Cruncher

Cruncher class

from tensorflow.summary import SummaryWriter, create_file_write, scalar # type: ignore

Cruncher

 Cruncher (agent:tspace.agent.dpg.DPG, truck:tspace.config.vehicles.Truck,
           driver:tspace.config.drivers.Driver, resume:bool=False,
           infer_mode:bool=False, data_dir:Optional[pathlib.Path]=None, tr
           ain_summary_writer:Optional[tensorflow.python.ops.summary_ops_v
           2.SummaryWriter]=None, logger:Optional[logging.Logger]=None,
           dict_logger:Optional[dict]=None)

*Cruncher is the processing unit of the data flow.

It consumes data (pd.DataFrame) from the observe pipeline and produces data (pd.DataFrame) into the flash pipeline.

Attributes:

- agent: abstract base class DPG, available DDPG, RDPG
- truck: Truck object
- driver: Driver object
- resume: bool, whether to resume training
- infer_mode: bool, whether only inferring or with training and inferring
- data_dir: Path, the local path to save all generated data
- train_summary_writer: SummaryWriter, Tensorflow training writer
- logger: Logger
- dict_logger: logger format specs*

source

Cruncher.__post_init__

 Cruncher.__post_init__ ()

Set logger, Tensorflow data path and running mode

source

Cruncher.filter

 Cruncher.filter (in_pipeline:tspace.dataflow.pipeline.queue.Pipeline[pand
                  as.core.frame.DataFrame], out_pipeline:tspace.dataflow.p
                  ipeline.queue.Pipeline[pandas.core.frame.DataFrame],
                  start_event:Optional[threading.Event],
                  stop_event:Optional[threading.Event],
                  interrupt_event:Optional[threading.Event],
                  flash_event:Optional[threading.Event],
                  exit_event:Optional[threading.Event])

*Consume data from the pipeline

Args: in_pipeline: Pipeline, the input pipeline out_pipeline: Pipeline, the output pipeline start_event: Event, start event stop_event: Event, stop event interrupt_event: Event, interrupt event flash_event: Event, flash event exit_event: Event, exit event*

	Type	Details
in_pipeline	Pipeline	input pipeline
out_pipeline	Pipeline	output pipeline
start_event	Optional	input event start
stop_event	Optional	input event stop
interrupt_event	Optional	input event interrupt
flash_event	Optional	input event flash
exit_event	Optional	input event exit
Returns	None