
                "environment. Please make sure that it is installed by "
                "either using our CLI: `zenml integration install "
                f"{flavor_model.integration}` or by manually installing "
                f"its requirements: `pip install "
                f"{integration_requirements}`"
            ) from e

get_connector() -> Optional[ServiceConnector]

Returns the connector linked to this stack component.

Returns:

Type	Description
Optional[ServiceConnector]	The connector linked to this stack component.

Raises:

Type	Description
RuntimeError	If the stack component does not specify connector requirements or if the connector linked to the component is not compatible or not found.

Source code in src/zenml/stack/stack_component.py

def get_connector(self) -> Optional["ServiceConnector"]:
    """Returns the connector linked to this stack component.

    Returns:
        The connector linked to this stack component.

    Raises:
        RuntimeError: If the stack component does not specify connector
            requirements or if the connector linked to the component is not
            compatible or not found.
    """
    from zenml.client import Client

    if self.connector is None:
        return None

    if self._connector_instance is not None:
        # If the connector instance is still valid, return it. Otherwise,
        # we'll try to get a new one.
        if not self._connector_instance.has_expired():
            return self._connector_instance

    if self.connector_requirements is None:
        raise RuntimeError(
            f"Unable to get connector for component {self} because this "
            "component does not declare any connector requirements in its. "
            "flavor specification. Override the "
            "`service_connector_requirements` method in its flavor class "
            "to return a connector requirements specification and try "
            "again."
        )

    if self.connector_requirements.resource_id_attr is not None:
        # Check if an attribute is set in the component configuration
        resource_id = getattr(
            self.config, self.connector_requirements.resource_id_attr
        )
    else:
        # Otherwise, use the resource ID configured in the component
        resource_id = self.connector_resource_id

    client = Client()
    try:
        self._connector_instance = client.get_service_connector_client(
            name_id_or_prefix=self.connector,
            resource_type=self.connector_requirements.resource_type,
            resource_id=resource_id,
        )
    except KeyError:
        raise RuntimeError(
            f"The connector with ID {self.connector} linked "
            f"to the '{self.name}' {self.type} stack component could not "
            f"be found or is not accessible. Please verify that the "
            f"connector exists and that you have access to it."
        )
    except ValueError as e:
        raise RuntimeError(
            f"The connector with ID {self.connector} linked "
            f"to the '{self.name}' {self.type} stack component could not "
            f"be correctly configured: {e}."
        )
    except AuthorizationException as e:
        raise RuntimeError(
            f"The connector with ID {self.connector} linked "
            f"to the '{self.name}' {self.type} stack component could not "
            f"be accessed due to an authorization error: {e}. Please "
            f"verify that you have access to the connector and try again."
        )

    return self._connector_instance

get_docker_builds(deployment: PipelineDeploymentBase) -> List[BuildConfiguration]

Gets the Docker builds required for the component.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentBase	The pipeline deployment for which to get the builds.	required

Returns:

Type	Description
List[BuildConfiguration]	The required Docker builds.

Source code in src/zenml/stack/stack_component.py

def get_docker_builds(
    self, deployment: "PipelineDeploymentBase"
) -> List["BuildConfiguration"]:
    """Gets the Docker builds required for the component.

    Args:
        deployment: The pipeline deployment for which to get the builds.

    Returns:
        The required Docker builds.
    """
    return []

get_pipeline_run_metadata(run_id: UUID) -> Dict[str, MetadataType]

Get general component-specific metadata for a pipeline run.

Parameters:

Name	Type	Description	Default
run_id	UUID	The ID of the pipeline run.	required

Returns:

Type	Description
Dict[str, MetadataType]	A dictionary of metadata.

Source code in src/zenml/stack/stack_component.py

def get_pipeline_run_metadata(
    self, run_id: UUID
) -> Dict[str, "MetadataType"]:
    """Get general component-specific metadata for a pipeline run.

    Args:
        run_id: The ID of the pipeline run.

    Returns:
        A dictionary of metadata.
    """
    return {}

get_settings(container: Union[Step, StepRunResponse, StepRunInfo, PipelineDeploymentBase, PipelineDeploymentResponse, PipelineRunResponse]) -> BaseSettings

Gets settings for this stack component.

This will return None if the stack component doesn't specify a settings class or the container doesn't contain runtime options for this component.

