From: Glen Petitpas 
To: Jun-Hui Zhao 
Subject: Re: Pipeline errors


Hi

imsize fixed the invert problem.

Thanks!
Glen

Jun-Hui Zhao wrote:
> Hi Glen,
>      I assumed that this problem has been fixed after set
> 'imsize=???x???'.
>      Please let me know if you have any further problems with
> your pipeline script.
>      Jun-Hui
> On Wed, 28 Jul 2010, Glen Petitpas wrote:
> 
>> Hi
>>
>> I tried four iterations of "invert" these fixes, and none of them seemed
>> to work for me. I was testing on 100728_13:24:22
>>
>> Option 1: I tried adding the options=flagchan0 to the smalod command.
>> With just that I got:
>> ### Fatal Error [invert]:  Unrecognised primary beam type -28
>>
>> Option 2: I removed the options=flagchan0 from smalod and added
>> manually unflaging the window using uvflag.
>> With that I got:
>> ### Fatal Error [invert]:  Invalid preamble "time/baseline".
>>
>> Option 3: I tried both the flagchan0 option AND the uvflag option
>> With that I got:
>> ### Fatal Error [invert]:  Unrecognised primary beam type -28
>>
>> Option 4: Ram suggested trying neither, and I got
>> ### Fatal Error [invert]:  Invalid preamble "time/baseline".
>>
>> Cheers
>> Glen
>>
>> Jun-Hui Zhao wrote:
>>> Hi Glen:
>>>     Both bugs are general to all the 4GHz mode data.
>>> The bug in the uvlist can be simply fixed by increasing the limit of
>>> spectral windows. To get rid of the second problem in mapping (invert),
>>> either set options=flagch0 in smalod or using uvflag to reset
>>> the flag states for channel 1 (window 1):
>>> uvflag vis=100727_11:33:23_rxl.lsb select='window(1)' flagval=unflag
>>>     There was a bug in options=flagch0 of smalod.
>>>     Both bugs have been fixed and a new beta version (1.1.0) has been
>>> released. Please ask Ram to get the updated miriad from
>>> http://www.cfa.harvard.edu/sma/miriad/download/
>>> and install it. Then please give a try again.
>>>     Thanks.
>>>     Jun-Hui
>>>
>>>
>>> On Tue, 27 Jul 2010, Glen Petitpas wrote:
>>>
>>>> Hi
>>>>
>>>> I tried my quickLook pipeline today on Rams test version and running
>>>> on rtdchilo2. It went pretty well, but there were still some issues.
>>>>
>>>> "smalod" seems to have worked without errors.
>>>>
>>>> "uvlist" Fatal Errors reporting "Too many windows"
>>>>
>>>>
>>>> This command is used for the autoscaling of the axes on the
>>>> summary plots. Not essential, but useful and will need to be fixed
>>>> eventually since I use "uvlist" for many reasons other than my pipeline.
>>>>
>>>> I also image the primary quasar to get an estimate of the beam.  To do
>>>> this first I create a continuum uv file of the main quasar using:
>>>>
>>>>  > uvaver vis=100727_11:33:23_rxl.lsb.tsys select=source"(0423-013)"
>>>> out=tmp_lsbcal0.uv line=channel,1,1,1,1
>>>> (where 100727_11:33:23_rxl.lsb.tsys is just the tsys corrected data file)
>>>>
>>>> Then I invert this using:
>>>>  > invert vis=tmp_lsbcal0.uv map=tmpll.map beam=tmpll.beam
>>>> options=systemp cell=0.1
>>>>
>>>> and I get
>>>> Invert: version 1.0 19-apr-09
>>>> Reading the visibility data ...
>>>> ### Fatal Error [invert]:  Invalid preamble "time/baseline".
>>>>
>>>> This error is a biggie.
>>>>
>>>> As far as I can tell, these are currently the only errors I get. I cannot
do
>>>> the higher level debugging until these are fixed.
>>>>
>>>> It seems we are very close to having the pipeline working again the
>>>> way it used to.
>>>>
>>>> Cheers
>>>> Glen