Bookeeping release (but one less dot in the version string)

Our latest release was more than a year ago! That's crazy!!
Development has been a bit slow recently. This is because I've been working mainly with Brendon Brewer's DNest4 and I haven't found the time to integrate it. This integration is on the roadmap and I want it to be quite tight. In the meanwhile this 0.1 release represents one step up the version ladder and shows that OPEN is maturing.

Here's an abridged changelog:

• add Travis CI integration
• add quantifiedcode integration
• badges for these things in the README
• multiband periodogram for data with offset
• improved star name finder, but I'm sure it's not bullet-proof yet
• add restrict by S/N option
• new plot for several quantities together
• window function as function of frequency
• chunks option for correlate
• AIC and BIC for nest best fit
• new Hoeffding periodogram
• pretty publication-ready plots
• improve bootstrap FAP calculation
• glscalculation is distributed with OpenMP
• specify the system when running the per command

Are we there yet?

Wow so many new things!
200+ commits and 6+ months later, here we are. Not yet at our final destination, but gaining up some momentum. Here's some of what's happened:

• working with multiple observatory offsets
  • individual offsets in plot of observations
• fix log-likelihood expression that has been slowing things down (and making them wrong)
• numpy arrays in window function calculation (nice speedup)
• a new option for restrict command to remove by index
• a very nice --gui option for the restrict command
  • with correct side effects on extras and system provenance
• introducing the MultiNest submodule
  • bayesian evidence calculation done right!
  • full resuming and restarting
  • automatic model selection run for up to 3 planets
• introducing RJDNest3 submodule
  • this is very experimental! As in it actually doesn't work...
• new options for planet creator
• treat special case ecc=0 where pericentre is undefined, so that t0 has physical meaning
• catching ecc=0 (may speed things up a bit)
• periodogram FAP calculation by bootstrap
• option in per command to avoid plotting, useful for macros
• --quiet option for read command for no output at all
• stellar mass as attribute of system
• remove makefiles from repository; build is now done with SCons
  • allow linking with system LAPACK from package liblapack-dev
  • don't need setup.py anymore
  • allow fortran compiler path as a build option
  • find f2py executable with SCons
• rename commands.py so it doesn't conflict with std library commands module
• rotation period from MH2008
• phase plots!
• CLEAN periodogram extension (with code from Annelies Mortier); new clean command
• new lopast and julian extensions
• save figures from plot and per
• plot residuals of fit
• show years on top axis of some plots, new conversion from bjd
• ask to remove linear trend off of FWHM when doing periodogram and plotting
• add resid (residuals) option for periodograms
• BGLS implemented
• new option syntax for per; fixes issue #4
• introducing the new GP module; needs gfortran > v4.6 (maybe 4.8?)
  (heavily inpired from dfm's george)
  • train GP hyperparameters beforehand (on FWHM or logRhk)
  • tests with 4 hyperparameters: ExpSquared * ExpSine2 kernels, following Haywood2014
  • implements product of kernels
  • using LAPACK routines for performance with bigger matrices
  • allows for GP only, without planets
  • wrapper allowing python calls to GP routines

Starting volunteer tests

This is a special release.
Many new things were added and many were dropped but the goal of the release is to provide a starting point to test OPEN in other machines. A lot of bug fixes are expected in the coming months.

Another Brick in the Wall

• Fix issue with .rdb files.

Extras names and order are now read directly from the header of the file. This fixes an issue with non-standard column names and orders because, apparently, there is no standard...
This does break the (new) saverdb command.

• add_noise and remove_rotation commands

Not yet ready for anything serious (and actually quite ad-hoc methods, so let's see...)

• Fixes in periodograms and plots
• Parallel evaluation of gls and bls periodograms
• Handle tilde expansion in filenames
• Very basic genyorbit command
• 'verbose' option for read
• Plenty of options for periodogram command
• Initial implementations of the DE and the DEMC algorithms
• New macros
• Basic initial implementations of an MCMC class and some result analysis code

First real release

Lots of new features were added for this release. We basically progressed from proof-of-concept material down to (almost) ready to use software.
The IO part of OPEN is basically settled as is most of the high-level organization. I am also happy with the calculation of the periodograms and the overall performance. Data manipulation is half-way there with the new 'restrict' options.
The place where things are still going slow is in the core fitting routines, which is ok, this being the hardest part. Nevertheless, the genetic algorithm and the multinest analysis are taking baby steps towards perfection and will be subject to a lot of work in the near future.

Pre-alpha release

This is a proof-of-concept release. Now going into design and implementation considerations.