Revise article

[gerlero]

As part of openjournals/joss-reviews#7633, make the following proposed changes to the paper.

From #316 (comment)

• PAPER: Comment on why an external parser was written instead of C++ bindings for OpenFOAM's own parser.
• PAPER: Mention that looping over a TimeDirectory will iterate over the reconstructed time folders.
• PAPER: soften language on statement of what files FoamFile can parse & refer readers to FoamFile API docs.
• PAPER: fix inconsistency numpy.ndarray vs np.ndarray.

From #317 (comment):

• PAPER: mention other OpenFOAM-related Python packages in the statement of need, including fluidfoam and fluidsimfoam.

From #318 (comment):

• PAPER: mention hardware used to measure the reported speedup.

[AndreWeiner]

@gerlero the PR looks good from my point of view.

[Failxxx]

Looks good to me 👍🏻