Add docs and tests for `frames` pivot aggregation

Add docs and tests for frames pivot aggregation

core/src/main/kotlin/org/jetbrains/kotlinx/dataframe/api/frames.kt contains two small public APIs without KDocs and without dedicated unit tests:

• Pivot<T>.frames(): DataRow<T>
• PivotGroupBy<T>.frames(): DataFrame<T>

The function is briefly mentioned on the pivot site page in the aggregation grammar and examples:
https://kotlin.github.io/dataframe/pivot.html#examples-of-aggregation

What it does

frames() aggregates pivot groups without reducing their rows. Each pivot key becomes a resulting column, and the corresponding grouped rows are stored as frame values.

For Pivot, the result is a DataRow.

For PivotGroupBy, the result is a DataFrame with one row per grouping key.

Scope

• Add concise KDocs for both frames() overloads in frames.kt.
• Add one basic unit test for Pivot.frames().
• Improve the pivot site docs with a short explanation, input/output tables, and a runnable example for frames().

Suggested KDoc content

/**
 * Aggregates this [Pivot] by keeping each pivot group as a frame.
 *
 * Each pivot key becomes a column in the resulting [DataRow], and each value is
 * a [DataFrame] containing rows that belong to the corresponding pivot group.
 */

For PivotGroupBy.frames(), use the same idea but mention that the result is a DataFrame
with one row per grouping key.

Basic test scenario

• Given a dataframe with a Boolean pivot column, e.g. isHappy.
• When calling df.pivot { isHappy }.frames().
• Then the result contains columns for pivot keys, e.g. true and false.
• And each resulting value is a dataframe with rows belonging to that pivot group.

Test sketch

@Test
fun `pivot frames keeps pivot groups as frame values`() {
    val df = dataFrameOf("name", "isHappy", "age")(
        "Alice", true, 20,
        "Bob", false, 30,
        "Charlie", true, 40,
    )

    val result = df.pivot("isHappy").frames()

    result.columnNames() shouldBe listOf("true", "false")

    val happy = result["true"] as DataFrame<*>
    val unhappy = result["false"] as DataFrame<*>

    happy.rowsCount() shouldBe 2
    unhappy.rowsCount() shouldBe 1
}

Acceptance criteria

• Pivot.frames() and PivotGroupBy.frames() have KDocs.
• There is at least one dedicated unit test for frames().
• The pivot documentation includes a clear frames() example with input/output tables.

Start with core/src/main/kotlin/org/jetbrains/kotlinx/dataframe/api/frames.kt and review the existing pivot aggregation tests before adding coverage for both pivot-key results and grouped frame values. Update the pivot site page at the aggregation examples section with the requested explanation, tables, and runnable example. Done means both overloads have KDocs, a dedicated unit test passes, and the site documentation includes a clear frames() example.