Add new wild cluster bootstrap methods as in MNW (2022)

[s3alfisc]

Paper: http://qed.econ.queensu.ca/faculty/mackinnon/working-papers/qed_wp_1485.pdf

[s3alfisc]

First commit into new branch, fast_n_reliable, with implementations of WCRx3 and WCUx3.

[s3alfisc]

Next steps:

• [x] test, test, test
• [x] cleanup, delete objects currently not required
• [x] add functionality to boottest.fixest, boottest.felm
• [x] add WCRx1, WCRx3
• [ ] add option for fixed effect
• [ ] add option for WLS
• [ ] add option for multi-param hypotheses
• [ ] add option to compute confidence intervals
• [x] change the UI of boottest() - likely soft-deprecate impose_null function argument & instead introduce algo_type argument with options WCR11, WCR31 etc and engine with options R or WildBootTests.jl.

Additionally, implement changes to summclust::summclust (which is used to compute CRV3 t-stats), which currently computes measures of leverage for all k covariates, as it can be slow.

[s3alfisc]

Close this for now. I will open new issues for currently not-yet implemented functionality (fixed effect support, CIs, etc).