output_quantities¶
- lsst.verify.output_quantities(package_name, quantities, suffix=None, include_metrics=False, metrics_package='verify_metrics')¶
Output measurements, as
astropy.units.Quantity
objects, from a pipeline task execution to alsst.verify
-formatted JSON file.- Parameters:
- package_name
str
Name of the package producing measurements. This name is used two ways:
Make fully-qualified metric names from keys in the
quantities
dictionary. For example, if aquantities
dict has a key-value pair{'PA1': 5 * u.mmag}
andpackage_name='validate_drp'
, the fully-qualified metric name is'validate_drp.PA1'
.As a filename prefix for the output JSON file.
- quantities
dict
ofastropy.units.Quantity
values Dictionary of measurements as plain
astropy.units.Quantity
instances. Each key is the name of a metric. If metric names are not fully-specified (inpackage.metric
format), the package name can be provided with thepackage_name
argument.- suffix
str
, optional Additional suffix to add to the output JSON filename:
{package_name}_{suffix}.verify.json
The suffix may be used to distinguish measurement output files from different tasks in the same package.
- include_metrics
bool
, optional Metric and specification definitions are included in the JSON output if set to
True
. The metric and specification definitions are loaded from a metric package indicated by themetrics_package
argument. Normally tasks do not need to include metric definitions if a post-processing step is used. Default:False
.- metrics_package
str
, optional Name of the metrics package to obtain metrics from if
include_metrics
isTrue
. Default is'verify_metrics'
.
- package_name
- Returns:
- filename
str
Filename where the JSON file was written.
- filename
See also
Notes
This function is designed for lightweight
lsst.verify
framework usage. Rather than maintainingJob
andMeasurement
objects, a task can simply record metric measurements asastropy.units.Quantity
objects. Withoutput_quantities
, the task can output these measurements in a standardizedlsst.verify
JSON format. Post-processing tools can load this data for local analysis, or submit it to the https://squash.lsst.codes dashboard service.Tasks that need to include
Blob
s,Measurement.extras
or queryMetric
objects should create aJob
instance and useJob.write
instead.