odc.stac.bench.BenchmarkContext

class odc.stac.bench.BenchmarkContext(cluster_info, npix, nbytes, dtype, shape, chunks, crs, transform, scenario='', temporal_id='', method='undefined', extras=<factory>)[source]

Benchmark Context Metadata.

Normalized representation of the task being benchmarked and the environment it is benchmarked in.

__init__(cluster_info, npix, nbytes, dtype, shape, chunks, crs, transform, scenario='', temporal_id='', method='undefined', extras=<factory>)

Methods

__init__(cluster_info, npix, nbytes, dtype, ...)

render_timing_info(times[, col_width])

Render textual representation of timing result.

render_txt([col_width])

Render textual representation for human consumption.

to_pandas_dict()

Extract parts one would need for analysis of results.

Attributes

chunk_signature

Render textual representation of chunk shapes.

data_signature

Render textual representation of data shape and type.

method

Some human name for method, (stackstac, odc-stac, rio-xarray)

resolution

Extract resolution.

scenario

Some human name for this benchmark data

temporal_id

Time period covered by data in human readable form

cluster_info

client.scheduler_info().copy()

nworkers

Number of workers in the cluster

nthreads

Number of threads in the cluster across all workers

total_ram

Total RAM across all workers of the cluster

npix

Total number of output pixels across all bands/timeslices

nbytes

Total number of output bytes across bands/timeslices

dtype

Data type of a pixel, assumed to be the same across bands

shape

Normalized shape in time,band,y,x order

chunks

Normalized chunk size in time,band,y,x order

crs

Projection used for output

transform

Linear mapping from pixel coordinates to CRS coordinates

extras

Any other parameters to capture