Implement heritability estimators

[shz9]

I'd like to provide implementation for summary statistics-based and individual-level data heritability estimators. This is important for many downstream models as well as for data interpretation.

Here are some popular heritability estimators that we can easily implement with the magenpy interfaces:

• LD Score regression (ldsc). We already have a simple estimator implemented, but it would be nice to have a fully functional implementation that incorporates annotations, estimates standard errors, etc.
• The GWASH estimator: This is an interesting estimator proposed by Schwartzman et al. a few years ago based on important work by Lee Dicker. The estimator seems simple enough to implement with the magenpy LD interfaces.
• The Haseman-Elston heritability estimator based on individual-level data.

We may add other estimators down the line if there is interest.