Parameters:

Name	Type	Description	Default
container	Union[Step, StepRunResponse, StepRunInfo, PipelineDeploymentBase, PipelineDeploymentResponse, PipelineRunResponse]	The Step, StepRunInfo or PipelineDeployment from which to get the settings.	required

Returns:

Type	Description
BaseSettings	Settings for this stack component.

Raises:

Type	Description
RuntimeError	If the stack component does not specify a settings class.

Source code in src/zenml/stack/stack_component.py

def get_settings(
    self,
    container: Union[
        "Step",
        "StepRunResponse",
        "StepRunInfo",
        "PipelineDeploymentBase",
        "PipelineDeploymentResponse",
        "PipelineRunResponse",
    ],
) -> "BaseSettings":
    """Gets settings for this stack component.

    This will return `None` if the stack component doesn't specify a
    settings class or the container doesn't contain runtime
    options for this component.

    Args:
        container: The `Step`, `StepRunInfo` or `PipelineDeployment` from
            which to get the settings.

    Returns:
        Settings for this stack component.

    Raises:
        RuntimeError: If the stack component does not specify a settings
            class.
    """
    if not self.settings_class:
        raise RuntimeError(
            f"Unable to get settings for component {self} because this "
            "component does not have an associated settings class. "
            "Return a settings class from the `@settings_class` property "
            "and try again."
        )

    key = settings_utils.get_stack_component_setting_key(self)

    all_settings = (
        container.config.settings
        if isinstance(
            container,
            (Step, StepRunResponse, StepRunInfo, PipelineRunResponse),
        )
        else container.pipeline_configuration.settings
    )

    if key in all_settings:
        return self.settings_class.model_validate(dict(all_settings[key]))
    else:
        return self.settings_class()

get_step_run_metadata(info: StepRunInfo) -> Dict[str, MetadataType]

Get component- and step-specific metadata after a step ran.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that was executed.	required

Returns:

Type	Description
Dict[str, MetadataType]	A dictionary of metadata.

Source code in src/zenml/stack/stack_component.py

def get_step_run_metadata(
    self, info: "StepRunInfo"
) -> Dict[str, "MetadataType"]:
    """Get component- and step-specific metadata after a step ran.

    Args:
        info: Info about the step that was executed.

    Returns:
        A dictionary of metadata.
    """
    return {}

prepare_pipeline_deployment(deployment: PipelineDeploymentResponse, stack: Stack) -> None

Prepares deploying the pipeline.

This method gets called immediately before a pipeline is deployed. Subclasses should override it if they require runtime configuration options or if they need to run code before the pipeline deployment.

Parameters:

Name	Type	Description	Default
deployment	PipelineDeploymentResponse	The pipeline deployment configuration.	required
stack	Stack	The stack on which the pipeline will be deployed.	required

Source code in src/zenml/stack/stack_component.py

def prepare_pipeline_deployment(
    self,
    deployment: "PipelineDeploymentResponse",
    stack: "Stack",
) -> None:
    """Prepares deploying the pipeline.

    This method gets called immediately before a pipeline is deployed.
    Subclasses should override it if they require runtime configuration
    options or if they need to run code before the pipeline deployment.

    Args:
        deployment: The pipeline deployment configuration.
        stack: The stack on which the pipeline will be deployed.
    """

prepare_step_run(info: StepRunInfo) -> None

Prepares running a step.

Parameters:

Name	Type	Description	Default
info	StepRunInfo	Info about the step that will be executed.	required

Source code in src/zenml/stack/stack_component.py

def prepare_step_run(self, info: "StepRunInfo") -> None:
    """Prepares running a step.

    Args:
        info: Info about the step that will be executed.
    """

StackComponentConfig(warn_about_plain_text_secrets: bool = False, **kwargs: Any)

Bases: BaseModel, ABC

Base class for all ZenML stack component configs.

Ensures that secret references don't clash with pydantic validation.

StackComponents allow the specification of all their string attributes using secret references of the form {{secret_name.key}}. This however is only possible when the stack component does not perform any explicit validation of this attribute using pydantic validators. If this were the case, the validation would run on the secret reference and would fail or in the worst case, modify the secret reference and lead to unexpected behavior. This method ensures that no attributes that require custom pydantic validation are set as secret references.

