splash reader for phantom-hdf5 files

[danieljprice]

this is needed, some work sitting in pull request

[dliptai]

submitted pull request #42 to merge updated hdf5-phantom branch into master.

Before approving pull request, it should be checked that reading phantom hdf5 dumps returns the same arrays as reading the phantom binary dump. e.g. pxyzu from GR is currently not being read

[danieljprice]

see comments on #42, this code needs to be rewritten in c consistent with the other hdf5 readers in splash, and also needs to be written in a way that does not hard wire the dataset names in the hdf5 reader, so it does not need ongoing maintenance any time a new quantity is written by phantom. Sorry for not being clearer on requirements here in the first place.