gabraganca / s4 Goto Github PK

The identification are still all messed up. Very strange...

We need to get rid of IDL in order to have a full open source approach.

Since it is is a plot resource, it would probably be best to not have any console interactivity. Maybe this interactivity should be on the user script.

Or even better: it could have an option to bring a Qt interativity.

This package has so many code that is not used. Some of it can be discarded.

Time to go hunting!

The setup.py is not finding ifort on my pc at ON, and so is not compiling Synspec and Rotin3.

Travis is not building the environment prior to running the tests. It is throwing the following error:

The command "wget $securityBaseURL/c/cups/libcups2_1.7.2-0ubuntu1.7_amd64.deb -O libcups2.deb" failed and exited with 8 during .

I tried to bind Travis to S4 but I could not get it working.

After install, when I run the program calling "sagui" in terminal, the program returns this error message:

ERROR: IOError: [Errno 2] No such file or directory: '/home/julio/.s4/synthesis/synplot/fort.11' [numpy.lib.npyio]
Traceback (most recent call last):
  File "/usr/local/bin/sagui", line 553, in <module>
    main()
  File "/usr/local/bin/sagui", line 546, in main
    mw = MainWindow()
  File "/usr/local/bin/sagui", line 44, in __init__
    self.wdg = Widget()
  File "/usr/local/bin/sagui", line 123, in __init__
    self.set_default()
  File "/usr/local/bin/sagui", line 432, in set_default
    self.synplot()
  File "/usr/local/bin/sagui", line 159, in synplot
    self.syn.run()
  File "/usr/local/lib/python2.7/dist-packages/s4/synthesis/synplotwrapper.py", line 99, in run
    self.spectrum = np.loadtxt(self.path + 'fort.11')
  File "/usr/lib/python2.7/dist-packages/numpy/lib/npyio.py", line 715, in loadtxt
    fh = iter(open(fname, 'U'))
IOError: [Errno 2] No such file or directory: '/home/julio/.s4/synthesis/synplot/fort.11'

The program tries to call .s4/synthesis/synplot/fort.11 but the file is missing.

I installed the S4 again and the first run of sagui returns:

[julio@bd2093ea ~]$ sagui
/usr/lib64/python2.7/site-packages/gtk-2.0/gtk/__init__.py:122: RuntimeWarning: PyOS_InputHook is not available for interactive use of PyGTK
  set_interactive(1)
ERROR: IOError: Calculated spectrum is not available. Check if syn(spec|plot) ran correctly. [s4.synthesis.synplotwrapper]
Traceback (most recent call last):
  File "/usr/bin/sagui", line 513, in <module>
    main()
  File "/usr/bin/sagui", line 508, in main
    mw = MainWindow()
  File "/usr/bin/sagui", line 41, in __init__
    self.wdg = Widget()
  File "/usr/bin/sagui", line 139, in __init__
    self.synplot()
  File "/usr/bin/sagui", line 172, in synplot
    self.syn.run()
  File "/usr/lib/python2.7/site-packages/s4/synthesis/synplotwrapper.py", line 97, in run
    'syn(spec|plot) ran correctly.')
IOError: Calculated spectrum is not available. Check if syn(spec|plot) ran correctly.

IDL and Intel Fortran are installed and working.

Try adding a boxcar smooting to the spectra.line_position function to improve detection of the line position on spectra of los S/N.

According to this document, SYNSPEC version is 51, and we are running v49. Should we update?

The synthetic spectrum is plotted underneath the observed spectrum. This makes very difficult to check if the fit is good or not.

I guess would it be better to do the opposite: synthetic spectra over the observed one.

Now that GDL can run Synplot (#31), we can start writing some tests.

The latest version of GDL is able to run Synplot. So, this should be made the default option instead of IDL.

I should probably write a method to load FITS spectrum as a ndarray.

The flux can be loaded easily with astropy.io.fits. For the wavelength I will have to calculate the values using the CRPIX1, CRVAL1 and CDELT1 entries on the header. The header can also be easily loaded also with the astropy.io.fits.

Travis CI runs their VMs on Ubuntu 14.04 images, which is an issue for us because it does not contain GDL on its repository, forcing us to download and install all dependencies one by one.

The repository where the dependency is located eventually get updated which breaks the Travis buildings (see #39 ).

One solution would be running an updated version of Ubuntu inside a Docker container, as it is suggested in travis-ci/travis-ci#5821

This docs can shed some light on how to do it.

The S4 can only open a spectrum if the spectrum is on the folder or if the entire path is given.

Please, add os.path.realpath to obtain the entire path of the spectrum before loading it.

@juliotux asked me to write a function to calculate the equivalent width of a line.

Actually the function is already implemented (s4.spectools.get_equiv_width()) but lacks a correction in the case that the user is using a scale factor.

I probably should add this to SAGUI as well.

Make that the line identification argument (line_id) to be an argumento fo the plot function and not from the object.

When using the plot method of the s4.synthesys.Synplot class on the ipython notebook, it is generating two plots.

I guess that the JSON class should be better at a IO module.