Working with time-ordered data (TOD)

The TOD class is the basic container for detector and pointing data. Populating a TOD with data can be done a number of different ways, so loading TODs is normally done through the moby2.scripting.get_tod function, which is able to decode complicated parameters and handle the peculiarities of different experiments.

Quick start

To load all of the detector and pointing data for a dirfile:

import moby2
filename = '1234560000.1234569999.ar1'
tod = moby2.scripting.get_tod({'filename': filename})

Look at the data and pointing info:

>>> print tod.data.shape
(1056, 256000)
>>> print tod.data[0]         # First detector data
[1.34e3, 1.33e3, ...]
>>> print tod.az
[...]
>>> print tod.alt
[...]

To load only certain samples, use the start and end arguments. These have the same semantics as array indices [start:end], including the handling of negative arguments:

# Load samples 200 to 1200 samples
tod = moby2.TOD.from_dirfile({'filename': filename,
                              'start': 200, 'end': 1200})

# Load entire TOD except for the last 400 samples:
tod = moby2.TOD.from_dirfile({'filename', 'end': -400})

To select particular detectors, use det_uid or rows and cols with read_block:

# Load detector channel 123
tod = moby2.TOD.from_dirfile({'filename': filename, 'det_uid': [123]})

# Load multiplexing row 2
tod = moby2.TOD.from_dirfile({'filename': filename, 'rows': [2]})

# Load only (row 2, col 3) and (row 8, col 0) by passing block_read=False
tod = moby2.TOD.from_dirfile({'filename': filename,
                              'rows': [2,8], 'cols': [3,0],
                              'block_read': False})

Full class documentation

class moby2.TOD(n_samples, n_dets=None, create_data=True, create_pointing=True, info=None)

abuses = None

A place to log processing operations

alt = None

Boresight altitude in radians (n_samp)

az = None

Boresight azimuth in radians (n_samp)

copy(copy_data=True, copy_pointing=True, copy_id=True, copy_cuts=True, copy_info=True, resample=1, resample_filter=True, resample_offset=None)

Return a copy of the TOD. A complete copy is made of all detector and pointing data unless the corresponding copy_* field is passed in.

This routine can also resample a TOD. The TOD is resampled at every Nth index, where N is a (possibly non-integral) value given by the resample parameter. If resample_filter=True, then a low-pass filter is used to resample the detector data… but that only currently works if resample is a power of 2.

ctime = None

Time, as unix timestamp (n_samp)

cuts = None

A preferred TODCuts object for the TOD

data = None

The detector data, a numpy array of shape (n_det,n_samp)

det_uid = None

Detector uids; integer array of shape (n_dept)

enc_flags = None

Pointing validity indicator (n_samp)

classmethod from_dirfile(*args, **kwargs)

get_hk(fields, filename=None, start=None, count=None, dtype='float32', fix_gaps=False)

Load housekeeping (or other) fields from a dirfile. Assuming that the TOD was loaded using TOD.for_dirfile, this function will look to the same dirfile and load the data corresponding to the same sample indices.

“fields” can be a string, or a list of string. The function will return an array or a list of arrays, respectively.

If you need to override the dirfile name, pass “filename”. If you want different starting or stopping indices, use “start” and “count”. To manipulate the output data type, set dtype (our wrapping of getdata probably only supports float32, float64, int32, int64, and maybe some unsigned ints).

Set fix_gaps=True to attempt patching of non-400 Hz ACTPol HK data in cases where the encoder tick dropped out for a while. Missing samples will be interpolated.

info = None

A TODInfo object providing misc. TOD metadata

classmethod init_from_dirfile(filename=None, start=None, end=None, channel_names=[], hk_spec=[], extractor=None, read_data=True)

Create a TOD object with data taken from a (probably ACT-style) dirfile. Since we use libactpol calls, the dirfile can be slimmed and zipped.

nsamps = None

n_samp