bucky.model.io

Monte carlo output handler.

Module Contents

Classes

BuckyOutputWriter

Class to manage the writing of raw output files and all the threading that comes with it.

Functions

init_write_thread(**kwargs)	Init write thread w/ a nonblocking stream.
merge_parquet_dataset(data_dir, **kwargs)	Merge parquet datasets w/ fastparquet.
write_parquet_dataset(df_data, data_dir, stream, pinned_mem)	Write a dataframe of MC output to parquet.

class bucky.model.io.BuckyOutputWriter(output_base_dir, run_id, data_format='parquet')

Class to manage the writing of raw output files and all the threading that comes with it.

close(self)

Cleanup and join write thread.

abstract write_historical_data(self)

TODO Write historical data used.

write_mc_data(self, data_dict)

Write the data from one MC to the output dir.

write_metadata(self, adm_mapping, str_dates, fitted_timeseries=None)

Write metadata to output dir.

abstract write_par_files(self)

TODO Copy parameter specs.

write_params(self, seed, params)

TODO WIP Write MC parameters per iteration.

bucky.model.io.init_write_thread(**kwargs)

Init write thread w/ a nonblocking stream.

bucky.model.io.merge_parquet_dataset(data_dir, **kwargs)

Merge parquet datasets w/ fastparquet.

bucky.model.io.write_parquet_dataset(df_data, data_dir, stream, pinned_mem)

Write a dataframe of MC output to parquet.