From cohen at lpta.in2p3.fr Thu Jun 5 07:42:36 2008 From: cohen at lpta.in2p3.fr (Cohen-Tanugi Johann) Date: Thu, 05 Jun 2008 13:42:36 +0200 Subject: [AstroPy] [SciPy-user] FITS images with header-supplied axes? Message-ID: <4847D12C.10107@lpta.in2p3.fr> hi there, I just registered to this mailing list, and I was amused to see that one of the last threads actually match a question I have asked 2 days ago to the help at stsci regarding numdisplay and WCS. I don't know if anything has come out of the discussion here yet, but I have the feeling that the following basic prototyped tool is right around the corner : - open a fits file or else get a numpy array - get a WCS object from the fits file or else creating it by hand - display in matplotlib the resulting combination (numpy array, wcs astronomical axes) astLib seems to do just that with the astPlot.FinderChart macro though I do not understand the use yet astLib swigifies WCSTools, while pywcs takes the better approach IMHO of wrapping wcslib directly. numdisplay sends to ds9 any numpy array, but that direciton does not seem very promising to me.... But again, as I just said the prototyped use above is essentially available, though not in a straightforward practical way. It would be nice to close the loop so that plotting astronomical data with matplotlib becomes an easy task. best, Johann From lee.j.joon at gmail.com Thu Jun 19 16:33:55 2008 From: lee.j.joon at gmail.com (Jae-Joon Lee) Date: Thu, 19 Jun 2008 16:33:55 -0400 Subject: [AstroPy] pywcsgrid : python module to draw and label a curvilinear coordinate grid. Message-ID: <6e8d907b0806191333m4dc9825ctf7e935ac9d8448b1@mail.gmail.com> Hi all, This is to introduce a python module "pywcsgrid", with which you can draw and label a curvilinear coordinate grid (using fits header information). It gives you grid lines as numpy arrays (and also tick locations and labels), therefore you may use any plotting package you want (I personally use matplotlib). This module is developed for my personal use (so the code quality is low) and should be regarded as beta version. But I hope others find this useful. I have put some more information in the following page with some example plots. http://docs.google.com/Doc?id=dgpj7zzh_158f9rc29fv The module is actually a wrapper around PGSBOX library in WCSLIB. While the original PGSBOX uses the pgplot library for plotting, pywcsgrid links PGSBOX with a custom version of pgplot. For example, when PGSBOX calls pgline for grid drawing, pgline in the custom pgplot is called which stores line positions as numpy array. The module depends on "pywcs". Any comments or suggestions will be appreciated. Regards, -JJ From lee.j.joon at gmail.com Thu Jun 19 16:33:55 2008 From: lee.j.joon at gmail.com (Jae-Joon Lee) Date: Thu, 19 Jun 2008 16:33:55 -0400 Subject: [AstroPy] pywcsgrid : python module to draw and label a curvilinear coordinate grid. Message-ID: <6e8d907b0806191333m4dc9825ctf7e935ac9d8448b1@mail.gmail.com> Hi all, This is to introduce a python module "pywcsgrid", with which you can draw and label a curvilinear coordinate grid (using fits header information). It gives you grid lines as numpy arrays (and also tick locations and labels), therefore you may use any plotting package you want (I personally use matplotlib). This module is developed for my personal use (so the code quality is low) and should be regarded as beta version. But I hope others find this useful. I have put some more information in the following page with some example plots. http://docs.google.com/Doc?id=dgpj7zzh_158f9rc29fv The module is actually a wrapper around PGSBOX library in WCSLIB. While the original PGSBOX uses the pgplot library for plotting, pywcsgrid links PGSBOX with a custom version of pgplot. For example, when PGSBOX calls pgline for grid drawing, pgline in the custom pgplot is called which stores line positions as numpy array. The module depends on "pywcs". Any comments or suggestions will be appreciated. Regards, -JJ From lee.j.joon at gmail.com Fri Jun 20 15:52:14 2008 From: lee.j.joon at gmail.com (Jae-Joon Lee) Date: Fri, 20 Jun 2008 15:52:14 -0400 Subject: [AstroPy] pywcsgrid : python module to draw and label a curvilinear coordinate grid. In-Reply-To: <6e8d907b0806191333m4dc9825ctf7e935ac9d8448b1@mail.gmail.com> References: <6e8d907b0806191333m4dc9825ctf7e935ac9d8448b1@mail.gmail.com> Message-ID: <6e8d907b0806201252y707db365xcd1048b9a8a39249@mail.gmail.com> I noticed that there are some license issues and disabled the download. pywcsgrid includes "pgnumb.f" from the pgplot library which does not allow redistribution. License of WCSLIB seems a bit confusing. It says LGPL in some places and GPL in some others. I'll try to resolve these things