Skip to main content

Quick Run

This guide takes you from a fresh install to a first run. The quick mode is a smoke test: it shows that the pipeline works end to end, not what the data say.

Prerequisites

The package is installed (Installation) and you are in the repository root.


Step 1 — Place the data​

Four files go in data/raw/ (the folder is git-ignored apart from its README):

Pantheon+SH0ES.dat
Pantheon+SH0ES_STAT+SYS.cov
desi_gaussian_bao_ALL_GCcomb_mean.txt
desi_gaussian_bao_ALL_GCcomb_cov.txt
python scripts/download_data.py # fetches only the two DESI files

The two Pantheon+SH0ES files are external products and must be obtained separately from the official release and copied in by hand (Pantheon+SH0ES).

Step 2 — Smoke test​

python src/xi_protocol1b_bao_run.py --quick # or: make quick · scripts/run_quick.sh

Quick mode runs 500 steps with burn-in 100 and 32 walkers for each of the three models.

Step 3 — Medium run (optional)​

python src/xi_protocol1b_bao_run.py --nsteps 5000 --burn-in 1000 --nwalkers 64

Step 4 — Final run​

python src/xi_protocol1b_bao_run.py --nsteps 100000 --burn-in 20000 --nwalkers 96

This is the command behind the published results. Capture the console log if you want to keep the progress and diagnostics:

bash scripts/run_long_final.sh # same command, output copied to run_100000_96.log

The Ξosc chain is the long one; scripts/run_xi_only_long.ps1 repeats it for Ξosc alone (--models Xi_osc) when only that convergence check is needed. Wall-clock time depends on the machine and is not recorded in the exported facts.


Expected outputs​

Every run writes to the default results/ and figures/ folders — including quick runs. A new run therefore overwrites the previous one there.

Do not cite a quick run

A quick run's summary_protocol1b_combined.txt reads Quick mode: True, and its numbers differ from the final ones — a 500-step chain is not converged for Ξosc. The authoritative results live in results/protocol1b_pantheon_desi/ (Outputs).

For a converged Ξosc chain the practical criterion is N/50 > τmax; for the final run this is 2,000 > 1504 (Convergence Diagnostics).

Troubleshooting​

ErrorCause
Missing Pantheon filePantheon+SH0ES.dat not in data/raw/ (or pass --sne-data)
Required Pantheon+ columns not foundfile lacks zHD, MU_SH0ES, MU_SH0ES_ERR_DIAG
Unexpected Pantheon covariancethe .cov file is truncated or not the STAT+SYS matrix
No BAO rows could be readDESI mean file in an unexpected format

Next steps​

Outputs — which files to trust and how to regenerate the tables.