zenml-nightly 0.70.0.dev20241125__py3-none-any.whl → 0.70.0.dev20241201__py3-none-any.whl

zenml/integrations/modal/step_operators/modal_step_operator.py

@@ -0,0 +1,242 @@
+#  Copyright (c) ZenML GmbH 2024. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at:
+#
+#       https://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
+#  or implied. See the License for the specific language governing
+#  permissions and limitations under the License.
+"""Modal step operator implementation."""
+
+import asyncio
+from typing import TYPE_CHECKING, Dict, List, Optional, Tuple, Type, cast
+
+import modal
+from modal_proto import api_pb2
+
+from zenml.client import Client
+from zenml.config.build_configuration import BuildConfiguration
+from zenml.config.resource_settings import ByteUnit, ResourceSettings
+from zenml.enums import StackComponentType
+from zenml.integrations.modal.flavors import (
+    ModalStepOperatorConfig,
+    ModalStepOperatorSettings,
+)
+from zenml.logger import get_logger
+from zenml.stack import Stack, StackValidator
+from zenml.step_operators import BaseStepOperator
+
+if TYPE_CHECKING:
+    from zenml.config.base_settings import BaseSettings
+    from zenml.config.step_run_info import StepRunInfo
+    from zenml.models import PipelineDeploymentBase
+
+logger = get_logger(__name__)
+
+MODAL_STEP_OPERATOR_DOCKER_IMAGE_KEY = "modal_step_operator"
+
+
+def get_gpu_values(
+    settings: ModalStepOperatorSettings, resource_settings: ResourceSettings
+) -> Optional[str]:
+    """Get the GPU values for the Modal step operator.
+
+    Args:
+        settings: The Modal step operator settings.
+        resource_settings: The resource settings.
+
+    Returns:
+        The GPU string if a count is specified, otherwise the GPU type.
+    """
+    if not settings.gpu:
+        return None
+    gpu_count = resource_settings.gpu_count
+    return f"{settings.gpu}:{gpu_count}" if gpu_count else settings.gpu
+
+
+class ModalStepOperator(BaseStepOperator):
+    """Step operator to run a step on Modal.
+
+    This class defines code that can set up a Modal environment and run
+    functions in it.
+    """
+
+    @property
+    def config(self) -> ModalStepOperatorConfig:
+        """Get the Modal step operator configuration.
+
+        Returns:
+            The Modal step operator configuration.
+        """
+        return cast(ModalStepOperatorConfig, self._config)
+
+    @property
+    def settings_class(self) -> Optional[Type["BaseSettings"]]:
+        """Get the settings class for the Modal step operator.
+
+        Returns:
+            The Modal step operator settings class.
+        """
+        return ModalStepOperatorSettings
+
+    @property
+    def validator(self) -> Optional[StackValidator]:
+        """Get the stack validator for the Modal step operator.
+
+        Returns:
+            The stack validator.
+        """
+
+        def _validate_remote_components(stack: "Stack") -> Tuple[bool, str]:
+            if stack.artifact_store.config.is_local:
+                return False, (
+                    "The Modal step operator runs code remotely and "
+                    "needs to write files into the artifact store, but the "
+                    f"artifact store `{stack.artifact_store.name}` of the "
+                    "active stack is local. Please ensure that your stack "
+                    "contains a remote artifact store when using the Modal "
+                    "step operator."
+                )
+
+            container_registry = stack.container_registry
+            assert container_registry is not None
+
+            if container_registry.config.is_local:
+                return False, (
+                    "The Modal step operator runs code remotely and "
+                    "needs to push/pull Docker images, but the "
+                    f"container registry `{container_registry.name}` of the "
+                    "active stack is local. Please ensure that your stack "
+                    "contains a remote container registry when using the "
+                    "Modal step operator."
+                )
+
+            return True, ""
+
+        return StackValidator(
+            required_components={
+                StackComponentType.CONTAINER_REGISTRY,
+                StackComponentType.IMAGE_BUILDER,
+            },
+            custom_validation_function=_validate_remote_components,
+        )
+
