Fix conversion from floats

[barucden]

Numbers such as 0.3 or 1.1 cannot be represented exactly using Float32 or Float64. We can all digits by converting them to BigFloat:

julia> big(0.3)
0.299999999999999988897769753748434595763683319091796875

julia> big(1.1)
1.100000000000000088817841970012523233890533447265625

This PR updates the Decimal constructors so that floating point numbers are converted to Decimals exactly in the sense that

julia> hash(1.1) == hash(Decimal(1.1))
true

Fixes #85

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.69%. Comparing base (514bf8a) to head (4d37211).

Additional details and impacted files

@@           Coverage Diff           @@
##           master      #86   +/-   ##
=======================================
  Coverage   99.68%   99.69%           
=======================================
  Files          11       11           
  Lines         319      329   +10     
=======================================
+ Hits          318      328   +10     
  Misses          1        1           

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