Mock code for simulating atmosphere monitoring detectors

sim_atm_dets.py

"""
Pseudocode example as a talking point for discussion.

Code similar to this could be dropped into a copy of `toast_so_sim`, or into a notebook where the data is being loaded
manually with the LoadContext operator.
"""

#-------------------------------------------------------------------------------------------------------------
# Load data into observations.  Either in a notebook with LoadContext directly or in a workflow script calling
# `load_or_simulate_observing()`.
#-------------------------------------------------------------------------------------------------------------


# Now for each observation we are going to:
#
#     1. Delete any detector data, since this contains the wrong number of detectors.
#     2. Add new detectors to the focalplane with a custom bandpass
#

# Wipe any starting detector data (it will get re-created by the simulation operators).
toast.ops.Delete(detdata=[defaults.det_data, defaults.det_flags]).apply(data)

# Detectors to add (assuming same names on all wafers)
monitor_dets = {
    "det_atm_01": {
        "quats": np.array([0, 0, 0, 1]), # Change to the actual boresight offset...
        "bandcenter": 90.0 * u.GHz,
        "bandwidth": 5.0 * u.GHz,
        "psd_net": 0.05 * u.K * np.sqrt(1 * u.second),
        # Add other properties too...
    },
    "det_atm_02": { # you get the idea...
    },
}

for obs in data.obs:
    # Note that by default for SO, each observation is a single wafer.
    
    # Get the focalplane
    fp = obs.telescope.focalplane
    
    # Get the astropy table of detector properties
    det_props = fp.detector_data
    
    # Add rows for our monitor dets.  This modifies the table in place.
    for mon_name, mon_props in monitor_dets.items():
        # Copy an existing row, modify with this current dets properties, then add the new row...
        det_props.add_row( ... )
    
    # Make a new focalplane with this updated detector table
    new_fp = toast.Focalplane(
        detector_data=det_props,
        field_of_view=fp.field_of_view,
        sample_rate=fp.sample_rate,
        thinfp=fp.thinfp,
    )
    
    # Replace the observation focalplane with this new one
    obs.telescope.focalplane = new_fp

#-------------------------------------------------------------------------------------------------------------
# Simulate atmosphere for all detectors (including the atmosphere monitor dets).  This can be done manually in a
# notebook or in a workflow script by calling `wrk.simulate_atmosphere_signal()`.  See toast_so_sim for an example
#-------------------------------------------------------------------------------------------------------------

#-------------------------------------------------------------------------------------------------------------
# Simulate instrumental noise for all detectors (including the monitor dets).
#-------------------------------------------------------------------------------------------------------------

# Now the monitor dets have the desired signal, and we flag them so that the rest of the simulated sky components
# do not get added to those dets.

for obs in data.obs:
    obs.set_local_detector_flags(
        {x: defaults.det_mask_invalid for x in monitor_dets}
    )

#-------------------------------------------------------------------------------------------------------------
# Simulate the rest of the sky components...
#-------------------------------------------------------------------------------------------------------------

# Data reduction.  In practice, one would need to calibrate these monitor dets differently...
# Then use the monitor dets to build the spatial model of the atmosphere power across the focalplane.  This can
# then be used as a 2D filter or a destriping template.

Note, if the monitor dets are replacing normal detectors, just modify existing rows of the table, rather than adding rows.

Another note- simulate the AZSS before flagging the dets.