Parameters:

Name	Type	Description	Default
warn_about_plain_text_secrets	bool	If true, then warns about using plain-text secrets.	False
**kwargs	Any	Arguments to initialize this stack component.	{}

Raises:

Type	Description
ValueError	If an attribute that requires custom pydantic validation is passed as a secret reference, or if the name attribute was passed as a secret reference.

Source code in src/zenml/stack/stack_component.py

def __init__(
    self, warn_about_plain_text_secrets: bool = False, **kwargs: Any
) -> None:
    """Ensures that secret references don't clash with pydantic validation.

    StackComponents allow the specification of all their string attributes
    using secret references of the form `{{secret_name.key}}`. This however
    is only possible when the stack component does not perform any explicit
    validation of this attribute using pydantic validators. If this were
    the case, the validation would run on the secret reference and would
    fail or in the worst case, modify the secret reference and lead to
    unexpected behavior. This method ensures that no attributes that require
    custom pydantic validation are set as secret references.

    Args:
        warn_about_plain_text_secrets: If true, then warns about using
            plain-text secrets.
        **kwargs: Arguments to initialize this stack component.

    Raises:
        ValueError: If an attribute that requires custom pydantic validation
            is passed as a secret reference, or if the `name` attribute
            was passed as a secret reference.
    """
    for key, value in kwargs.items():
        try:
            field = self.__class__.model_fields[key]
        except KeyError:
            # Value for a private attribute or non-existing field, this
            # will fail during the upcoming pydantic validation
            continue

        if value is None:
            continue

        if not secret_utils.is_secret_reference(value):
            if (
                secret_utils.is_secret_field(field)
                and warn_about_plain_text_secrets
            ):
                logger.warning(
                    "You specified a plain-text value for the sensitive "
                    f"attribute `{key}` for a `{self.__class__.__name__}` "
                    "stack component. This is currently only a warning, "
                    "but future versions of ZenML will require you to pass "
                    "in sensitive information as secrets. Check out the "
                    "documentation on how to configure your stack "
                    "components with secrets here: "
                    "https://docs.zenml.io/getting-started/deploying-zenml/secret-management"
                )
            continue

        if pydantic_utils.has_validators(
            pydantic_class=self.__class__, field_name=key
        ):
            raise ValueError(
                f"Passing the stack component attribute `{key}` as a "
                "secret reference is not allowed as additional validation "
                "is required for this attribute."
            )

    super().__init__(**kwargs)

Attributes

is_local: bool property

Checks if this stack component is running locally.

Concrete stack component configuration classes should override this method to return True if the stack component is relying on local resources or capabilities (e.g. local filesystem, local database or other services).

Examples:

• Artifact Stores that store artifacts in the local filesystem
• Orchestrators that are connected to local orchestration runtime services (e.g. local Kubernetes clusters, Docker containers etc).

Returns:

Type	Description
bool	True if this config is for a local component, False otherwise.

is_remote: bool property

Checks if this stack component is running remotely.

Concrete stack component configuration classes should override this method to return True if the stack component is running in a remote location, and it needs to access the ZenML database.

This designation is used to determine if the stack component can be used with a local ZenML database or if it requires a remote ZenML server.

Examples:

• Orchestrators that are running pipelines in the cloud or in a location other than the local host
• Step Operators that are running steps in the cloud or in a location other than the local host

Returns:

Type	Description
bool	True if this config is for a remote component, False otherwise.

is_valid: bool property

Checks if the stack component configurations are valid.

Concrete stack component configuration classes should override this method to return False if the stack component configurations are invalid.

Returns:

Type	Description
bool	True if the stack component config is valid, False otherwise.

required_secrets: Set[secret_utils.SecretReference] property

All required secrets for this stack component.

Returns:

Type	Description
Set[SecretReference]	The required secrets of this stack component.

Functions

Functions

Modules

stack_validator

Implementation of the ZenML Stack Validator.

Classes

StackValidator(required_components: Optional[AbstractSet[StackComponentType]] = None, custom_validation_function: Optional[Callable[[Stack], Tuple[bool, str]]] = None)

A StackValidator is used to validate a stack configuration.

Each StackComponent can provide a StackValidator to make sure it is compatible with all components of the stack. The KubeflowOrchestrator for example will always require the stack to have a container registry in order to push the docker images that are required to run a pipeline in Kubeflow Pipelines.

Initializes a StackValidator instance.

