Support for product types

[jserranohidalgo]

Description

The main motivation for this PR is increasing the support for product types (i.e. case classes) in doric. This mainly includes:

• Creating literal columns from (non-custom) product types.
• Deserializing struct columns into their corresponding (non-custom) case class instances.
• Deserializing columns into custom case class instances (e.g. creating a User instance from a string containing name and age).
• Accessing fields of user-defined product types (not only of RowColumns).
• Safe-access: if the specified field is missing, a compilation error ensues.

Some other changes in the LiteralSparkType and SparkType type classes have been also included. For instance:

• Spark types now include a nullable field. This is for ensuring that the datatype calculated by doric (statically, through implicits), matches exactly the datatype calculated reflectively by Spark.
• Add implicit Spark type instances for some new Spark datatypes: DateType, TimeStampType, ...
• Literal columns are created through typedlit in order to serialize product types (and combinations thereof) out-of-the-box.

Related Issues

Resolves:

• Static & dynamic field accessors for product types #213

Partially solves:

• Coverage of Spark types #228

Related but NOT SOLVED by this PR:

• Custom types within product types #250
• Field access for custom product types  #249

How Has This Been Tested?

Tests for Spark and Literal Spark types have been divded into four major specs:

• Spark and Literal Spark Types tests - just checks that the implicit approach by doric and the reflective method by Spark, give the same datatypes.
• Custom types tests. Tests related to custom type mappings.
• Serialization tests. Tests related to the creation of literal columns for Scala values.
• Deserialization tests- Tests related to the creation of Scala values from DataFrame rows.

[codecov]

Codecov Report

Merging #251 (f7a5ce5) into main (888b56a) will increase coverage by 0.01%.
The diff coverage is 97.78%.

@@            Coverage Diff             @@
##             main     #251      +/-   ##
==========================================
+ Coverage   97.28%   97.29%   +0.01%     
==========================================
  Files          58       58              
  Lines        1028     1107      +79     
  Branches       10       10              
==========================================
+ Hits         1000     1077      +77     
- Misses         28       30       +2     

Flag	Coverage Δ
spark-2.4.x	93.40% <91.85%> (-0.51%)	⬇️
spark-3.0.x	96.58% <97.78%> (+0.07%)	⬆️
spark-3.1.x	97.40% <97.78%> (+0.01%)	⬆️
spark-3.2.x	97.40% <97.78%> (+0.01%)	⬆️
spark-3.3.x	97.40% <97.78%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
core/src/main/scala/doric/sem/Errors.scala	92.31% <ø> (ø)
core/src/main/scala/doric/types/SparkType.scala	96.00% <96.88%> (+0.62%)	⬆️
.../src/main/scala/doric/types/LiteralSparkType.scala	98.65% <98.36%> (-1.35%)	⬇️
core/src/main/scala/doric/DoricColumn.scala	66.67% <100.00%> (-4.76%)	⬇️
core/src/main/scala/doric/syntax/DStructs.scala	100.00% <100.00%> (ø)
...c/main/scala/doric/syntax/LiteralConversions.scala	100.00% <100.00%> (ø)
core/src/main/scala/doric/types/NumericType.scala	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 888b56a...f7a5ce5. Read the comment docs.

[jserranohidalgo]

@alfonsorr @eruizalo los dos ignore que quedan en los tests se resolverán con el ticket #250
salvo que tengáis algún comentario, por mi parte no veo nada más a añdir/cambiar