Skip to content

Support custom job manager classes in initialize_from_df  #667

@soxofaan

Description

@soxofaan

@mbuchhorn asked for feature to properly support custom job manager classes in initialize_from_df. Which is currently hard-linked to MultiBackendJobManager

# TODO: option to provide custom MultiBackendJobManager subclass with custom normalize?
if self.exists():
if on_exists == "skip":
return self
elif on_exists == "error":
raise FileExistsError(f"Job database {self!r} already exists.")
else:
# TODO handle other on_exists modes: e.g. overwrite, merge, ...
raise ValueError(f"Invalid on_exists={on_exists!r}")
df = MultiBackendJobManager._normalize_df(df)

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions