Gravar-mail: Benchmarking of long-read correction methods