pipeline.infrastructure.renderer.regression

The regression module contains base classes and plugin registry for the pipeline’s regression test value extractor framework.

This module contains two classes:

• RegressionExtractor: the base class for extractor plug-ins

• RegressionExtractorRegistry: the registry and manager for plug-ins

Tasks provide and register their own extractors that each extends RegressionExtractor. These extractor plug-ins analyse the results of a task, extracting pertinent values and writing them to a dict. The keys of the output dict identify the value; the values of the dict are the extracted values themselves.

The pipeline QA framework is activated whenever a Results instance is accepted into the pipeline context. The pipeline QA framework operates by calling is_handler_for(result) on each registered QAPlugin, passing it the the accepted Results instance for inspection. QAPlugins that claim to handle the Result are given the Result for processing. In this step, the QA framework calls QAPlugin.handle(context, result), the method overridden by the task-specific QAPlugin.

Functions

extract_qa_score_regression(prefix, result)
extract_regression_results(context)
get_prefix(result)
key_intersection(d1, d2)	Compare keys of two dicts, returning duplicate keys.
union(d1, d2)	Return the union of two dicts, raising an exception if duplicate keys are detected in the input dicts.

Classes

ApplycalRegressionExtractor()	Regression test result extractor for applycal tasks.
FluxcalflagRegressionExtractor()	Regression test result extractor for hifa_gfluxscaleflag.
GcorFluxscaleRegressionExtractor()	Regression test result extractor for hifa_gcorfluxscale
RegressionExtractor()	RegressionExtractor is the mandatory base class for all regression test result extractors.
RegressionExtractorRegistry()	The registry and manager of the regression result extractor framework.