This program ensures that all directories and files are present and will extract the data from the fits files into internal directories for future faster access. On clusters it will also distribute the data among the different nodes. This program is only launched once at the begninning of the processing, as long as the data/position/flag are unchanged.

This program determines the map parameters and size and compute the pixel index of all bolometers within this map, properly handling the flagged data into separate maps (if needed). It can also perfom coordinate transformation between J2000 Equatorial and Galactic system.

This program will process data written by sanePre, saneProc and saneInv to compute the maximum likelihood solution of the map-making equation. It is performed iteratively, using a conjugate gradient approach.

The signal table contains the calibrated signal in unit of the final map. It can include glitches and other artefacts as long as their are masked. The mask table should include all non physical effect. The convention used for the mask is 0 for good data, any other value would be treated as bad data.

The lon and lat contains either the right ascension and declinaison or the galactic longitude and latitude, of all bolometers at all timestamps, following the ordering of the signal and mask images. They must be in degree. This format is referred as the HIPE format in the ini file below.

The offsets table consists of 3 columns, names, dX, dY, and nBolo lines, describing the relative offsets of the bolometers. The reference position contains 3 rows, lon, lat & phi, and nTime lines, describing the position on the sky of the reference offset and the position angle between the bolometers and sky plane. The offsets should be provided using a TAN projection. This format is referred as the SANEPIC format in the ini file

The channel table has the same order as the nBolo dimension in the signal and mask images. This is only an indexing of the channels and will be used later to select group of bolometers to produce maps with SANEPIC, for e.g. jackknife maps.

Some keyword are searched in the primary header, namely EQUINOX, DATE-OBS, RADESYS, RESTWAV, RESTFRQ, TIMESYS, CREATOR, INSTRUME, OBJECT, TELESCOP, OBSERVER, UNIT. This keyword will be propagated into the final fits file.

The common section is used to describe the fields used by all programs, mainly input output and temporary directories, lists of fits files and bolometers

The format of the input files and an explanation of the suffix mechanism will follow

All quantities related to positions and map will be here, pixel size, projection center, projection type, coordinate type, coordinates transformation (galactic/equatorial), file format, masks and mapping of flagged data

It is also possible to give a mask file as input, SANEPIC will then use it to remove crossing constrains on non-zero pixels and use its World Coordinate System header to construct the final map.

This is mainly a path and a suffix to the noise power spectra

All parameters related to the pre-processing of the data is summarized here, including the sampling frequency, order of the polynomial baseline, the cut-off frequency of the high-pass butherworth filter, the threshold frequency to be applied on the noise power spectra.

Most of the key parameter are set in [saneProc], so here are only the maximum number of iteration, a possible dumping of results during processing and the filename prefix of the maps

This file lists the input fits data files, one per line. An optionnal column can be added to assign file to processor/rank when using the parallelization per frame (note that the fits file can be compressed).

This file lists the bolometers used in the SANEPIC processing, this can be a subset of the bolometer present in the data files. You can define either a common list for all data file (bolo_global_file), or a different list for each data file following the *_suffix mechanism. In this case your lists should be named file basename+bolo_suffix, i.e. if your data file is named fileA.fits, then the corresponding list of bolometer file will be named fileA.bolo.

These files contain the auto- and cross- power spectra of the noise. Thus for bolometer $i$ and $j$, the cross-power spectra between pure noise timeline (in Fourier space) $d_i$ and $d_j$ is defined as (see Patanchon et al. 2008)

$P_{ij}(w) = \frac{1}{n_{q}} \sum_{w_{min}(q)}^{w_{max}(q)}{\tilde{d}_{iw}^* \tilde{d}_{jw}}$

The fits file must contains 3 extensions, channels, equivalent to the table found in the data fits file, frequency a 1D image of size nFreq+1 listing the bins edge in frequency for the power spectra and finally PowerSpectra, which, for simplicity, contains all auto and cross power spectra, stored in a single 2D image of size nFreq x (nBolos x nBolos)

Thus, a typical (1.2 MB) Noise Noise Power Spectra file, describing nBolos=41 with nFreq=99 will look like :

This file list the threshold frequencies (in Hz), applied to the noise-noise power spectra, for each input fits data files or frames. It must follows the same order. Thus a typical file would read as :