+    def get_docker_builds(
+        self, deployment: "PipelineDeploymentBase"
+    ) -> List["BuildConfiguration"]:
+        """Get the Docker build configurations for the Modal step operator.
+
+        Args:
+            deployment: The pipeline deployment.
+
+        Returns:
+            A list of Docker build configurations.
+        """
+        builds = []
+        for step_name, step in deployment.step_configurations.items():
+            if step.config.step_operator == self.name:
+                build = BuildConfiguration(
+                    key=MODAL_STEP_OPERATOR_DOCKER_IMAGE_KEY,
+                    settings=step.config.docker_settings,
+                    step_name=step_name,
+                )
+                builds.append(build)
+
+        return builds
+
+    def launch(
+        self,
+        info: "StepRunInfo",
+        entrypoint_command: List[str],
+        environment: Dict[str, str],
+    ) -> None:
+        """Launch a step run on Modal.
+
+        Args:
+            info: The step run information.
+            entrypoint_command: The entrypoint command for the step.
+            environment: The environment variables for the step.
+
+        Raises:
+            RuntimeError: If no Docker credentials are found for the container registry.
+            ValueError: If no container registry is found in the stack.
+        """
+        settings = cast(ModalStepOperatorSettings, self.get_settings(info))
+        image_name = info.get_image(key=MODAL_STEP_OPERATOR_DOCKER_IMAGE_KEY)
+        zc = Client()
+        stack = zc.active_stack
+
+        if not stack.container_registry:
+            raise ValueError(
+                "No Container registry found in the stack. "
+                "Please add a container registry and ensure "
+                "it is correctly configured."
+            )
+
+        if docker_creds := stack.container_registry.credentials:
+            docker_username, docker_password = docker_creds
+        else:
+            raise RuntimeError(
+                "No Docker credentials found for the container registry."
+            )
+
+        my_secret = modal.secret._Secret.from_dict(
+            {
+                "REGISTRY_USERNAME": docker_username,
+                "REGISTRY_PASSWORD": docker_password,
+            }
+        )
+
+        spec = modal.image.DockerfileSpec(
+            commands=[f"FROM {image_name}"], context_files={}
+        )
+
+        zenml_image = modal.Image._from_args(
+            dockerfile_function=lambda *_, **__: spec,
+            force_build=False,
+            image_registry_config=modal.image._ImageRegistryConfig(
+                api_pb2.REGISTRY_AUTH_TYPE_STATIC_CREDS, my_secret
+            ),
+        ).env(environment)
+
+        resource_settings = info.config.resource_settings
+        gpu_values = get_gpu_values(settings, resource_settings)
+
+        app = modal.App(
+            f"zenml-{info.run_name}-{info.step_run_id}-{info.pipeline_step_name}"
+        )
+
+        async def run_sandbox() -> asyncio.Future[None]:
+            loop = asyncio.get_event_loop()
+            future = loop.create_future()
+            with modal.enable_output():
+                async with app.run():
+                    memory_mb = resource_settings.get_memory(ByteUnit.MB)
+                    memory_int = (
+                        int(memory_mb) if memory_mb is not None else None
+                    )
+                    sb = await modal.Sandbox.create.aio(
+                        "bash",
+                        "-c",
+                        " ".join(entrypoint_command),
+                        image=zenml_image,
+                        gpu=gpu_values,
+                        cpu=resource_settings.cpu_count,
+                        memory=memory_int,
+                        cloud=settings.cloud,
+                        region=settings.region,
+                        app=app,
+                        timeout=86400,  # 24h, the max Modal allows
+                    )
+
+                    await sb.wait.aio()
+
+            future.set_result(None)
+            return future
+
+        asyncio.run(run_sandbox())

zenml/io/filesystem.py

@@ -54,11 +54,11 @@ class BaseFilesystem(ABC):
 
     @staticmethod
     @abstractmethod
-    def open(name: PathType, mode: str = "r") -> Any:
+    def open(path: PathType, mode: str = "r") -> Any:
         """Opens a file.
 
