Skip to content

eholmbeck/smhr-rpa

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Build Status

Spectroscopy Made Harder

Gotta pay back that tech debt.

Authors

  • Andrew R. Casey (Monash)
  • Alex Ji (Carnegie Observatories)

VERSION NOTE:

  • The branch refactor-scatterplot has an updated and improved GUI (as of Jan 2020). These have not been merged into master yet but should be soon.
  • The master branch is currently frozen to a version from about July 2019.
  • v0.22 (formerly branch better-errors) is a frozen version that is the result of a big update on May 30, 2019. It is considered a stable version.
  • v0.2 is a frozen development version, v0.21 is a slightly more recently frozen version.
  • v0.1 is the current stable version. Things are working and it is being used for papers.

If you are new to SMHR, you should use the branch refactor-scatterplot. Note v0.1 and v0.2 files are not compatible, but there is a script to convert old save files into new save files. There is not a way to convert files from the old SMH to new SMHR.

Note about this version

  • This is a fork of the original SMHr code that has updates and additions for use by the R-Process Alliance.
  • Direct questions to Erika Holmbeck (Carnegie Observatories) or Alex Ji (University of Chicago).

Installation

This is one way that I (Alex) got things running from a fresh mac install and I'm putting it here as a record of some things I had to do. Default anaconda does not officially support pyside, but I think conda-forge does. We may switch to that, or a different version of PySide, in the future. For now, this works...

  • Download full anaconda with python 2.7. If you run into trouble later, specifically install an old anaconda version from the anaconda archive around October 2016.
  • conda install pyside: this will install pyside v1.1.1
  • install qt4:
brew tap cartr/qt4
brew tap-pin cartr/qt4
brew install qt@4
  • conda install matplotlib=1.5.1 (fixing this is very painful because I need to update to qt5; I have started doing this but it will take a long time)
  • conda install numpy=1.11.3 (this is an issue with pyside)
  • conda install scipy=0.19.0 (this is an issue with newer versions of scipy, not sure yet why but it segfaults in interpolate.griddata)
  • conda install qt=4.8.7 (you may have to uninstall and downgrade some things for this to work; it should be safe to upgrade those later) (Recent versions of anaconda have removed this version of QT. We have not yet found a solution but are working on it...it appears to mess up the Open File Dialogs, and we will try to come up with a workaround.)
  • Clone smhr
  • Go into the smhr directory and python setup.py develop
  • Go to smhr/smh/gui and open with ipython __main__.py. It should crash with a message about libpyside, saying something is not found. This is because a file is named wrong within anaconda.
    • Go to ~/anaconda/lib (or the equivalent if you made a separate environment), and make a symlink or file copy for libpyside-python2.7.1.1.dylib from one that looks almost the same (e.g. libpyside-python2.7.1.1.1.dylib)
    • Try to open SMHR again, and it will crash. Do the same thing for libshiboken when you see this error message.
    • If you have multiple anaconda installations: make sure that you are doing all of this with the correct anaconda (e.g. anaconda2)
  • If you have problems with qt_menu.nib, use ~/anaconda/bin/python.app __main__.py instead of ipython. If you have multiple anaconda installations, make sure you change anaconda to anaconda2 or whatever you installed it to. (I have fixed this on my laptop by copying it somewhere but I cannot find where. Some possible places that could help):
~/anaconda/python.app/Contents/Resources/qt_menu.nib
~/anaconda/pkgs/launcher-1.0.0-1/launcherapp/Contents/Resources/qt_menu.nib
~/anaconda/pkgs/launcher-1.0.0-2/launcherapp/Contents/Resources/qt_menu.nib
~/anaconda/pkgs/python.app-1.2-py27_3/pythonapp/Contents/Resources/qt_menu.nib
~/anaconda/pkgs/python.app-1.2-py27_4/pythonapp/Contents/Resources/qt_menu.nib
~/anaconda/Launcher.app/Contents/Resources/qt_menu.nib
/usr/local/Cellar/qt/4.8.7_1/lib/QtGui.framework/Versions/4/Resources/qt_menu.nib
  • There are sometimes some problems with segfaults due to the GUI library. We hope this will go away when we move away from pyside. Sorry.

Some notes from Henrique Reggiani that may be helpful to some:

I am running everything under MAC OS X 10.14.6
My Python 2.7 install is the system install (I do not use anaconda and I only have one python install per version, so I manage packages as needed).

I install qt4 from macports (system install). I already have qt5 but it did not create any problems.
I had to export in my .profile/.bash_rc the path to the qmake executable:
(export PATH="$PATH:/opt/local/libexec/qt4/bin") – this is the system path

Then PySide was not finding the libpyside-python2.7.1.2.dylib. The libraries were all correctly placed and named. To solve the problem you go into $PYTHONPATH/site-packages/PySide and run this script:

MOOG

It is currently recommended that you use this version of MOOG: https://github.com/alexji/moog17scat

This version is not the most efficient, and it computes synthetic spectra only to about 0.003 accuracy. It is modified from the 2017 February version of MOOG from Chris Sneden's website (there has been an update made in Nov 2019 that Alex has not investigated yet). It includes Jennifer Sobeck's scattering routines (turned on and off with the flag scat, which is not true in the default MOOG 2017) and the fixes to the Barklem damping that were implemented in the 2014 MOOG refactoring.

Note that by default right now, we require you to have an executable called MOOGSILENT callable from your $PATH environment variable. Specifically, we use the version of MOOG that you get from which MOOGSILENT. We may change this in the future.

About

Spectroscopy Made Hard(er)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 99.7%
  • Shell 0.3%