General fixes in prep for repository opening

[cjosey]

So, I finally have confirmation OpenDeplete will live up to its name and be open source.

This PR adds a number of fixes I wanted to get into the code before I open it.

• Memory copying is a bit more correct, which was causing bugs in multistage integrators.
• Results now require .h5. This makes automation easier.
• I updated the documentation a bit. It's not good, but it's not broken. Small victories.
• I added an on-by-default semi-infinitely-dilute nuclide addition for all nuclides with reaction rates. I would prefer actual infinitely dilute tallies, but that takes effort.

The last one allows production of nuclides far along a repeated (n,gamma) chain, such as californium.

Once I merge this in, the repo goes public.

[cjosey]

Alright, I've reviewed as much as I can. I'm opening it up.