datapipe_testbench.benchmarks.dl1#

Benchmarks using DL1 information.

Classes

`HillasIntensityResolutionBenchmark`([...])	Benchmark post-cleaning Hillas intensity.
`PixelIntensityResolutionBenchmark`([...])	Benchmark pixel-wise intensities at the DL1 level, before cleaning.

class HillasIntensityResolutionBenchmark(chunk_size=10000, max_chunks=None)[source]#

Bases: ResolutionBenchmark

Benchmark post-cleaning Hillas intensity.

Initialize Resolution Benchmark.

Parameters:

input_data_levelstrq: name of input data in the InputDataset. Must be one of the fields in that structure.
reco_columnstr: name of column in the event data file to use as the reconstructed quantity
true_axisaxis: Histogram axis for the true axis of the dispersion. The name of the axis must be the name of a column in the input events data
per_tel_typebool: True if this should produce results per telescope type
resolution_methodResolutionMethod: How to compute resolution
reco_symbolstr, optional: Alternate label for the quantity, which can be in latex. If specified, this will be used in the plot labels.
num_delta_binsint: how many bins to use in the relative difference (delta) axis
delta_rangetuple[float, float]: bin range for the delta axis
chunk_sizeint: how many events to load per chunk. Higher is faster, but uses more memory.
max_chunksint: maximum number of chunks to process, or None to process all.
filter_functionCallable: function of f(true_values, reco_values) -> bool to filter what gets filled.
resolution_requirement_tablestr, optional: name of table specifying the requirement in the datapipe_testbench resources. If specified, it will be plotted
bias_requirement_tablestr, optional: name of table specifying the requirement in the datapipe_testbench resources. If specified, it will be plotted
bias_plot_rangetuple[float, float], optional: range to use when plotting bias (in units of the reco_column). If not specified it will be automatic.

class PixelIntensityResolutionBenchmark(chunk_size=10000, max_chunks=None)[source]#

Bases: ResolutionBenchmark

Benchmark pixel-wise intensities at the DL1 level, before cleaning.

Initialize Resolution Benchmark.

Parameters:

input_data_levelstrq: name of input data in the InputDataset. Must be one of the fields in that structure.
reco_columnstr: name of column in the event data file to use as the reconstructed quantity
true_axisaxis: Histogram axis for the true axis of the dispersion. The name of the axis must be the name of a column in the input events data
per_tel_typebool: True if this should produce results per telescope type
resolution_methodResolutionMethod: How to compute resolution
reco_symbolstr, optional: Alternate label for the quantity, which can be in latex. If specified, this will be used in the plot labels.
num_delta_binsint: how many bins to use in the relative difference (delta) axis
delta_rangetuple[float, float]: bin range for the delta axis
chunk_sizeint: how many events to load per chunk. Higher is faster, but uses more memory.
max_chunksint: maximum number of chunks to process, or None to process all.
filter_functionCallable: function of f(true_values, reco_values) -> bool to filter what gets filled.
resolution_requirement_tablestr, optional: name of table specifying the requirement in the datapipe_testbench resources. If specified, it will be plotted
bias_requirement_tablestr, optional: name of table specifying the requirement in the datapipe_testbench resources. If specified, it will be plotted
bias_plot_rangetuple[float, float], optional: range to use when plotting bias (in units of the reco_column). If not specified it will be automatic.

datapipe_testbench.benchmarks.dl1#

This Page