Run API

pydantic model mlopus.mlflow.RunApi

Bases: Run, EntityApi

Run metadata with MLflow API handle.

field end_time: datetime | None [Required]

Start time in local timezone.

field exp: ExpApi [Required]

Parent experiment.

field id: str [Required]

Run ID in the pattern [w-]+

field metrics: Dict[str, Any] [Required]

Nested metrics dict with float leaf-values.

field name: str [Required]

Run name.

field params: Dict[str, Any] [Required]

Nested params dict with JSON-serializable leaf-values.

field repo: str [Required]

Artifacts repo URL.

field start_time: datetime | None [Required]

End time in local timezone.

field status: RunStatus | None [Required]

Run status.

field tags: Dict[str, Any] [Required]

Nested tags dict with JSON-serializable leaf-values and keys in the pattern: [w-/]+

class Config

Bases: object

Pydantic class config.

cache_artifact(path_in_run='')

Pull run artifact from MLflow server to local cache.

Parameters:

path_in_run (str) –

See path_in_run

Return type:

Path

cache_meta()

Fetch latest metadata of this run and save to cache.

Return type:

RunApi

clean_cached_artifact(path_in_run='')

Clean cached artifact for this run.

Parameters:

path_in_run (str) –

See path_in_run

create_child(name=None, tags=None, repo=None)

Declare a new child run to be used later.

Parameters:

• name (Optional[str]) – See schema.Run.name.

• tags (Optional[Mapping]) – See schema.Run.tags.

• repo (Union[str, Url, None]) – See repo.

end_run(succeeded=True)

End experiment run.

Return type:

RunApi

export_artifact(target, path_in_run='')

Export run artifact cache to target.

See also

• mlopus.mlflow.BaseMlflowApi.export_run_artifact()

Parameters:

• target (Path) – Cache export path.

• path_in_run (str) –

  See path_in_run

Return type:

Path

export_meta(target)

Export metadata cache for this run.

Parameters:

target (Path) – Cache export path..

Return type:

RunApi

find_model_versions(query=None, sorting=None)

Search model versions belonging to this run with query in MongoDB query language.

Parameters:

• query (Optional[Dict[str, Any]]) – Query in MongoDB query language.

• sorting (Optional[List[Tuple[str, Literal[1, -1]]]]) – Sorting criteria (e.g.: [(“asc_field”, 1), (“desc_field”, -1)]).

Return type:

Iterator[ModelVersionApi]

get_artifact(path_in_run='')

Get local path to run artifact.

See also

• mlopus.mlflow.BaseMlflowApi.get_run_artifact()

Parameters:

path_in_run (str) –

See path_in_run

Return type:

Path

list_artifacts(path_in_run='')

List artifacts in this run’s artifact repo.

Parameters:

path_in_run (str) –

See path_in_run

Return type:

Union[List[ObjMeta], ObjMeta]

load_artifact(loader, path_in_run='')

Load run artifact.

See also

• mlopus.mlflow.BaseMlflowApi.load_run_artifact()

Parameters:

• loader (Callable[[Path], TypeVar(A)]) – Loader callback.

• path_in_run (str) –

  See path_in_run

Return type:

TypeVar(A)

log_artifact(source, path_in_run=None, keep_the_source=None, allow_duplication=None, use_cache=None)

Publish artifact file or dir to this experiment run.

See also

• mlopus.mlflow.BaseMlflowApi.log_run_artifact()

Parameters:

• source (Union[Path, Callable[[Path], None]]) –

  See source

• path_in_run (Optional[str]) –

  See path_in_run

• keep_the_source (Optional[bool]) –

  See keep_the_source

• allow_duplication (Optional[bool]) –

  See allow_duplication

• use_cache (Optional[bool]) –

  See use_cache

log_metrics(metrics)

Log metrics to this experiment run.

Parameters:

metrics (Mapping) – See schema.Run.metrics.

Return type:

RunApi

log_model_version(model, source, path_in_run=None, keep_the_source=None, allow_duplication=False, use_cache=None, version=None, tags=None)

Publish artifact file or dir as model version inside this experiment run.

Parameters:

• model (Model | str) –

  Model name or object.

• source (Union[Path, Callable[[Path], None]]) –

  See source

• path_in_run (Optional[str]) –

  See path_in_run

• keep_the_source (Optional[bool]) –

  See keep_the_source

• allow_duplication (bool) –

  See allow_duplication

• use_cache (Optional[bool]) –

  See use_cache

• version (Optional[str]) –

  See version

• tags (Optional[Mapping]) –

  Model version tags.

  See schema.ModelVersion.tags

Return type:

ModelVersionApi

Returns:

New model version metadata with API handle.

log_params(params)

Log params to this run.

Parameters:

params (Mapping) – See schema.Run.params.

Return type:

RunApi

place_artifact(target, path_in_run='', overwrite=False, link=True)

Place run artifact on target path.

See also

• mlopus.mlflow.BaseMlflowApi.place_run_artifact()

Parameters:

• target (Path) – Target path.

• path_in_run (str) –

  See path_in_run

• overwrite (bool) –

  See overwrite

• link (bool) –

  See link

resume()

Resume this experiment run.

Return type:

RunApi

set_tags(tags)

Set tags on this run.

Parameters:

tags (Mapping) – See schema.Run.tags.

Return type:

RunApi

start_child(name=None, tags=None, repo=None)

Start a new child run.

Parameters:

• name (Optional[str]) – See schema.Run.name.

• tags (Optional[Mapping]) – See schema.Run.tags.

• repo (Union[str, Url, None]) – See repo.

update()

Use API to get latest data for this entity and update this entity in place.

Return type:

EntityApi

property children: Iterator[RunApi]

Child runs.

property repo_url: Url

Artifacts repo URL.

property url: str

The URL to this experiment run.