API Reference¶

class asimov_pesummary.pesummary.PESummary(production, category=None)[source]¶

Bases: Pipeline

A postprocessing pipeline add-in using PESummary.

A production using this pipeline may be either a regular analysis (post-processing a single upstream production’s samples) or an asimov SubjectAnalysis (combining several productions’ samples into one set of summary pages). SubjectAnalysis productions marked refreshable: true are automatically resubmitted by asimov’s monitor loop whenever their resolved source analyses change; when that happens here, only the newly-added analyses are passed to summarypages, using its own --add_to_existing/--existing_webdir flags to append them to the already-published pages rather than recombining everything from scratch. If an analysis is ever removed from the combined set, summarypages cannot retract a label from an existing page in place, so a full rebuild is triggered instead.

property config_template¶

A minimal bundled config template.

Asimov’s generic manage build step calls production.make_config() to render a production’s own .ini from a template, whenever one doesn’t already exist in the event repository (see repository.find_prods()). Without this, asimov falls back to looking for a pesummary.ini template inside its own package, which doesn’t exist there any more – PESummary moved out of asimov core into this plugin – so make_config() raises TemplateNotFound (see FakeCBCPipeline.config_template in testing.py, which hit and documented this same gap first). The rendered file is what summarypages --config receives (see _submit_single_analysis/_submit_subject_analysis); PESummary’s own actual behaviour is controlled by the explicit CLI flags built from self.meta elsewhere in this class, so this only needs to be a valid, renderable ini for provenance/display purposes.

results()[source]¶

Fetch the results file from this post-processing step.

A dictionary of results will be returned with the description of each results file as the key. These may be nested if it makes sense for the output, for example skymaps.

For example:

{'metafile': '/home/asimov/working/samples/metafile.hd5',
 'skymaps': {'H1': '/another/file/path', ...}
}

Returns:

dict

A dictionary of the results.

collect_assets()[source]¶

Advertise this pipeline’s combined metafile, in case a further downstream step ever needs to consume it.

build_dag(user=None, dryrun=False)[source]¶

No-op: PESummary has no separate build step. All of the work happens in submit_dag, but asimov’s generic manage build submit CLI unconditionally calls build_dag on every pipeline before submit_dag, so this must exist.

submit_dag(dryrun=False)[source]¶

Run PESummary on the results of this job.