Add KDocs for function `Split`

Add KDocs for non-deprecated split APIs

The file core/src/main/kotlin/org/jetbrains/kotlinx/dataframe/api/split.kt holds the public split operation and its Split.* / SplitWithTransform.* builder chain. Almost all of it is undocumented — only Split.by(regex) and Split.match(...) currently have KDocs.

Please add concise KDocs for the current non-deprecated APIs only. Deprecated and binary-compatibility (@AccessApiOverload, string-ReplaceWith) overloads are out of scope.

Scope

Non-deprecated APIs in split.kt, by stage of the chain:

• Entry points — DataFrame.split(...) (selector and vararg String) and DataColumn<Iterable<*>>.splitInto(...).
• Builder types — the Split<T, C> class (incl. cast()), the SplitWithTransform<T, C, R> class, and the ColumnNamesGenerator typealias.
• Split.* configuration — by(...) (char / string delimiters, iterable transform; the regex overload is already documented), default(...), match(...) (already documented).
• SplitWithTransform.* / Split.* terminals (produce a DataFrame) — into(...), inward(...), intoColumns(), intoRows(...), inplace(), and SplitWithTransform.default(...).

Out of scope: all deprecated / @AccessApiOverload overloads, the already-documented Split.by(regex) and Split.match(...), and internal/@PublishedApi helpers.

KDocs should explain

• The overall flow: split { columns } selects columns to split → an optional by/match/default step defines how each value is split into parts → a terminal (into / inward / intoColumns / intoRows / inplace) decides where the parts go.
• What each terminal produces:
  • into(...) — new sibling columns named by names (with extraNamesGenerator for the overflow);
  • inward(...) — the same, but nested inside a column group under the source column;
  • intoColumns() — splits a column of DataFrame/rows into columns of a group;
  • intoRows(...) — explodes each split into separate rows (dropEmpty behavior);
  • inplace() — keeps the parts as a List in the original column.
• How replacement column names are chosen: given names, then extraNamesGenerator, then the auto-generated splitN fallback (see the existing by(regex) example).
• by/match semantics: by splits by delimiters / a custom transform; match extracts regex capturing groups; default supplies a value for missing parts.
• That entry points return a lazy Split builder (no work until a terminal is called), and cast() only refines the generic type.

Acceptance criteria

• Every public, non-deprecated declaration in split.kt has a concise KDoc (the file is fully covered), including the Split and SplitWithTransform classes and the ColumnNamesGenerator typealias.
• No documentation is added for deprecated or @AccessApiOverload overloads.
• KDocs distinguish the terminals (into vs inward vs intoColumns vs intoRows vs inplace) and their resulting shape.
• KDocs explain the names / extraNamesGenerator naming logic.
• Style and cross-references are consistent with the existing Split.by(regex) / Split.match KDocs and with the site docs at split.html.