2.1. mbl.workflow.etl#

Classes

ETL()

class mbl.workflow.etl.ETL[source]#

Bases: object

class Metadata(database: str = 'random_heisenberg', ed_table: str = 'ed', tsdrg_table: str = 'folding_tsdrg', s3_path: str = 's3://many-body-localization/gap_ratio', gap_ratio_table: str = 'gap_ratio')[source]#

Bases: object

Parameters

  • database (str) –

  • ed_table (str) –

  • tsdrg_table (str) –

  • s3_path (str) –

  • gap_ratio_table (str) –

Return type

None

database: str = 'random_heisenberg'#
ed_table: str = 'ed'#
tsdrg_table: str = 'folding_tsdrg'#
s3_path: str = 's3://many-body-localization/gap_ratio'#
gap_ratio_table: str = 'gap_ratio'#
__init__(database='random_heisenberg', ed_table='ed', tsdrg_table='folding_tsdrg', s3_path='s3://many-body-localization/gap_ratio', gap_ratio_table='gap_ratio')#
Parameters

  • database (str) –

  • ed_table (str) –

  • tsdrg_table (str) –

  • s3_path (str) –

  • gap_ratio_table (str) –

Return type

None

classmethod create_gap_ratio_table(params)[source]#

CREATE TABLE AS SELECT (CTAS) approach to dump results from fetch_gap_ratio() into another table. This method will run in parallel with ray.

Parameters

params (Sequence[Dict]) – List of dictionary that contains kwargs to fetch_gap_ratio().

Return type

pandas.core.frame.DataFrame

Returns:

Notes

  1. Both averaging orders will be considered.

    There is no need to provide the order in params.

  2. Performance may be bounded by the bandwidth of internet.

    Please adjust the number of cpus in ray.init() accordingly.