schrodinger.seam.testing.test_pipelines module¶

class schrodinger.seam.testing.test_pipelines.BenchFlow(num_outputs: int, fanout_rate: int, output_size_bytes: int, with_local_exec_dofn: bool = False)¶

Bases: PTransform

A workflow for benchmarking the performance and memory of a pipeline.

This workflow generates random strings of a given size and number of outputs.

__init__(num_outputs: int, fanout_rate: int, output_size_bytes: int, with_local_exec_dofn: bool = False)¶

static generate_random_strings(_: None, size: int, num_out: int) → Iterable[str]¶

class LocalRandomStringsDoFn(*unused_args, **unused_kwargs)¶: Bases: _LocalOnlyDoFn, LocalRandomStringsDoFn

expand(pbegin)¶

class schrodinger.seam.testing.test_pipelines.GPUTransform(label: Optional[str] = None)¶

Bases: PTransform

expand(pcoll)¶

class schrodinger.seam.testing.test_pipelines.CPUTransform(label: Optional[str] = None)¶

Bases: PTransform

expand(pcoll)¶

class schrodinger.seam.testing.test_pipelines.GPUExampleWorkflow(output_path: str)¶

Bases: PTransform

A simple workflow that demonstrates the use of GPU resources in a Beam pipeline.

Writes a JSON file with the hostname of the machine where the GPU transforms and the CPU transform are executed.

__init__(output_path: str)¶

expand(pcoll)¶