Reduction Problems

Forum for data reduction problems/solutions in the analysis.

Problem 1: Errors in processv2 at startup - I (Noel ) performed the cvs update earlier today, my IDL code is given for a basic analysis of one star on 090603.

Solution 1: The "hann" keyword is now replaced with "nohann" so that in general you shouldn't have to worry about this. The pavo_default_params_090311 is indeed in CVS. Try a "CVS update"... it works for me, but if this doesn't add the file, add to this solution and try "cvs checkout control/idl/pavo" from your root directory and CVS will definitely add the file.
Note to solution: The "cvs update" did not work alone, nor did the "cvs checkout control/idl/pavo"; however, copying the file to the current directory and removing the /hann worked ok. (Noel) Secondly - Note that you need to compile the code processv2 and ppspec in order for it to not crash.
Ah - further note: This problem only occurs if the analysis is done in a directory where the code is not... something we haven't done from Sydney. A note is made in the analysis manual... and someone will try and rectify this.

IDL> processv2, '090603', [1,1,0],00000, 01199, '090603_tauHer_1.txt',00111, 01152+indgen(5), /hann, root_dir='/raid/', foregroundtype=1,/plot
% Keyword HANN not allowed in call to: PROCESSV2
% Execution halted at: $MAIN$
IDL> processv2, '090603', [1,1,0],00000, 01199, '090603_tauHer_1.txt',00111, 01152+indgen(5), root_dir='/raid/', foregroundtype=1,/plot
Using parameter file: pavo_default_params_090311
% Compiled module: GETPAVOPARAMS.
% OPENR: Error opening file. Unit: 100
File: pavo_default_params_090311
No such file or directory
% Execution halted at: GETPAVOPARAMS 36
/chara/observe/control/idl/pavo/getpavoparams.pro
% PROCESSV2 95
/chara/observe/control/idl/pavo/processv2.pro
% $MAIN$

Problem 2: an unexpected crash in the code. - experienced on data taken on 20090804

IDL> processv2, '090804', [1,1,0],01512, 02626, '090804_1.txt',01512, 02022+indgen(5), root_dir='/raid/', foregroundtype=1, /plot
Using parameter file: pavo_default_params_090710
% LOADCT: Loading table Rainbow + white
using foreground type: 1
Number of distinct star data blocks: 4
% READFITS: Now reading 128 by 512 by 50 array
...
% READFITS: Now reading 128 by 512 by 50 array
% Stop encountered: PROCESSV2 255 /chara/observe/control/idl/pavo/processv2.pro
% Program caused arithmetic error: Floating divide by 0
% Program caused arithmetic error: Floating underflow
% Program caused arithmetic error: Floating illegal operand

Solution: Star used to calculate rotation and offset was not bright enough for the software. Changed star and it worked fine.

Problem 3: unexpected crash when running in Atlanta, not from MtW. Any ideas?
(The file in question DOES exist)
IDL> processv2, '090511', [1,1,1], 9548, 19691, '090511_data.txt', 12600, 12975, root_dir='/nfs/morgan/users/richardson/PAVO/', foregroundtype=0,/plot
Using parameter file: pavo_default_params_090311
% LOADCT: Loading table Rainbow + white
using foreground type: 0
% READFITS: Now reading 128 by 512 by 50 array
% READFITS: Now reading 128 by 512 by 50 array
offset: -0.937428
rotation: 1.60000
% SPAWN: Error managing child process.
No such file or directory
% HEADFITS: ERROR - Extension past EOF
error opening file/nfs/morgan/users/richardson/PAVO/090511/pavo09548.fits.Z
Error getting observing status
% Stop encountered: PROCESSV2 161 /nfs/morgan/users/richardson/PAVO/control/idl/pavo/processv2.pro

Solution: (maybe...) This looks like a problem with the '.Z' extension. You can uncompress the fits files and run processv2 with the "/z" option.
Check that uncompress works as "uncompress" or with "gzip -cd". The program readfits from astrolib uses "gzip -cd" to uncompress the files.

Two things to note... perhaps it is an issue with compressed files BUT a) the code had already run through to the point of calculating the offset and rotation. The weird thing is that it started at the point where it goes in depth with the plots and analysis of the fringes. b) [and not a real issue for Mike, but more of a note for GSU people] our computers does not have an easy way to uncompress .Z files. I will speak with the system admin on Monday about that.

"uncompress" is now installed on Atlanta computers, and it works locally!

Solution notes: The offset and rotation are not necessarily calculated.... this initial part of the analysis is saved in processv2.pro so that if you repeat the analysis, lots of time-consuming stuff doesn't have to be re-done. It was implemented by Dan Huber to make /fast go faster. But indeed... the fact that "% READFITS" worked twice seems to indicate that the compressed data are a problem in headfits and not readfits. HEADFITS uses fxposit.pro, which in turn uses "uncompress". So e.g. a unix "alias uncompress gunzip" should work, if you can't get uncompress. Or just modify gxposit.pro in astrolib, replacing the word "uncompress" with "gunzip".