         Args:
-            name: The path to the file.
+            path: The path to the file.
             mode: The mode to open the file in.
 
         Returns:

zenml/io/local_filesystem.py

@@ -55,18 +55,18 @@ class LocalFilesystem(BaseFilesystem):
     SUPPORTED_SCHEMES: ClassVar[Set[str]] = {""}
 
     @staticmethod
-    def open(name: PathType, mode: str = "r") -> Any:
+    def open(path: PathType, mode: str = "r") -> Any:
         """Open a file at the given path.
 
         Args:
-            name: The path to the file.
+            path: The path to the file.
             mode: The mode to open the file.
 
         Returns:
             Any: The file object.
         """
         encoding = "utf-8" if "b" not in mode else None
-        return open(name, mode=mode, encoding=encoding)
+        return open(path, mode=mode, encoding=encoding)
 
     @staticmethod
     def copyfile(

zenml/materializers/built_in_materializer.py

@@ -28,7 +28,7 @@ from typing import (
 )
 
 from zenml.artifact_stores.base_artifact_store import BaseArtifactStore
-from zenml.enums import ArtifactType
+from zenml.enums import ArtifactType, VisualizationType
 from zenml.logger import get_logger
 from zenml.materializers.base_materializer import BaseMaterializer
 from zenml.materializers.materializer_registry import materializer_registry
@@ -415,6 +415,23 @@ class BuiltInContainerMaterializer(BaseMaterializer):
                 self.artifact_store.rmtree(entry["path"])
             raise e
 
+    # save dict type objects to JSON file with JSON visualization type
+    def save_visualizations(self, data: Any) -> Dict[str, "VisualizationType"]:
+        """Save visualizations for the given data.
+
+        Args:
+            data: The data to save visualizations for.
+
+        Returns:
+            A dictionary of visualization URIs and their types.
+        """
+        # dict/list type objects are always saved as JSON files
+        # doesn't work for non-serializable types as they
+        # are saved as list of lists in different files
+        if _is_serializable(data):
+            return {self.data_path: VisualizationType.JSON}
+        return {}
+
     def extract_metadata(self, data: Any) -> Dict[str, "MetadataType"]:
         """Extract metadata from the given built-in container object.
 

zenml/materializers/structured_string_materializer.py

@@ -19,22 +19,23 @@ from typing import Dict, Type, Union
 from zenml.enums import ArtifactType, VisualizationType
 from zenml.logger import get_logger
 from zenml.materializers.base_materializer import BaseMaterializer
-from zenml.types import CSVString, HTMLString, MarkdownString
+from zenml.types import CSVString, HTMLString, JSONString, MarkdownString
 
 logger = get_logger(__name__)
 
 
-STRUCTURED_STRINGS = Union[CSVString, HTMLString, MarkdownString]
+STRUCTURED_STRINGS = Union[CSVString, HTMLString, MarkdownString, JSONString]
 
 HTML_FILENAME = "output.html"
 MARKDOWN_FILENAME = "output.md"
 CSV_FILENAME = "output.csv"
+JSON_FILENAME = "output.json"
 
 
 class StructuredStringMaterializer(BaseMaterializer):
     """Materializer for HTML or Markdown strings."""
 
-    ASSOCIATED_TYPES = (CSVString, HTMLString, MarkdownString)
+    ASSOCIATED_TYPES = (CSVString, HTMLString, MarkdownString, JSONString)
     ASSOCIATED_ARTIFACT_TYPE = ArtifactType.DATA_ANALYSIS
 
