Home

Welcome to the SoFiA Image Pipeline wiki!

Troubleshooting

The -m option doesn't produce a combined image.

Try typing which convert. If you have karma/kvis installed (common for viewing HI radio cubes), then it may return:
/usr/local/karma/bin/convert

If this is the case, you need to find where ImageMagick is installed. If you installed with Homebrew the path may be something like:
/usr/local/Cellar/imagemagick/7.1.0-13/bin/convert

To create the combined image, add the correct path after the -m option on the command line.

The mask is not being plotted on the Position-Velocity Diagram

Try installing the latest version of pvextractor from github (as of this note 2 Aug 2022, 0.4 is needed but not available through pypi):
python3 -m pip install git+https://github.com/radio-astro-tools/pvextractor

Minimum requirements for plotting HI contours on an optical image

Source catalog with the following properties:

Naming scheme: <catalogname>_cat.txt or <catalogname>_cat.xml

Catalog contents:

• A commented header similar enough to SoFiA in that the header starts on line 18. (Need to work on making this more flexible!). A sample can be found here.
• The following column names: id, name, ra, dec, x, y, x_min, y_min, x_max, y_max, rms, and <one of freq,v_opt,v_rad, or v_app>.
  • ra, dec are in decimal degrees (float).
  • x, y is the source center in pixels (float).
  • x(y)_min(max) are the min/max in pixels of the source (int), used to determine if a larger figure should be retrieved for an individ source.
  • freq, v_opt, v_rad, or v_app are the systemic frequency/velocity of the galaxy.

Moment0 FITS file for each source with the following properties:

Naming scheme: <catalogname>_cubelets/<catalogname>_<id>_mom0.fits

FITS header contents (these may be non-standard for mom0 maps, but are required for estimating meaningful column density contours):

• SPECSYS or VELREF to refer to the spectral coord system of the original data.
• CTYPE3 to determine the spectral units of the original data.
  • If a cubelet is provided the coordinate system is determined from that so ctype3 isn't needed in a mom0 map. Alternatively, one could determine this from the catalog column names (e.g the existence of 'freq', 'v_opt', 'v_rad', or 'v_app' to indicate the center of the galaxy. (Requires some development).

Command line arguments:

• -cw <channel width in native data units>. This can be the channel width of the original data, or simply the km/s or Hz over which the HI column density should be calculated for a given signal-to-noise (which can be specified with -snr). Noise value is taken from the rms column of the catalog.

Known citations to the SoFiA Image Pipeline software

• A serendipitous discovery of H I-rich galaxy groups with MeerKAT, Glowacki M., Albrow L., Reynolds T., Elson E., Mahony E.~K., Allison J.~R., 2024, MNRAS, 529, 3469. doi:10.1093/mnras/stae684