Does this tool take in fully-processed bams as input or unprocessed bams?

[claudiadast]

When running bamhash on a bam file, should the bam be fully processed (i.e. sorted, deduped, and recalibrated)?

I just tested bamhash on a fully-processed bam file and its source paired-end fastq files, however the resulting hashes differ, so I'm wondering if it's because the input bam was fully processed.

UPDATE:

I just re-ran bamhash on a non-fully processed bam file, and got the following results:

bam result:
2490f971d6f15fa2	764438888

source fastq result:
2490f971d6f15fa2	382219444

What do the two columns represent? Is it enough that one of the columns match between the files?