An Asimov plugin for running PyCBC Inference parameter estimation.

📚 Full documentation and tutorial

This plugin enables Asimov to submit and manage pycbc_inference jobs for gravitational wave parameter estimation. It provides:

• Configuration templating: renders a pycbc_inference config file from a bundled Liquid template, populated from a production's ledger metadata (config_template), for use by asimov manage build.
• Scheduler integration: submits jobs via Asimov's own scheduler abstraction (self.scheduler, from asimov.pipeline.Pipeline / asimov.scheduler_utils), which supports both HTCondor and Slurm -- rather than talking to a scheduler's API directly.
• Status tracking: monitors job completion by checking for a genuine, readable posterior samples file.
• Checkpoint-aware resurrection: pycbc_inference checkpoints its own progress; a failed or evicted job is resubmitted so it resumes from that checkpoint rather than starting over.
• Dependency-driven post-processing: once a job completes, its samples are available (via collect_assets()) to any downstream production with a needs: dependency on it -- for example an asimov-pesummary production. This plugin doesn't submit that job itself; Asimov's own dependency resolution does.

Requires asimov>=0.7. This plugin does not depend on the pycbc Python package itself -- like the sibling asimov-lalinference and asimov-bayeswave plugins, it shells out to the pycbc_inference executable, which is expected to be installed separately (e.g. via conda) in the environment pointed at by Asimov's [pipelines] environment config option.

pip install asimov-pycbc

For development:

git clone https://github.com/etive-io/asimov-pycbc
cd asimov-pycbc
pip install -e ".[test]"

Once installed, Asimov discovers this plugin automatically via its asimov.pipelines entry point. Add a production to an event with pipeline: pycbc:

kind: analysis
name: pycbc-imrphenompv2
pipeline: pycbc
waveform:
  approximant: IMRPhenomPv2
  reference frequency: 20
sampler:
  sampler: dynesty
  sampler kwargs:
    nlive: 2000
    dlogz: 0.1
scheduler:
  accounting group: ligo.dev.o4.cbc.pe.pycbc
  processes: 8

asimov manage build submit will render a pycbc_inference ini from the bundled template (unless one already exists in the event repository) and submit the job.

Strain data is read from a production's data metadata using the same conventions as the other Asimov gravitational-wave pipeline plugins (e.g. asimov-gwdata, asimov-lalinference, asimov-bayeswave):

• data.frame types / data.channels -- real data via datafind lookup
• data.cache files -- a pre-built LAL-format frame cache
• data.frame files -- explicit frame file paths
• data.fake strain (e.g. aLIGOZeroDetHighPower) -- simulated Gaussian noise, no real strain data required

This means a production populated by a data-retrieval step (for example asimov-gwdata, which writes data.cache files/data.frame files into a production's metadata) is picked up automatically, without any extra wiring: data-retrieval -> pycbc -> pesummary is expressed entirely through Asimov's own needs: dependency mechanism.

This plugin does not run PESummary (or anything else) itself. When a pycbc production completes, after_completion() only marks it finished -- post-processing is expressed as a separate production with a needs: dependency on it, e.g.:

kind: analysis
name: pycbc-test-pesummary
pipeline: pesummary
needs:
  - pycbc-test
waveform:
  approximant: IMRPhenomD
  reference frequency: 30
  minimum frequency:
    H1: 30
postprocessing:
  pesummary:
    multiprocess: 2

Asimov's own dependency resolution builds and submits pycbc-test-pesummary once pycbc-test reaches finished; PESummary picks up its samples via pycbc-test's collect_assets() (through production._previous_assets()). Install asimov-pesummary into the same environment as pycbc itself -- summarypages needs to import pycbc to read pycbc_inference's native HDF5 format.

Unit tests (mocked, no external dependencies):

pip install -e ".[test]"
pytest

.github/workflows/e2e.yml also runs a genuine end-to-end test: a real pycbc_inference run (using PyCBC's own simulated --fake-strain noise so no real strain data is required) submitted through a real HTCondor scheduler, waiting for a real, parseable posterior samples file -- then a real asimov-pesummary production, wired up via needs:, that consumes those samples and produces a real, parseable summarypages output.

.github/workflows/docs.yml also checks that the subcommand and flags used by every asimov ... command shown in docs/*.rst still exist on the live, installed asimov CLI, on every pull request (scripts/lint_tutorial_commands.py). It doesn't validate positional arguments or actually run anything, but it does mean a renamed subcommand or flag in a future asimov release shows up as a CI failure here rather than as a tutorial that silently stops working:

pip install -e .
python scripts/lint_tutorial_commands.py

Contributions welcome! Please submit issues or pull requests to etive-io/asimov-pycbc.

MIT -- see LICENSE.