WIP: SingleStageSurveyDesign for one cluster, one strata and one weights

[smishr]

• Add class for single-stage approximation of multistage sampling
• Works with cluster strata and weights
• Can optionally provide probs instead of weights with different constructor
• Testing for the above, show method

Closes #152 once correctly implemented.

[smishr]

Currently sampsize and cluster variables are different to the analogous design in R, when tested with NHANES and YRBS datasets.

data(yrbs)
yrbs_design <- svydesign(id=~psu, weight=~weight, strata=~stratum, data=yrbs)
data(nhanes)
dnhanes <- svydesign(id=~SDMVPSU, strata=~SDMVSTRA, weights=~WTMEC2YR, nest=TRUE, data=nhanes)

[codecov-commenter]

Codecov Report

Merging #155 (515eb24) into AnalyticSolution (264dcf2) will decrease coverage by 3.44%.
The diff coverage is 35.00%.

@@                 Coverage Diff                  @@
##           AnalyticSolution     #155      +/-   ##
====================================================
- Coverage             74.46%   71.02%   -3.44%     
====================================================
  Files                    13       13              
  Lines                   419      459      +40     
====================================================
+ Hits                    312      326      +14     
- Misses                  107      133      +26     

Impacted Files	Coverage Δ
src/show.jl	0.00% <0.00%> (ø)
src/SurveyDesign.jl	85.08% <60.86%> (-3.53%)	⬇️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[iuliadmtru]

#152 is now closed because it became outdated in the singledesign branch. Should this PR be closed as well?

[smishr]

i think some of the features of this PR can be merged with AnalyticalSolution branch