LALInference pipeline integration for Asimov.

This package provides a plugin for Asimov 0.7+ that enables integration with the LALInference parameter estimation pipeline for gravitational wave data analysis.

• 🔌 Plugin Architecture: Seamlessly integrates with Asimov via entry points
• 🚀 Scheduler-agnostic: Automated DAG generation and job submission via Asimov's HTCondor/Slurm scheduler API
• 📈 Result Collection: Automatic collection of posterior samples and results
• 🧪 Well Tested: Unit tests plus a genuine end-to-end test (real lalinference_pipe DAG generation and HTCondor execution against LALInference's own simulated noise)

LALInference itself is only distributed via conda-forge — there is no PyPI wheel for it — so installing this plugin is a two-step process:

conda install -c conda-forge lalinference
pip install asimov-lalinference

conda install -c conda-forge lalinference
git clone https://github.com/transientlunatic/asimov-lalinference.git
cd asimov-lalinference
pip install -e .

pip install -e ".[docs,test]"

Once installed (alongside a working lalinference_pipe, from the conda-forge lalinference package), the LALInference pipeline is automatically available in Asimov via its entry-point registry — no further configuration is required beyond a normal Asimov production blueprint. See the documentation for a full example blueprint.

• Python >= 3.10 (the conda-forge lalinference feedstock does not build for older Pythons)
• asimov >= 0.7.0
• conda-forge::lalinference — install separately with conda install -c conda-forge lalinference (there is no asimov[gw] extra in asimov core today that would pull this in automatically; each GW pipeline plugin, including this one, currently needs to be installed explicitly)

lalinference_pipe generates a small lalinf_touch_output placeholder script per engine node and wires it in via custom +PreCmd/+PreArguments job classads, so that any declared output file the real executable doesn't happen to write for a given run (e.g. the SNR summary file, which isn't always produced) still exists as an empty placeholder by the time HTCondor transfers job outputs — otherwise the transfer hard-fails with something like:

Transfer output files failure ... Details: 1 total failures: first failure: reading from
file .../lalinferencenest-...-1.hdf5_snr.txt: (errno 2) No such file or directory

+-prefixed classads aren't a native HTCondor pre-exec hook — real IGWN/LIGO condor pools apparently have a USER_JOB_WRAPPER configured site-wide that interprets this convention, but a vanilla HTCondor install (e.g. the htcondor/mini test image this repo's own CI uses, or your own self-hosted pool) does not, and lalinf_touch_output silently never runs. If you hit the error above on a pool you control, configure a USER_JOB_WRAPPER that reads PreCmd/PreArguments off the job ad (available via the $_CONDOR_JOB_AD file HTCondor provides to every job) and runs it before the real executable — see .github/workflows/e2e.yml in this repo for a working example.

MIT License - see LICENSE file for details.

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.