Workflow: the inspiral analysis workflow generator (pycbc.workflow)


Pycbc’s workflow module is a tool used to create the workflows needed to perform coincident analyses of gravitational-wave data searching for compact-binary-coalescences using matched-filtering.

The pycbc workflow runs through a number of stages, which are designed to be as independent of each other as possible, while also allowing an integrated end-to-end workflow to be constructed.

Documentation of the workflow module and how to run it can be found below.

Please see the following poster, presentied at the March LVC meeting, 2014, for an introduction to the workflow module (referred to as “ahope”). Especially see the following top-level workflow generation model.


Workflow module documentation

The following contains a list of the sub-modules in pycbc’s workflow module, a bried description of what each does, and a link to documentation on how to use each module. Collectively these pages should provide complete details on how to set up a workflow from scratch.

Basics and overview

The following page gives a description of how the workflow module expects configuration files to be layed out and how the basic command-line interface functions. If writing a workflow for the first time we recommend you read through this page!

Generating segments

Obtain the science segments and data-quality segments from making queries to a segment database.

Obtaining data

Run queries to the datafind server to find the needed frames and test these for consistency if desired

Injection generation

Generate injection files for use later in the analysis

Template bank

Construct a template bank, or banks, of CBC waveforms that will be used to matched-filter the data with.

Split table

Split an output file into numerous parts to allow parallel analysis. Normally used to split the template bank up to allow matched-filtering in parallel


Perform the matched-filters and calculate any signal-based consistency tests that should be calculated.


Determine if “triggers” seen in one detector are also seen in other detectors. Also check for coincidence between time-slid triggers for background evaluation