Nullable accessors and fixed convertTo

[Jolanrensen]

Let's say you have a Dataframe of Location:

@DataSchema
public data class Location(
    val name: String,
    val gps: Gps?,
)

@DataSchema
public data class Gps(
    val latitude: Double,
    val longitude: Double,
)

val locations: DataFrame<Location> = listOf(
    Location("Home", Gps(0.0, 0.0)),
    Location("Away", null),
).toDataFrame()

In the current implementation working with gps is very difficult. Since it is generated as DataRow<Gps?>, the extension accessors like latitude and longitude don't work. Plus, in the dataframe, the gps property is shown as gps: {latitude:Double?, longitude:Double?}, which further shows that the accessor would be helpful.
The generated accessors do not clash with the non-nullable variants.

Now possible:

locations.filter {
    (gps.latitude ?: Double.MIN_VALUE) >= 0.0
}

Also fixing #176. might need some more convertTo tests, but the example from the issue now works.

[koperagen]

let's use rebase instead of merge to pull changes from dev branch in the future. i think that would make 1 less commit in the history

[koperagen]

Do we need it?

[Jolanrensen]

Maybe, I'm not sure. I needed toNullable, but for completeness if we need it in the future I thought it best to also create a reverse option. Might save time to figure out how it works again