     def load(self, data_type: Type[STRUCTURED_STRINGS]) -> STRUCTURED_STRINGS:
@@ -94,6 +95,8 @@ class StructuredStringMaterializer(BaseMaterializer):
             filename = HTML_FILENAME
         elif issubclass(data_type, MarkdownString):
             filename = MARKDOWN_FILENAME
+        elif issubclass(data_type, JSONString):
+            filename = JSON_FILENAME
         else:
             raise ValueError(
                 f"Data type {data_type} is not supported by this materializer."
@@ -120,6 +123,8 @@ class StructuredStringMaterializer(BaseMaterializer):
             return VisualizationType.HTML
         elif issubclass(data_type, MarkdownString):
             return VisualizationType.MARKDOWN
+        elif issubclass(data_type, JSONString):
+            return VisualizationType.JSON
         else:
             raise ValueError(
                 f"Data type {data_type} is not supported by this materializer."

zenml/model/model.py

@@ -13,14 +13,12 @@
 #  permissions and limitations under the License.
 """Model user facing interface to pass into pipeline or step."""
 
-import datetime
 from typing import (
     TYPE_CHECKING,
     Any,
     Dict,
     List,
     Optional,
-    Tuple,
     Union,
 )
 from uuid import UUID
@@ -41,7 +39,6 @@ if TYPE_CHECKING:
         ModelResponse,
         ModelVersionResponse,
         PipelineRunResponse,
-        StepRunResponse,
     )
 
 logger = get_logger(__name__)
@@ -60,7 +57,9 @@ class Model(BaseModel):
     ethics: The ethical implications of the model.
     tags: Tags associated with the model.
     version: The version name, version number or stage is optional and points model context
-        to a specific version/stage. If skipped new version will be created.
+        to a specific version/stage. If skipped new version will be created. `version`
+        also supports placeholders: standard `{date}` and `{time}` and any custom placeholders
+        that are passed as substitutions in the pipeline or step decorators.
     save_models_to_registry: Whether to save all ModelArtifacts to Model Registry,
         if available in active stack.
     """
@@ -337,12 +336,16 @@ class Model(BaseModel):
             metadata: The metadata to log.
         """
         from zenml.client import Client
+        from zenml.models import RunMetadataResource
 
         response = self._get_or_create_model_version()
         Client().create_run_metadata(
             metadata=metadata,
-            resource_id=response.id,
-            resource_type=MetadataResourceTypes.MODEL_VERSION,
+            resources=[
+                RunMetadataResource(
+                    id=response.id, type=MetadataResourceTypes.MODEL_VERSION
+                )
+            ],
         )
 
     @property
@@ -537,6 +540,8 @@ class Model(BaseModel):
         from zenml.models import ModelRequest
 
         zenml_client = Client()
+        # backup logic, if the Model class is used directly from the code
+        self.name = format_name_template(self.name)
         if self.model_version_id:
             mv = zenml_client.get_model_version(
                 model_version_name_or_number_or_id=self.model_version_id,
@@ -743,85 +748,6 @@ class Model(BaseModel):
             )
         )
 
