The Imager PIPELINE and REPORT tools

With ALMA or NOEMA, you may end up with an observational data set that contains several sources, each of them observed with a number of spectral windows of different spectral resolutions, covering many spectral lines.

The bookkeeping of such data sets can be intricate. To help the users to focus on the science, we have developped a PIPELINE (namely a suite of scripts named all-*.ima) that automates the whole imaging process in this case, and the REPORT command that can create printable (PDF format) files that summarize the data content and properties.

As a (strong) bonus, the PIPELINE builds up on all the knowledge and tools described in previous sections to provide an excellent solution given the data set, unless the source and/or data have very special properties.

This pipeline can be run through the PIPELINE command, or through a control panel launched by the PIPELINE /WIDGET command.

The pipeline control parameters are available in the ALL% global structure, and can be checked using the PIPELINE ? command.

The pipeline is intended to handle all spectral sub-bands for a single receiver tuning of the interferometer, with the data (possibly with several observing dates and configurations) in UVFITS or GILDAS UV format. Data from each sub-band and each source should be stored in a separate file.

The principle is to gather all UV tables in a sub-directory (named ./RAW/ by default), and to store the results of the various processing steps (Self Calibration, Spectral line extraction, Imaging) in different sub-directories (respectively named ./SELF/, ./TABLES/, ./MAPS/ by default). The REPORT command uses the same architecture, and generates PDF reports in the ./REPORT/ sub-directory.

The behaviour of the PIPELINE Imaging step depends on the pipeline mode, ALL%MODE, that can be controlled by the PIPELINE /MODE option. It also depends on whether a CATALOG is used or not, and on the ALL%RANGE[1:3] variable that controls the velocity span and resolution.

• CONTINUUM Mode:
  In this mode no data cube is produced. Only continuum images are created, with no attempt to remove any contaminating line emission.
• SURVEY Mode
  In this mode, there is also no spectral line emission filtering. Full data cubes are produced, at the spectral resolution specified by ALL%RANGE[3]. No continuum image is produced: continuum can be later extracted through the MAP_CONTINUUM command.
• SPLIT Mode
  
  • Without a spectral line catalog, the Imaging step will make no attempt to identify spectral lines, and thus no attempt to image each line separately. Instead, each “high spectral resolution” UV table will be imaged completely. An automatic estimate of the continuum level is made, and is imaged separately from the continuum-subtracted data. At the end of the process, the deconvolved “continuum” image is added back to the spectral image.“`Low spectral resolution” data only produce a continuum image. This mode is appropriate for e.g. low spectral resolution data, very wide lines, and/or nearly confusion limited data with many spectral lines.
  • With a line catalog, a user-specified velocity range is imaged separately around each line in the catalog that falls into the observed frequency coverage. Continuum and Line are separated, but no attempt is made to add back the continuum data at end.
• ALL Mode
  This mode is similar to the SPLIT mode, except that no attempt is made to produce continuum-free line data cubes. The line data cubes also include the continuum data. This mode is appropriate when the continuum emission has spatial variations of its spectral index.

Naming conventions apply to identify which data set covers which spectral line.