Parameters:

Name	Type	Description	Default
required_components	Optional[AbstractSet[StackComponentType]]	Optional set of stack components that must exist in the stack.	None
custom_validation_function	Optional[Callable[[Stack], Tuple[bool, str]]]	Optional function that returns whether a stack is valid and an error message to show if not valid.	None

Source code in src/zenml/stack/stack_validator.py

def __init__(
    self,
    required_components: Optional[AbstractSet[StackComponentType]] = None,
    custom_validation_function: Optional[
        Callable[["Stack"], Tuple[bool, str]]
    ] = None,
):
    """Initializes a `StackValidator` instance.

    Args:
        required_components: Optional set of stack components that must
            exist in the stack.
        custom_validation_function: Optional function that returns whether
            a stack is valid and an error message to show if not valid.
    """
    self._required_components = required_components or set()
    self._custom_validation_function = custom_validation_function

Functions

validate(stack: Stack) -> None

Validates the given stack.

Checks if the stack contains all the required components and passes the custom validation function of the validator.

Parameters:

Name	Type	Description	Default
stack	Stack	The stack to validate.	required

Raises:

Type	Description
StackValidationError	If the stack does not meet all the validation criteria.

Source code in src/zenml/stack/stack_validator.py

def validate(self, stack: "Stack") -> None:
    """Validates the given stack.

    Checks if the stack contains all the required components and passes
    the custom validation function of the validator.

    Args:
        stack: The stack to validate.

    Raises:
        StackValidationError: If the stack does not meet all the
            validation criteria.
    """
    missing_components = self._required_components - set(stack.components)
    if missing_components:
        raise StackValidationError(
            f"Missing stack components {missing_components} for "
            f"stack: {stack.name}"
        )

    if self._custom_validation_function:
        valid, err_msg = self._custom_validation_function(stack)
        if not valid:
            raise StackValidationError(
                f"Custom validation function failed to validate "
                f"stack '{stack.name}': {err_msg}"
            )

Functions

utils

Util functions for handling stacks, components, and flavors.

Classes

Functions

get_flavor_by_name_and_type_from_zen_store(zen_store: BaseZenStore, flavor_name: str, component_type: StackComponentType) -> FlavorResponse

Get a stack component flavor by name and type from a ZenStore.

Parameters:

Name	Type	Description	Default
zen_store	BaseZenStore	The ZenStore to query.	required
flavor_name	str	The name of a stack component flavor.	required
component_type	StackComponentType	The type of the stack component.	required

Returns:

Type	Description
FlavorResponse	The flavor model.

Raises:

Type	Description
KeyError	If no flavor with the given name and type exists.

Source code in src/zenml/stack/utils.py

def get_flavor_by_name_and_type_from_zen_store(
    zen_store: BaseZenStore,
    flavor_name: str,
    component_type: StackComponentType,
) -> FlavorResponse:
    """Get a stack component flavor by name and type from a ZenStore.

    Args:
        zen_store: The ZenStore to query.
        flavor_name: The name of a stack component flavor.
        component_type: The type of the stack component.

    Returns:
        The flavor model.

    Raises:
        KeyError: If no flavor with the given name and type exists.
    """
    flavors = zen_store.list_flavors(
        FlavorFilter(name=flavor_name, type=component_type)
    )
    if not flavors:
        raise KeyError(
            f"No flavor with name '{flavor_name}' and type "
            f"'{component_type}' exists."
        )
    return flavors[0]

validate_stack_component_config(configuration_dict: Dict[str, Any], flavor: Union[FlavorResponse, str], component_type: StackComponentType, zen_store: Optional[BaseZenStore] = None, validate_custom_flavors: bool = True) -> Optional[StackComponentConfig]

Validate the configuration of a stack component.

Parameters:

Name	Type	Description	Default
configuration_dict	Dict[str, Any]	The stack component configuration to validate.	required
flavor	Union[FlavorResponse, str]	The model or name of the flavor of the stack component.	required
component_type	StackComponentType	The type of the stack component.	required
zen_store	Optional[BaseZenStore]	An optional ZenStore in which to look for the flavor. If not provided, the flavor will be fetched via the regular ZenML Client. This is mainly useful for checks running inside the ZenML server.	None
validate_custom_flavors	bool	When loading custom flavors from the local environment, this flag decides whether the import failures are raised or an empty value is returned.	True

Returns:

Type	Description
Optional[StackComponentConfig]	The validated stack component configuration or None, if the
Optional[StackComponentConfig]	flavor is a custom flavor that could not be imported from the local
Optional[StackComponentConfig]	environment and the validate_custom_flavors flag is set to False.

Raises:

Type	Description
ValueError	If the configuration is invalid.
ImportError	If the flavor class could not be imported.
ModuleNotFoundError	If the flavor class could not be imported.

Source code in src/zenml/stack/utils.py

def validate_stack_component_config(
    configuration_dict: Dict[str, Any],
    flavor: Union[FlavorResponse, str],
    component_type: StackComponentType,
    zen_store: Optional[BaseZenStore] = None,
    validate_custom_flavors: bool = True,
) -> Optional[StackComponentConfig]:
    """Validate the configuration of a stack component.

    Args:
        configuration_dict: The stack component configuration to validate.
        flavor: The model or name of the flavor of the stack component.
        component_type: The type of the stack component.
        zen_store: An optional ZenStore in which to look for the flavor. If not
            provided, the flavor will be fetched via the regular ZenML Client.
            This is mainly useful for checks running inside the ZenML server.
        validate_custom_flavors: When loading custom flavors from the local
            environment, this flag decides whether the import failures are
            raised or an empty value is returned.

    Returns:
        The validated stack component configuration or None, if the
        flavor is a custom flavor that could not be imported from the local
        environment and the `validate_custom_flavors` flag is set to False.

    Raises:
        ValueError: If the configuration is invalid.
        ImportError: If the flavor class could not be imported.
        ModuleNotFoundError: If the flavor class could not be imported.
    """
    if isinstance(flavor, FlavorResponse):
        flavor_model = flavor
    else:
        if zen_store:
            flavor_model = get_flavor_by_name_and_type_from_zen_store(
                zen_store=zen_store,
                flavor_name=flavor,
                component_type=component_type,
            )
        else:
            flavor_model = Client().get_flavor_by_name_and_type(
                name=flavor,
                component_type=component_type,
            )

    try:
        flavor_class = Flavor.from_model(flavor_model)
    except (ImportError, ModuleNotFoundError):
        # The flavor class couldn't be loaded.
        if flavor_model.is_custom and not validate_custom_flavors:
            return None
        raise

    config_class = flavor_class.config_class
    # Make sure extras are forbidden for the config class. Due to inheritance
    # order, some config classes allow extras by accident which we patch here.
    validation_config_class: Type[StackComponentConfig] = type(
        config_class.__name__,
        (config_class,),
        {"model_config": {"extra": "forbid"}},
    )
    configuration = validation_config_class(**configuration_dict)

    if not configuration.is_valid:
        raise ValueError(
            f"Invalid stack component configuration. Please verify "
            f"the configurations set for {component_type}."
        )
    return configuration

warn_if_config_server_mismatch(configuration: StackComponentConfig) -> None

Warn if a component configuration is mismatched with the ZenML server.

Parameters:

Name	Type	Description	Default
configuration	StackComponentConfig	The component configuration to check.	required

Source code in src/zenml/stack/utils.py

def warn_if_config_server_mismatch(
    configuration: StackComponentConfig,
) -> None:
    """Warn if a component configuration is mismatched with the ZenML server.

    Args:
        configuration: The component configuration to check.
    """
    zen_store = Client().zen_store
    if configuration.is_remote and zen_store.is_local_store():
        if zen_store.type == StoreType.REST:
            logger.warning(
                "You are configuring a stack component that is running "
                "remotely while using a local ZenML server. The component "
                "may not be able to reach the local ZenML server and will "
                "therefore not be functional. Please consider deploying "
                "and/or using a remote ZenML server instead."
            )
        else:
            logger.warning(
                "You are configuring a stack component that is running "
                "remotely while connected to the local database. The component "
                "will not be able to reach the local database and will "
                "therefore not be functional. Please consider deploying "
                "and/or using a remote ZenML server instead."
            )
    elif configuration.is_local and not zen_store.is_local_store():
        logger.warning(
            "You are configuring a stack component that is using "
            "local resources while connected to a remote ZenML server. The "
            "stack component may not be usable from other hosts or by "
            "other users. You should consider using a non-local stack "
            "component alternative instead."
        )