-    def _prepare_model_version_before_step_launch(
-        self,
-        pipeline_run: "PipelineRunResponse",
-        step_run: Optional["StepRunResponse"],
-        return_logs: bool,
-    ) -> Tuple[str, "PipelineRunResponse", Optional["StepRunResponse"]]:
-        """Prepares model version inside pipeline run.
-
-        Args:
-            pipeline_run: pipeline run
-            step_run: step run (passed only if model version is defined in a step explicitly)
-            return_logs: whether to return logs or not
-
-        Returns:
-            Logs related to the Dashboard URL to show later.
-        """
-        from zenml.client import Client
-        from zenml.models import PipelineRunUpdate, StepRunUpdate
-
-        logs = ""
-
-        # copy Model instance to prevent corrupting configs of the
-        # subsequent runs, if they share the same config object
-        self_copy = self.model_copy()
-
-        # in case request is within the step and no self-configuration is provided
-        # try reuse what's in the pipeline run first
-        if step_run is None and pipeline_run.model_version is not None:
-            self_copy.version = pipeline_run.model_version.name
-            self_copy.model_version_id = pipeline_run.model_version.id
-        # otherwise try to fill the templated name, if needed
-        elif isinstance(self_copy.version, str):
-            if pipeline_run.start_time:
-                start_time = pipeline_run.start_time
-            else:
-                start_time = datetime.datetime.now(datetime.timezone.utc)
-            self_copy.version = format_name_template(
-                self_copy.version,
-                date=start_time.strftime("%Y_%m_%d"),
-                time=start_time.strftime("%H_%M_%S_%f"),
-            )
-
-        # if exact model not yet defined - try to get/create and update it
-        # back to the run accordingly
-        if self_copy.model_version_id is None:
-            model_version_response = self_copy._get_or_create_model_version()
-
-            client = Client()
-            # update the configured model version id in runs accordingly
-            if step_run:
-                step_run = client.zen_store.update_run_step(
-                    step_run_id=step_run.id,
-                    step_run_update=StepRunUpdate(
-                        model_version_id=model_version_response.id
-                    ),
-                )
-            else:
-                pipeline_run = client.zen_store.update_run(
-                    run_id=pipeline_run.id,
-                    run_update=PipelineRunUpdate(
-                        model_version_id=model_version_response.id
-                    ),
-                )
-
-            if return_logs:
-                from zenml.utils.cloud_utils import try_get_model_version_url
-
-                if logs_to_show := try_get_model_version_url(
-                    model_version_response
-                ):
-                    logs = logs_to_show
-                else:
-                    logs = (
-                        "Models can be viewed in the dashboard using ZenML Pro. Sign up "
-                        "for a free trial at https://www.zenml.io/pro/"
-                    )
-        self.model_version_id = self_copy.model_version_id
-        return logs, pipeline_run, step_run
-
     @property
     def _lazy_version(self) -> Optional[str]:
         """Get version name for lazy loader.

zenml/model/utils.py

@@ -52,29 +52,31 @@ def log_model_metadata(
             `model` in decorator.
 
     Raises:
-        ValueError: If no model name/version is provided and the function is not
-            called inside a step with configured `model` in decorator.
+        ValueError: If the function is not called with proper input.
     """
     logger.warning(
         "The `log_model_metadata` function is deprecated and will soon be "
         "removed. Please use `log_metadata` instead."
     )
 
-    if model_name and model_version:
-        from zenml import Model
+    from zenml import log_metadata
 
-        mv = Model(name=model_name, version=model_version)
+    if model_name and model_version:
+        log_metadata(
+            metadata=metadata,
+            model_version=model_version,
+            model_name=model_name,
+        )
+    elif model_name is None and model_version is None:
+        log_metadata(
+            metadata=metadata,
+            infer_model=True,
+        )
     else:
-        try:
-            step_context = get_step_context()
-        except RuntimeError:
-            raise ValueError(
-                "Model name and version must be provided unless the function is "
-                "called inside a step with configured `model` in decorator."
-            )
-        mv = step_context.model
-
-    mv.log_metadata(metadata)
+        raise ValueError(
+            "You can call `log_model_metadata` by either providing both "
+            "`model_name` and `model_version` or keeping both of them None."
+        )
 
 
 def link_artifact_version_to_model_version(
@@ -107,7 +109,7 @@ def link_artifact_to_model(
         model: The model to link to.
 
     Raises:
-        RuntimeError: If called outside of a step.
+        RuntimeError: If called outside a step.
     """
     if not model:
         is_issue = False

zenml/models/__init__.py

@@ -372,6 +372,10 @@ from zenml.models.v2.misc.auth_models import (
     OAuthRedirectResponse,
     OAuthTokenResponse,
 )
+from zenml.models.v2.misc.run_metadata import (
+    RunMetadataEntry,
+    RunMetadataResource,
+)
 from zenml.models.v2.misc.server_models import (
     ServerModel,
     ServerDatabaseType,
@@ -752,4 +756,6 @@ __all__ = [
     "ServiceConnectorInfo",
     "ServiceConnectorResourcesInfo",
     "ResourcesInfo",
+    "RunMetadataEntry",
+    "RunMetadataResource",
 ]

