Add unit tests for `concat` APIs

Add unit tests for concat APIs

Now it's only 10-15% of lines and methods are covered

core/src/test/kotlin/org/jetbrains/kotlinx/dataframe/api/concat.kt currently covers only a small subset of concat behavior.

The implementation in core/src/main/kotlin/org/jetbrains/kotlinx/dataframe/api/concat.kt contains multiple public overloads, and the site page lists several supported use cases:
https://kotlin.github.io/dataframe/concat.html

Add focused unit tests for the main overload groups and documented behavior.

Test scenarios as an option

DataFrame.concat

• df.concat(df1, df2) appends rows from multiple dataframes.
• df1 concat df2 appends rows from another dataframe.
• df.concat(frames: Iterable<DataFrame<T>>) appends rows from an iterable of dataframes.
• Concatenating dataframes with the same schema preserves column names, values, and row order.
• Concatenating dataframes with different schemas performs schema unification.
• Missing columns are filled with null.
• Columns with the same name and different compatible types are unified to the lowest common type.
• Concatenating with an empty dataframe preserves the expected schema and rows.
• Concatenating dataframes with no columns preserves row count correctly.

DataColumn.concat

• column.concat(other) appends values from another column.
• column.concat(col1, col2) appends values from multiple columns.
• The resulting column keeps the name of the first column.
• Columns with compatible but different value types are unified correctly.
• Concatenating with an empty column preserves expected values and type.
• DataColumn<DataFrame<T>>.concat() concatenates all dataframes stored in a frame column.
• DataColumn<Iterable<T>>.concat() flattens iterable values into a single list.

DataRow.concat

• row.concat(row1, row2) creates a dataframe from the receiver row and provided rows.
• Row order is preserved.
• Rows with the same schema are concatenated correctly.
• Rows with different schemas are unified consistently with dataframe concat.

Iterable.concat

• Iterable<DataFrame<T>>.concat() concatenates all dataframes in order.
• Empty iterable of dataframes returns an empty dataframe.
• Iterable<DataColumn<T>>.concat() concatenates all columns in order.
• Empty iterable of columns returns an empty column.
• Iterable<DataRow<T>?>.concat() creates a dataframe from rows.
• null rows in Iterable<DataRow<T>?>.concat() produce empty rows with one row, according to current implementation.

GroupBy.concat

• groupBy.concat() concatenates groups back into a dataframe.
• Group keys are not added as separate columns.
• Row order follows group order.
• Existing columns with the same names as key columns are preserved as group columns.

GroupBy.concatWithKeys

• groupBy.concatWithKeys() concatenates groups and adds missing key columns.
• Key column values are repeated for every row in the corresponding group.
• Key columns that already exist inside groups are not overwritten.
• Result values are checked, not only column names.

ReducedGroupBy.concat

• reducedGroupBy.concat() applies the reducer to each group and concatenates the produced dataframes.
• Result preserves reducer output order and schema.

Acceptance criteria

• concat.kt tests cover every public overload group from concat.kt.
• Tests cover all use cases listed on https://kotlin.github.io/dataframe/concat.html.
• Schema unification behavior is covered, including missing columns filled with null.
• GroupBy.concat() and GroupBy.concatWithKeys() are both tested with actual values.
• Existing empty-dataframe tests are preserved and extended where needed.