Commit summary from repository starlink
------------------------------------
(Tim Jenness) surf: Use GSL for fitting with SKYDIP2
100.0% applications/surf/src/
(Tim Jenness) surf: Add alternative skydip fitting routine based on GSL
97.0% applications/surf/sculib/
2.9% applications/surf/
(Tim Jenness) surf: Flag observations that have suspect input data
100.0% applications/surf/src/
(Tim Jenness) surf: Warn if the tel or amb temperatures can not be determined
100.0% applications/surf/src/
(Tim Jenness) surf: Fix FITS header to use for telescope temperature for SCUBA-2
100.0% applications/surf/src/
(Tim Jenness) surf: Set default variance to reasonable value for skydips
100.0% applications/surf/src/
(Ed Chapin) smurf: fix chunking error when down-sampling
100.0% applications/smurf/libsmf/
(Ed Chapin) smurf: memory leak
100.0% applications/smurf/libsmf/
Commits from repository starlink
-----------------------------
commit 41c6fb637eb8f7924bc79048f9b1e7d1f494f47d
Author: Tim Jenness <[log in to unmask]>
Date: Thu Oct 27 15:18:50 2011 -1000
surf: Use GSL for fitting with SKYDIP2
This does not change the SKYDIP command.
applications/surf/src/surf_skydip2.f | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
commit a9883f3e55913cf94d20f423d8277d50ab80a594
Author: Tim Jenness <[log in to unmask]>
Date: Thu Oct 27 14:55:29 2011 -1000
surf: Add alternative skydip fitting routine based on GSL
This can replace the SCULIB_FIT_SKYDIP fortran code. SURF originally
used a NAG routine but this was replaced around 1995 when NAG was
pulled from the Starlink software. The replacement algorithm does
mostly work okay except that it will not fix fixed parameters. The
GSL replacement will leave fixed parameters fixed but still has
a problem with free parameters going out of range because the
GSL routine also does not have box constraints but uses a
penalty function to discourage bad parameter choices.
Note that there is a LM fitting routine in SURF which does
not seem to be used. The original NAG routine supported
non-linear constraints as well as box constraints.
applications/surf/component.xml | 2 +-
applications/surf/configure.ac | 4 +-
applications/surf/sculib/Makefile.am | 4 +-
applications/surf/sculib/sculib_fit_skydip_gsl.c | 785 ++++++++++++++++++++
applications/surf/surf_link_adam | 1 +
5 files changed, 792 insertions(+), 4 deletions(-)
commit fa25a547b71d2c6647e0508ecc98f29efbe0a201
Author: Tim Jenness <[log in to unmask]>
Date: Wed Oct 26 16:43:41 2011 -1000
surf: Flag observations that have suspect input data
If the sky temperatures are negative this is bad.
If the sky is brighter at low elevations than high elevations
this is bad.
applications/surf/src/surf_skydip2.f | 84 ++++++++++++++++++++++++++++++++
1 files changed, 84 insertions(+), 0 deletions(-)
commit e7b4b414d404a452fb972d75800b23513a17491f
Author: Tim Jenness <[log in to unmask]>
Date: Wed Oct 26 15:39:44 2011 -1000
surf: Warn if the tel or amb temperatures can not be determined
applications/surf/src/surf_skydip2.f | 28 +++++++++++++++++++++++-----
1 files changed, 23 insertions(+), 5 deletions(-)
commit 2ab17b70e35582c6ebc99df49724cf7c692631e5
Author: Tim Jenness <[log in to unmask]>
Date: Wed Oct 26 14:25:40 2011 -1000
surf: Fix FITS header to use for telescope temperature for SCUBA-2
applications/surf/src/surf_skydip2.f | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
commit 6ce1f49057c596ea6e7a46141c9bdaf29f031edc
Author: Tim Jenness <[log in to unmask]>
Date: Wed Oct 26 14:22:38 2011 -1000
surf: Set default variance to reasonable value for skydips
If there is no variance data associated with the skydip input data
then use a default of 1.0 K rather than a default of 0.0 K.
applications/surf/src/surf_skydip2.f | 6 ++++--
1 files changed, 4 insertions(+), 2 deletions(-)
commit b4d5a667d2f8f6406e1a8eb5315070744826e88e
Author: Ed Chapin <[log in to unmask]>
Date: Thu Oct 27 14:24:21 2011 -0700
smurf: fix chunking error when down-sampling
The second call to smf_grp_related was missing the keymap so it
didn't know the new sample rate, and would get the wrong answer
when trying a new maximum length (in time) per chunk.
Also added in some optional debugging memory usage messages.
applications/smurf/libsmf/smf_iteratemap.c | 74 +++++++++++++++++++++++--
1 files changed, 68 insertions(+), 6 deletions(-)
commit 27a806507369964e279a2cc019cd0e7fe04f0dfa
Author: Ed Chapin <[log in to unmask]>
Date: Thu Oct 27 12:03:23 2011 -0700
smurf: memory leak
applications/smurf/libsmf/smf_fft_data.c | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
|