Accessing measurement set data

The CASA table tool allows access to all data contained in measurement set format. It is however a little cumbersome to use often in scripts and very cumbersome and verbose on the command line.

For this reason I have sketched out a simple, python-orientated, wrapper for data access in casata. This wrapper is contained in the module tools.data. This shows the basics of how it should work – patches to make it better would be very welcome.

The key features that I would like to have in this wrapper are:

• Nice, short names for commonly used parameters. So “ANTENNA1” is “a1”, etc

• Flexibility to get multiple columns at once. So by asking for:

  cols=["TIME", "DATA"]
  

  one gets (unsurprisingly) both the time and the data....

• Automatic interpolation of subsidiary tables onto the time base of the main table. So, if you ask for:

  cols=["DATA", "POINTING_OFFSET"]
  

  then the pointing offsets pulled out of “pointing” sub-table and interpolated to the time base of the main table

• Automatic join of common parameters. So if you ask for:

  spw=1
  

  You get spectral window 1

The basics are there now, but to make it really useful I need patches from people using it!

Examples

Here are some examples with uid___A002_Xa048f_X1.ms.

• Get timestamps and visibility data on baseline 0-1 for spw 1:

  >>> msin="/home/bnikolic/data/almagen/uid___A002_Xa048f_X1.ms"
  >>> time, visdata=data.vis(msin, ["TIME", "DATA", ], spw=1, a1=0, a2=1)
  >>> visdata
  array([[[ 0.00281871-0.01850254j,  0.00162236-0.01870679j,
            0.00240437-0.01870679j, ...,  0.00316449-0.01151703j,
            0.00369846-0.01152579j,  0.00382977-0.01135217j]]])
  

• Get and plot pointing offsets:

  >>> time, point1, point2 =data.vis(msin, ["TIME", "POINTING_OFFSET", ], spw=1, a1=0, a2=2)
  >>> pylab.plot(time, point1[0])
  [<matplotlib.lines.Line2D instance at 0x3efdc68>]
  >>> pylab.plot(time, point2[0])
  [<matplotlib.lines.Line2D instance at 0x3ad4098>]
  >>> pylab.savefig("uid___A002_Xa048f_X1_point.png")
  

  

• Get specific sub-scans (!)

  >>> time, visdata=data.vis(msin, ["TIME", "DATA", ], spw=1, a1=0, a2=2, subscan=1)
  >>> pylab.plot(time, numpy.abs(visdata[0,0,:]))
  [<matplotlib.lines.Line2D instance at 0x79d0ea8>]
  >>> time, visdata=data.vis(msin, ["TIME", "DATA", ], spw=1, a1=0, a2=2, subscan=2)
  >>> pylab.plot(time, numpy.abs(visdata[0,0,:]))
  [<matplotlib.lines.Line2D instance at 0x79d0f38>]
  >>> time, visdata=data.vis(msin, ["TIME", "DATA", ], spw=1, a1=0, a2=2, subscan=3)
  >>> pylab.plot(time, numpy.abs(visdata[0,0,:]))
  [<matplotlib.lines.Line2D instance at 0x3690c20>]
  >>> time, visdata=data.vis(msin, ["TIME", "DATA", ], spw=1, a1=0, a2=2, subscan=4)
  >>> pylab.plot(time, numpy.abs(visdata[0,0,:]))
  [<matplotlib.lines.Line2D instance at 0x44a6dd0>]
  >>> time, visdata=data.vis(msin, ["TIME", "DATA", ], spw=1, a1=0, a2=2, subscan=5)
  >>> pylab.plot(time, numpy.abs(visdata[0,0,:]))
  [<matplotlib.lines.Line2D instance at 0x3d6bc68>]
  >>> pylab.savefig("uid___A002_Xa048f_X1_subs.png")
  

  

Here are some new examples on other data set:

• Obtaining the closure phase:

  >>> data.closurePh(msin, [0,1,3], spw=7, field=0)
  array([[[ 52.87974425,  54.13733906,  54.97565168,  52.62370937,
            52.64148397,  54.30899442,  53.93281406,  52.61754189,
            52.61760492,  56.32891778,  55.98011178,  54.28151425,
            (... rest of data here)