comet_ml.integration.snowflake ¶

get_artifact_v1 ¶

get_artifact_v1(
    experiment, artifact_name, workspace=None, version_or_alias=None
)

Get the SQL query for a Snowflake Artifact.

Parameters:

experiment (Experiment) –
Instance of Experiment object.
artifact_name (str) –
Artifact name.
workspace (str, default: None ) –
Retrieve an artifact belonging to that workspace.
version_or_alias (Iterable[str], default: None ) –
Retrieve the artifact by the given alias or version.

log_artifact_v1 ¶

log_artifact_v1(
    experiment,
    artifact_name,
    sql,
    sample=None,
    sample_size=100,
    url=None,
    artifact_version=None,
    artifact_aliases=None,
)

Logs a Snowflake Artifact to Comet based on a SQL query and a sample Pandas DataFrame.

Parameters:

experiment (Experiment) –
Instance of an Experiment used to log the Artifact.
artifact_name (str) –
The name of the Artifact to create.
sql (str) –
The SQL used to generate the Snowflake dataset.
sample (Pandas DataFrame, default: None ) –
Sample of the data generated by the SQL query defined previously.
sample_size (int, default: 100 ) –
Number of rows to record from the sample dataframe defined previously. Defaults to 100.
artifact_version (str, default: None ) –
Version of the artifact. If none is provided Comet will auto-increment the version.
artifact_aliases (Iterable[str], default: None ) –
Aliases to associate with the Artifact.

log_snowpark_dataframe_v1 ¶

log_snowpark_dataframe_v1(
    experiment,
    artifact_name,
    dataframe,
    artifact_version=None,
    artifact_aliases=None,
    log_sample=False,
    sample_size=10,
)

Logs a Snowpark Dataframe as a Snowflake Artifact to Comet. The full dataset is not saved in Comet, instead we record both the SQL query and a small preview sample.

Parameters:

experiment (Experiment) –
Instance of an Experiment used to log the Artifact.
artifact_name (str) –
The name of the Artifact to create.
dataframe (Snowpark Dataframe) –
The Snowpark Dataframe to record.
artifact_version (str, default: None ) –
Version of the artifact. If none is provided Comet will auto-increment the version
artifact_aliases (Iterable[str], default: None ) –
Aliases to associate with the Artifact
log_sample (bool, default: False ) –
Whether a preview of the dataset should be saved to the Comet platform.
sample_size (int, default: 10 ) –
Number of rows to record from the sample dataframe defined previously.

Nov. 18, 2024