[Distributions] Implement Expon distribution

[shivasankarka]

Exponential Distribution

This PR introduces an implementation of the continuous exponential distribution, Expon.

The implementation supports the standard parameterisation with loc and scale (fields of Expon) enabling shifting and scaling of the exponential distribution.

Traits

This PR also introduces a new trait, RVContinuousLike, designed to parameterise continuous probability distributions (to mimic scipy.stats.rv_continous). At present, the trait defines only the core functionality common to continuous random variables. Expon is the first distribution to conform to this trait. Support will be extended to additional continuous distributions in subsequent PRs.

[shivasankarka]

@forfudan Please check if the format, method names etc look good and if I need to make any other changes. I tried to follow scipy.stats naming conventions and API wherever possible to keep it consistent. If everything looks good, I will open PR for the other distributions. Thanks!

[forfudan]

@shivasankarka I did some changes to your PR and merged it.

1. Expon → Exponential, to be more self-explanatory as other distributions.
2. RVContinuousLike → ContinuouslyDistributed, same as above, this is a not-so-pretty name of scipy (C-style)
3. Remove the Copyable and Movable trait from Exponential because the ContinuouslyDistributed trait alreay contains them.
4. Remove some unused imports.
5. For logcdf, use log(-expm1(-y)) instead of log1p(-exp(-y)) for better stability when x -> loc.