support function: current_timestamp, current_date, now

same functions in Spark SQL.

1. There is `unix_timestamp("")` as a temp solution, not convenient though. unix_timestamp returns unix epouch time in seconds as int64, one must write `timestamp (unix_timestamp("") * 1000)` to archive current_timestamp.
2. all `current_*` function should executed at begining of one SQL, means all `current_*` functions calls in same SQL should guarantee return the same value, no matter where the `current_*` call is.

Look for existing timestamp functions like `unix_timestamp` in the SQL function implementation. The new functions `current_timestamp`, `current_date`, and `now` need to be added, ensuring they return the same value for all calls within a single SQL statement. Check how query execution context is managed to guarantee consistency. Test by writing SQL queries that call these functions multiple times.

Do mô hình lập chỉ mục viết ra từ nội dung của issue.