Allow specifying separate dtypes for solution, objective, and measures

[btjanaka]

Description

Previously, dtype in the archive could only be a single value specifying the dtype for the solution, objective, and measures all at once. Now, it is allowed to be a dict specifying separate dtypes for each of these, e.g.,

    archive = GridArchive(
        solution_dim=3,
        dims=[10, 10],
        ranges=[(-1, 1), (-2, 2)],
        dtype={
            "solution": object,
            "objective": np.float32,
            "measures": np.float32,
        },
    )

Comments:

• This PR is a followup to #470 which allowed expressing the different dtypes present in the archive.
• Emitters used the dtype of their archive, e.g., archive.dtype. Now, they choose a specific field's dtype to use. The emitters implemented thus far all choose archive.dtypes["solution"], i.e., they have the same dtype as the solution.

TODO

• [x] Write docstrings
• [x] Implement change with parse_dtype function
• [x] Add tests

Status

• [x] I have read the guidelines in CONTRIBUTING.md
• [x] I have formatted my code using yapf
• [x] I have tested my code by running pytest
• [x] I have linted my code with pylint
• [x] I have added a one-line description of my change to the changelog in HISTORY.md
• [x] This PR is ready to go