add fastq-scan additional outputs - read lengths

[kapsakcj]

I oftentimes have to dig deep into the execution directories to find the fastq-scan output JSON to see mean and median read lengths

It would be nice to float the mean and median read lengths to 2 output columns.

Additionally output the JSON as a 3rd new column.

The average/mean read length is available in the cg_pipeline output TSV, which is an output column, but you have to download the file to view for each sample (bit of a pain)