Allow structs and list of structs in g/->dataset (#335)

[WaqasAliAbbasi]

Fixes #335, Works well for:

• maps
• list of maps

It's recursive so it supports maps inside maps as well.

Caveats for list of maps

• All maps in the list of maps need to be of the same schema. This won't work:

  g/->dataset [{:a 1 :b [{:b "c"} {:b "d" :e "e"}]}]
  ; Execution error (ArrayIndexOutOfBoundsException) at org.apache.spark.sql.catalyst.CatalystTypeConverters$StructConverter/toCatalystImpl (CatalystTypeConverters.scala:252).
  ; Index 1 out of bounds for length 1

• Keys of the maps after the first one are ignored. For example:

  (-> (g/->dataset [{:a 1 :b [{:b "c"} {:d "e"}]}])
    (g/collect-vals))
  ; ([1 ({:b "c"} {:b "e"})])

^Resolved the above caveats

[codecov-commenter]

Codecov Report

Merging #336 (0ca9f3f) into develop (e3c88ad) will decrease coverage by 0.127%. The diff coverage is 89.743%.

@@              Coverage Diff              @@
##           develop      #336       +/-   ##
=============================================
- Coverage   99.967%   99.839%   -0.128%     
=============================================
  Files           36        36               
  Lines         3085      3118       +33     
  Branches         0         4        +4     
=============================================
+ Hits          3084      3113       +29     
  Misses           1         1               
- Partials         0         4        +4     

Impacted Files	Coverage Δ
...rc/clojure/zero_one/geni/core/dataset_creation.clj	97.590% <89.743%> (-2.410%)	:arrow_down:

[anthony-khong]

Thank you @WaqasAliAbbasi! I'll make a new release today!