This function is deprecated and will be removed in a future release. Use log_metadata() with infer_artifact=True or specific artifact parameters instead.
The log_artifact_metadata() function logs metadata to artifact versions. It has been superseded by the more flexible log_metadata() function.
Signature def log_artifact_metadata ( metadata : Dict[ str , MetadataType], artifact_name : Optional[ str ] = None , artifact_version : Optional[ str ] = None , ) -> None
Parameters metadata
Dict[str, MetadataType]
required
Dictionary of metadata key-value pairs to log.
The name of the artifact. Can be omitted when called inside a step with only one output.
The version of the artifact. If not provided when called inside a step, metadata will be associated with the newly created artifact.
Migration Guide Old Way (Deprecated) from zenml import step, log_artifact_metadata import pandas as pd @step def process_data () -> pd.DataFrame: df = pd.DataFrame({ "value" : range ( 100 )}) # Old way - deprecated log_artifact_metadata( metadata = { "row_count" : len (df)}, artifact_name = "output" ) return df
New Way (Recommended) from zenml import step, log_metadata import pandas as pd @step def process_data () -> pd.DataFrame: df = pd.DataFrame({ "value" : range ( 100 )}) # New way - recommended log_metadata( metadata = { "row_count" : len (df)}, infer_artifact = True ) return df
Examples with New Approach Single Output Step from zenml import step, log_metadata import pandas as pd @step def create_dataset () -> pd.DataFrame: df = pd.DataFrame({ "data" : range ( 1000 )}) # Automatically infer the artifact (single output) log_metadata( metadata = { "num_rows" : len (df), "columns" : list (df.columns) }, infer_artifact = True ) return df
Multiple Outputs from typing import Tuple, Annotated from zenml import step, log_metadata import pandas as pd @step def split_data () -> Tuple[ Annotated[pd.DataFrame, "train" ], Annotated[pd.DataFrame, "test" ], ]: train = pd.DataFrame({ "x" : range ( 800 )}) test = pd.DataFrame({ "x" : range ( 200 )}) # Specify which output to attach metadata to log_metadata( metadata = { "split" : "train" , "size" : len (train)}, artifact_name = "train" , infer_artifact = True ) log_metadata( metadata = { "split" : "test" , "size" : len (test)}, artifact_name = "test" , infer_artifact = True ) return train, test
External Artifact from zenml import log_metadata # Log metadata to an existing artifact log_metadata( metadata = { "quality_checked" : True , "reviewer" : "data_scientist_1" }, artifact_name = "production_data" , artifact_version = "12" )
Why Migrate? The new log_metadata() function offers several advantages:
Unified interface - One function for all metadata logging (steps, runs, artifacts, models)Better inference - More intelligent automatic detection of targetsMore options - Can log to artifacts by ID, name+version, or inferenceClearer semantics - The infer_artifact parameter makes the behavior explicitFuture-proof - Active development and support
log_metadata Use the new unified metadata function
get_step_context Access step context
ArtifactConfig Configure artifacts
