Make InferenceData and Dataset type-unstable

[sethaxen]

Fixes #15 by removing type annotations from InferenceData and Dataset. This makes high-order operations on these types compile quickly (and possibly automatically precompile), while operations on arrays are still compiled for the specific dimensions of that array and so are still fast.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.75%. Comparing base (766efd4) to head (676e82d).
⚠️ Report is 40 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #24      +/-   ##
==========================================
+ Coverage   97.08%   97.75%   +0.66%     
==========================================
  Files          10        9       -1     
  Lines         309      312       +3     
==========================================
+ Hits          300      305       +5     
+ Misses          9        7       -2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sethaxen]

I'd like to in this PR also explore the feasibility of making Dataset have dictionary storage instead of NamedTuple. I expect this will eliminate even more compile time lag, e.g. show, and will allow new variables to be added to existing datasets.

[sethaxen]

Supporting mutable and type-unstable Dataset will require some changes to DimensionalData (see rafaqz/DimensionalData.jl#457), so that will wait for another PR in a future breaking release.