zenml/models/v2/core/artifact_version.py

@@ -576,6 +576,7 @@ class ArtifactVersionFilter(WorkspaceScopedTaggableFilter):
             ModelVersionArtifactSchema,
             ModelVersionSchema,
             PipelineRunSchema,
+            RunMetadataResourceSchema,
             RunMetadataSchema,
             StepRunInputArtifactSchema,
             StepRunOutputArtifactSchema,
@@ -679,10 +680,12 @@ class ArtifactVersionFilter(WorkspaceScopedTaggableFilter):
 
             for key, value in self.run_metadata.items():
                 additional_filter = and_(
-                    RunMetadataSchema.resource_id == ArtifactVersionSchema.id,
-                    RunMetadataSchema.resource_type
+                    RunMetadataResourceSchema.resource_id
+                    == ArtifactVersionSchema.id,
+                    RunMetadataResourceSchema.resource_type
                     == MetadataResourceTypes.ARTIFACT_VERSION,
-                    RunMetadataSchema.key == key,
+                    RunMetadataResourceSchema.run_metadata_id
+                    == RunMetadataSchema.id,
                     self.generate_custom_query_conditions_for_column(
                         value=value,
                         table=RunMetadataSchema,

zenml/models/v2/core/component.py

@@ -81,11 +81,6 @@ class ComponentBase(BaseModel):
         title="The stack component labels.",
     )
 
-    component_spec_path: Optional[str] = Field(
-        default=None,
-        title="The path to the component spec used for mlstacks deployments.",
-    )
-
 
 # ------------------ Request Model ------------------
 
@@ -155,10 +150,6 @@ class ComponentUpdate(BaseUpdate):
         title="The stack component labels.",
         default=None,
     )
-    component_spec_path: Optional[str] = Field(
-        title="The path to the component spec used for mlstacks deployments.",
-        default=None,
-    )
     connector: Optional[UUID] = Field(
         title="The service connector linked to this stack component.",
         default=None,
@@ -201,10 +192,6 @@ class ComponentResponseMetadata(WorkspaceScopedResponseMetadata):
         default=None,
         title="The stack component labels.",
     )
-    component_spec_path: Optional[str] = Field(
-        default=None,
-        title="The path to the component spec used for mlstacks deployments.",
-    )
     connector_resource_id: Optional[str] = Field(
         default=None,
         description="The ID of a specific resource instance to "
@@ -325,15 +312,6 @@ class ComponentResponse(
         """
         return self.get_metadata().labels
 
-    @property
-    def component_spec_path(self) -> Optional[str]:
-        """The `component_spec_path` property.
-
-        Returns:
-            the value of the property.
-        """
-        return self.get_metadata().component_spec_path
-
     @property
     def connector_resource_id(self) -> Optional[str]:
         """The `connector_resource_id` property.

zenml/models/v2/core/model_version.py

@@ -652,6 +652,7 @@ class ModelVersionFilter(WorkspaceScopedTaggableFilter):
 
         from zenml.zen_stores.schemas import (
             ModelVersionSchema,
+            RunMetadataResourceSchema,
             RunMetadataSchema,
             UserSchema,
         )
@@ -672,10 +673,12 @@ class ModelVersionFilter(WorkspaceScopedTaggableFilter):
 
             for key, value in self.run_metadata.items():
                 additional_filter = and_(
-                    RunMetadataSchema.resource_id == ModelVersionSchema.id,
-                    RunMetadataSchema.resource_type
+                    RunMetadataResourceSchema.resource_id
+                    == ModelVersionSchema.id,
+                    RunMetadataResourceSchema.resource_type
                     == MetadataResourceTypes.MODEL_VERSION,
-                    RunMetadataSchema.key == key,
+                    RunMetadataResourceSchema.run_metadata_id
+                    == RunMetadataSchema.id,
                     self.generate_custom_query_conditions_for_column(
                         value=value,
                         table=RunMetadataSchema,