fix out of range argument for np.arccos()

[YAY-C]

due to rounding error in computed cosine values (see get_cos())

• clip value returned by get_cos() to range domain [-1, 1]
• (misc) add verbosity level to e-SPAHM compute script

[YAY-C]

@briling sorry I just also included some verbosity level to suppress some warnings because I'm using these functions in a notebook and its a bit annoying (also it conflicts with the progress bar)
shouldn't change the normal behaviour!

[briling]

the second commit belongs to a different PR i think...

but importantly, it changes the behavior, because the check_nelec() function also raises an error if there is too many electrons to fit into existing basis. we added it after fixing the minao-without-ecp bug for Iodine, and it should catch similar bugs for other bases. so the options are:

1. rename the argument to ignore_too_many_electrons/allow_too_many_electrons or something (so it looks dangerous) but i think it's still dangerous
2. pass this verbose argument to check_nelec() so the error is raised but the warning is suppressed
3. option 2) but rename it to suppress_full_shell_warning or something (i think verbose is too vague)

[briling]

(see the comment)

[briling]

why do you have full-shell molecules anyway?

[liam-o-marsh]

I'm partial to a modified option 3 ("suppress_warnings", in case other warnings get added in the future)

[YAY-C]

@briling @liam-o-marsh
I moves the warning discussion to a new PR and will merge the arccos() fix that you approved!