Format BigQuery pipe syntax queries correctly

Normalized profile

What this task asks for

Add parsing and formatting support for BigQuery pipe syntax queries without changing traditional SQL formatting. BigQuery pipe syntax chains transformations via `|>` instead of nested clauses. Pipe queries start with standalone `FROM` and each subsequent `|>` step occupies its own line at base indentation. Clauses that the existing formatter treats as indented clauses (`WHERE`, `SELECT`, `ORDER BY`, `AGGREGATE`, `EXTEND`, `SET`, `DROP`) place their body on a new indented line after the keyword. Clauses that the existing formatter treats as one-line clauses (`LIMIT`, `JOIN` and its variants, `AS`) keep their content on the same line as the keyword. Pipe-exclusive clauses absent from standard SQL include `AGGREGATE` with an optional nested `GROUP BY` sub-clause requiring its own indentation level, `EXTEND` for computed columns, `SET` for replacing values, `DROP` for removing columns, and `AS` for naming intermediates. `|>` must tokenize as a distinct type, not bitwise `|` plus `>`.