Python Links

 		Python, IPython, matplotlib links provided on the left.

Startup 

bash-2.05b$ casapy
_______________________________________________
Available tools:

   cb (calibrater)       im (imager)
   ms (MS)               tb (table)
   mp (MS plot)          tp (table plot)

_______________________________________________
Available tasks:

  imager: clean, feather, invert, mosaic, (wproj)

_______________________________________________

type 'exp tool' or 'exp tool.function' for help


CASA [1]:
Some key interface differences between casa (python) and aips++ (glish) for those familiar with AIPS++.

  • No tool constructors (more data-centric).
  • No ':=' or ';' needed.
  • On-line help (exp)
  • CNTRL-C
  • Memory usage
    • casa 27M footprint (c.f. aips++ 103M starting)
    • pre-loading imager, calibrater, ms, and imager tasks (adds 1M to startup)
  • Startup time (casa: 1.9s (ballista), aips++: 10.1s (ballista))
  • Performance (approx. 15% faster on *tested* scripts)
  • Python is 0-based. Awkward but uniform. MS selection solves much of this.
  • Error handling (loquacious versus laconic) 
    glish

: Caught an exception! Event type=run exception=RecordInterface: field
         FLAG_LEVEL is unknown
         File:   servers.g, Line 1009
         Stack:    .(), ms.g line 419
         .()

python

Exception Reported: RecordInterface: field FLAG_LEVEL is unknown

For glish-users

  • Detailed functionality has been migrated over and is usable nearly identically as before. 
    Python imaging
im.open('ngc5921_src.split.ms')
im.setdata(mode='channel',nchan=[46],start=[1],step=[1],fieldid=[0])
im.setimage(nx=256,ny=256,cellx=quantity(15.,'arcsec'),celly=quantity(15.,'arcsec'),mode='channel',nchan=46,fieldid=0)
im.weight(type='briggs',rmode='norm',robust=0.5)
im.clean(algorithm='hogbom',niter=6000,gain=0.1,threshold=quantity(8.,'mJy'),model=['ngc5921_py.model'],image=['ngc5921_py.image'],residual=['ngc5921_py.residual'],mask=[''])

Glish imaging
im:=imager('ngc5921_src.split.ms')
im.setdata(mode='channel',nchan=46,start=1,step=1,fieldid=1)                    
im.setimage(nx=256,ny=256,cellx='15arcsec',celly='15arcsec',mode='channel',nchan=46,fieldid=1)
im.weight(type='briggs',rmode='norm',robust=0.5)
im.clean(algorithm='hogbom',niter=6000,gain=0.1,threshold='8mJy',model='gtest2.model',image='gtest2.image',residual='gtest2.residual',mask='')
  • Glish records -> Python dictionaries
    • dict.keys()
    • dict.sort()
    • dict.has_key('keyname')
    • dict.values()
    • dict['keyname']
    • del dict['keyname'] 
    

  •  range specifications

    CASA [11]: range(4,7)
  Out[11]: [4, 5, 6]

CASA [12]: range(1,3)
  Out[12]: [1, 2]

CASA [13]: range(4,7)+range(55,60)
  Out[13]: [4, 5, 6, 55, 56, 57, 58, 59]

    


  •  quantities (for now)

    CASA [4]: quan=quantity(5.,'km/s')

CASA [5]: quan.value
  Out[5]: 5.0

CASA [6]: quan.units
  Out[6]: 'km/s'


    

Functionality Update:


      

    •  Previously, parts of imager completed; calibrater stubbed

    •  Now:

      imager      30 methods,  8 untested, 6 not implemented, 1 abort, 0 fail
calibrater  16 methods,  1 untested, 3 not implemented, 0 abort, 1 fail
ms          29 methods,  6 untested; 3 not implemented, 0 abort, 3 fail
tables      stubbed
tableplot   stubbed
msplot      stubbed
---
autoflag    started
measures    started

      

    •  help system enabled with argument information


      In [1]: exp im.setdata
    Set the data parameters selection for subsequent processing :
      mode      = velocity
      nchan     = [ 1 ]
      start     = [ 0 ]
      step      = [ 1 ]
      mstart    = { value=0.0, units=km/s }
      mstep     = { value=0.0, units=km/s }
      spwid     = [ 0 ]
      fieldid   = [ 0 ]
      msselect
      async     = false
    ----------------------------------------

      
      


    

    
Example Reduction:


      

    •  NGC 5921

    •  All regression tests being ported

    


    
Tasks:

      

    •  Focus area for Focus Group!

    


    
Why Matplotlib?

      

    •  interface (zoom, cursor, etc)

    •  many point rendering

    •  publication quality (many formats bmp, eps, jpg, png, ps, svg)

    •  use TeX labelling

    


    
To Do:

      

    •  complete functionality for initial release

    •  improved logging

    •  matplotlib

    •  documentation

    •  ----

    •  interface working group in spring (March?)

    

    

    

Please send any comments or questions about CASA or AIPS++
to aips2-requests@nrao.edu
    

Copyright © 1995-2000,2001,2002,2003,2004 Associated Universities Inc.,
Washington, D.C.

    

This code is available under the terms of the GNU General Public Lincense




    




Modified on 
Wednesday, 02-Sep-2009 08:40:43 MDT