`Map<String, Any?>` to `DataRow` conversion with nested maps

While `map.toDataRow()` is very helpful, there's no easy way to fully unfold nested maps to to datarows, like

```kt
val map = mapOf(
"name" to "a",
"metadata" to
mapOf(
"country" to "Philippines",
"region" to mapOf("name" to "Caraga", "code" to "XIII"),
"population" to mapOf("value" to "12345", "year" to 2020),
"wrongMap" to mapOf(1 to 2, 4 to 4),
),
)
```
Similarly to `unfold` and `toDataFrame()`, we could add a `maxDepth` parameter that would allow us to specify the maximum depth of nesting we scan for and convert.
`convertKeysToString: Boolean` might be another nice addition because there's no way to tell if keys are strings or not until we encounter a `ClassCastException`.

`Iterable>.toDataFrame()` should probably gain the same treatment.

And if we're converting keys to strings anyways, let's make it work for `Map`, why not

Start by reading the implementations and tests for map.toDataRow(), unfold, toDataFrame(), and Iterable>.toDataFrame(). Clarify how maxDepth and convertKeysToString should behave for nested and non-string-keyed maps, then ensure the map and iterable conversions consistently produce the intended DataRows and have coverage for the example cases.

Written by the indexing model from the issue text.