Running the Code

There are several distinct parts of the code that correspond to different steps:

1. Generating analysis products (beam transfers, KL modes, PS estimators)
2. Simulating data (creating maps, generating timestreams)
3. Analysing data (making maps, estimating powerspectra)

Below I'll try and give a brief overview of how to run each.

Generating Products

There is a reasonably stream lined approach to generating all the intermediate products required to do forecasting and data analysis. There is a reasonably obvious (but at the moment undocumented) configuration file format that allows you to set everything up and run easily. There is a well documented example here.

When you've setup a params file, you can fun it with:

python ./scripts/simulate.py param_file_name.yaml

If you're on a cluster (such as SciNet) and have a large problem, you'll probably want to run this as a batch job. If you run

python ./scripts/mkjob_simulate.py param_file_name.yaml

it will create the required directories, make a job script, and then submit it to the queue listed in the file.

This runs through and creates the beamtransfers. The relevant directories that it creates are:

./bt/                      # Beam transfers
./bt/beam_f/               # Frequency ordered
./bt/beam_f/00/            # First frequency
./bt/beam_f/00/beam.hdf5   # All the beams for frequency 00.

./bt/beam_m/               # m-ordered beams
./bt/beam_m/30/            # All the beams for m=30
./bt/beam_m/30/beam.hdf5   # The B matrices for m=30
./bt/beam_m/30/svd.hdf5    # The SVD beams

Simulated Data

simulations_21cm: make sky, foregrounds

python ./simulations_21cm/scripts/mk_sky_21cm.py

import healpy m1 = hdf5 file freq X pixel healpy.mollview(m1[0])

healpy.anafast(m1[0]) for ang. power spec.

for timestream: /drift/pipeline/mk_timestream

python mk_timestream.py loc_config_file output_dir --map 21cm_map.hdf5 --noise n_days

pipeline: testpipe.py