Fix argument order in the numpy, numba and parallel Monte Carlo wrappers

[Cashubski]

Summary

Four of the six Monte Carlo methods of EquityVanillaOption return NaN for every input:

call = EquityVanillaOption(Date(1, 1, 2027), 105.0, OptionTypes.EUROPEAN_CALL)
call.value_mc(value_dt, 100.0, discount_curve, dividend_curve, BlackScholes(0.3), 20000, 42)             # 10.9112
call.value_mc_numpy_only(value_dt, 100.0, discount_curve, dividend_curve, BlackScholes(0.3), 20000, 42)  # nan
call.value_mc_numba_only(...)      # nan
call.value_mc_numba_parallel(...)  # nan
call.value_mc_numpy_numba(...)     # nan

Cause. Every model function in black_scholes_mc.py has the signature (s, t, k, r, q, v, opt_type, num_paths, seed, use_sobol). The four broken wrappers call them with (s, t, k, opt_type, r, q, v, ...), so the model receives r=opt_type, q=r, v=q, opt_type=vol; the volatility is a float that matches no option type and the result is NaN. value_mc and value_mc_nonumba_nonumpy pass the arguments in the right order and work.

Fix. Pass the arguments in the model functions' order in all four wrappers.

Test. test_monte_carlo_variants_agree_with_analytic prices the same call with all six variants and a fixed seed and checks each is finite and within sampling error of the closed form. Full unit suite passes (1,059 tests).

[domokane]

Thanks.