and any advice. will be welcomed. Regards, -JJ On Thu, Jun 19, 2008 at 4:33 PM, Jae-Joon Lee wrote: > Hi all, > > This is to introduce a python module "pywcsgrid", with which you can > draw and label a curvilinear coordinate grid (using fits header > information). It gives you grid lines as numpy arrays (and also tick > locations and labels), therefore you may use any plotting package you > want (I personally use matplotlib). > > This module is developed for my personal use (so the code quality is > low) and should be regarded as beta version. But I hope others find > this useful. > > I have put some more information in the following page with some example plots. > > http://docs.google.com/Doc?id=dgpj7zzh_158f9rc29fv > > The module is actually a wrapper around PGSBOX library in WCSLIB. > While the original PGSBOX uses the pgplot library for plotting, > pywcsgrid links PGSBOX with a custom version of pgplot. For example, > when PGSBOX calls pgline for grid drawing, pgline in the custom pgplot > is called which stores line positions as numpy array. > The module depends on "pywcs". > > Any comments or suggestions will be appreciated. > Regards, > > -JJ > From jh at physics.ucf.edu Wed Jun 25 11:36:43 2008 From: jh at physics.ucf.edu (Joe Harrington) Date: Wed, 25 Jun 2008 11:36:43 -0400 Subject: [AstroPy] [cesare.tirabassi@gmail.com: [Bug 133784] Re: saods9 is way out of date wrt upstream (4.0b7 vs 4.13)] Message-ID: Dear astrofolk, The ds9 package from SAO is perhaps the most widely used viewer in the biz, and it's pretty common for people to use it with numpy or IRAF. There's a poor substitute for it in IDL (called atv). It has been available in Debian-derived Linux distros for some years, but is outdated. I flagged this fact in a bug report last year, only to see the (idle) Debian maintainer formally abandon the package instead of updating it. Having a package available in the Debian repo means that normal Linux users can install it with a few mouse clicks in a GUI. It means that the install is fully reversible and that updates are automatic. It eliminates the need to manage a /usr/local, which can be difficult for non-system-savvy users. However, abandoned packages do not long remain in a distro, and old packages are sometimes worse than none at all. The message below is the first move for Debian to formally remove the package from its repos. The reasons cited in the note below are certainly accurate: the package is hard to build and does not use the standard tools. They are also why we need to find a Debian maintainer: either one of us steps forward and solves the problem of making a useful package out of it, or we each do it individually for ourselves. I hope that someone will step forward to take on this important task, or that SAO will step up and release OS packages like .debs, or at least release sources based on standard tools and that can be reliably built by users. While I am not familiar with making .deb files, my understanding is that it is easier than building ds9 from sources. Having taken on the numpy docs, I cannot do this myself as well. Will anyone step forward to take on packaging ds9? --jh-- Prof. Joseph Harrington Department of Physics MAP 414 4000 Central Florida Blvd. University of Central Florida Orlando, FL 32816-2385 jh at physics.ucf.edu ------- Start of forwarded message ------- MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" From: Cesare Tirabassi To: jh at alum.mit.edu Reply-To: Bug 133784 <133784 at bugs.launchpad.net> Date: Wed, 25 Jun 2008 12:01:18 -0000 Subject: [Bug 133784] Re: saods9 is way out of date wrt upstream (4.0b7 vs 4.13) We have version 5.2 now from upstream. This is a nightmare, its a 25 MB tarball which contains many libraries already in the archive (amongst them: blt, tcl, tk, tkimg, tktable, zlib), some of these in binary format, at least another modified from source (this was replaced in Debian by a previous fortran version), its own build system which is a (worse) homemade version of a complete autotools set. The package has been orphaned in Debian since February 2008, and I have serious doubts that, due to its nature, it will be adopted in the near future. My preference would be to wait until Intrepid alpha 6 release (currently scheduled for September 18th) and if there is no update from Debian just remove the obsolete source and binaries from the archive. ** Changed in: saods9 (Ubuntu) Assignee: Cesare Tirabassi (norsetto) => (unassigned) Status: In Progress => Confirmed - -- saods9 is way out of date wrt upstream (4.0b7 vs 4.13) https://bugs.launchpad.net/bugs/133784 You received this bug notification because you are a direct subscriber of the bug. ------- End of forwarded message -------