See work by Scott Dodelson
The URL for the obs_decam package is here.
There is currently only one "mapper" defined for the analysis of DECam data, the DecamInstcalMapper. As the name suggests, this operates using the instcal (or Instrument Calibrated) data products. These have the instrumental signature removed, the background still in the image, and some measure of astrometric and photometric zeropointing. It also requires the associated dqmask (mask file) and wtmap (weight map, stored as inverse variance) files for the same EXPNUM (visit), and requires that they are in the same subdirectory, which serves as the main input repository for processing. In particular, it requires:
The first step in this process is to build a registry that stores all the possible CCD images that can be analyzed. Since each DECam filename is unique, and they do not share a common string that associates the 3 files (instcal, dqmask, wtmap) together, we associate the 3 files by using their common EXPNUM in the header. These file names are stored in the registry. If there are not the requisite 3 files (instcal, dqmask, wtmap) having a common EXPNUM, this EXPNUM will not be entered in the registry.
The registry is built through the command
Things to note
ingestImagesDecam.py is located in your path at
. The name of the output directory is given above as decam/. It is required that there is a file _mapper in this directory with the content: lsst.obs.decam.decamMapper.DecamInstcalMapper
. The "link" mode creates symbolic links to the original instcals. Other options are to "move" and "copy"
. Finally, point the script at the instcals you intend to analyze. Note the script expects the dqmask/ and wtmap/ directories at the same level as instcal/. Also note that any
.fits.[fz/gz] files must be uncompressed beforehand
This will create the file
decam/registry.sqlite3 which contains the table raw with the schema:
These fields are defined here.
The next step is to process the images. This looks like:
Finally, to cluster the source measurements into lightcurves, run: