bump version to 2.3

using meshgrid for the data
added interpolation function

(unsure how this worked before)

new function plt.get_dst returns a dst object
function plt.save_dst forces you to provide a file name (and use get_dst)

- current convert A vs mA
- forgot to write mass (= mc2)
- using file write() rather than print function
print was not working correctly on both python2 and 3
this has been tested to give correct result on both

(was not working on my python2)
added plt.save_dst()
fixed pack function in dst.save() for python3

replaced xrange -> range (py3 compatibility)

made sure forward compatibility is preserved

fix for IS energy levels (use keV instead of MeV when max(E)<0.1)

resembles TraceWin distribution plot

this might not work in all cases just yet

by default false. if true overwrite existing group

also some cosmetic cleaning

can now save toutatis files correctly

toutatis distributions has a 7th line of zeros
now can read both correctly

if merging files with differing length of arrays, assume
that the shorter one lost all beam

close buffer in case trying to save file multiple times..

new function which saves the density data into a HDF5 file format
useful after e.g. combining multiple data..

seems number of header lines vary..

in dst and plt
also cosmetic changes (spaces, ordering)

dictionary return now excludes lost particles
this was essential for the twiss calculation apparently