Getting split reads from MetaPhlAn 4.2.4?

Hi,
I understand from the discussions on the forum that there is not yet support for long reads in HUMAnN.

However, I have run MetaPhlAn 4.2.4 with the option split_reads on nanopore data and I was wondering if it would be possible to get a fastq file with the split reads used by MetaPhlAn that could be used as input to HUMAnN? Of course it would be better to be able to use full length long reads but maybe this option could be faster to implement?

Thanks in advance!