Add ability to store multiple cells per HDF5

[WardLT]

Fixes #61

• [x] Add warning if metadata differs between cells added to a H5
• [x] Add an option that reads any data from an h5, even if a user doesn't know the prefixes

[coveralls]

Pull Request Test Coverage Report for Build 8973181701

Details

• 53 of 54 (98.15%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.7%) to 81.959%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
batdata/data.py	53	54	98.15%
<!--	Total:	53	54	98.15%	-->

Totals
Change from base Build 8784024460:	0.7%
Covered Lines:	686
Relevant Lines:	837

---

💛 - Coveralls

[WardLT]

Good point a out the potential annoyance of having to lookup prefixes before reading. How about I add an 'auto' option to the read which will grab data belonging to the first prefix then warn if there are more than one available?

[victorventuri]

I think the 'auto' idea is good, my question there is: how is the "first" prefix identified? I mean, there's no inherent order to them, right?

I also think there should be a 'all' option, which basically loads all prefixes available. If you think it's wise, the 'all' can also output a warning for memory constraints.

[WardLT]

@victorventuri , I opted to add a "load by index" option to supply the need for "any." prefix=0 will load the first prefix, alphabetically.

Does this give you the features you'd need?