add hashes and rearrange mult handling

[loriab]

Description

This PR is preparatory to #318.

• NEW: adds a check to the basic Molecule pydantic model so multiplicities <1 get caught. This route is accessed when Molecule(..., validate=False). I believe this is a universally good check, but it is a new restriction. Any weird positron, fractional electron, etc. reason to not catch less-than-one mult?
• add many hash checks (incl. all the chgmult variations in validate_and_fill_chgmult testing) to molecules, so we'll know if code starts changing the hash and thereby disrupting the QCA server.
• qcel.molparse has a different error message for less-than-one fragment_multiplicities you try to put through it. qcel.molparse is accessed whenever Molecule(...) or Molecule(..., validate=True) is called.
• qcel.molparse newly allows floats that are ints (e.g., 1.0) for multiplicity. previously it would raise an error about not being an int.
• qcel.molparse newly has a dedicated error for any less-than-one molecular_multiplicity you try to put through it. Previously it raised the usual "Inconsistent" error.
• more tests that chgmult are processing as expected, including some with fractional charges (https://github.com/MolSSI/QCElemental/pull/343/files#diff-29ef453faf072f4ee853122f0f64f6ba594c00dc44599a1f5138946619d9951fR90-R96) for the first time.

Changelog description

Status

• Code base linted
• Ready to go

[loriab]

@awvwgk , is enforcing a lower bound of 1.0 on multiplicity a sensible thing to do even for fractional electron, or is it restricting?

[awvwgk]

I believe a lower bound of 1.0 is sensible, I can't think of a usecase even for a fractional electron which could not be represented this way.

Derk @DerkKooi, do you know any case that is not representable with this lower bound of multiplicity?