# Error handling.
set -ue

function stats() {
    # Summarizes value in the clustering file.
    grep at | awk ' { print $5 } ' | tr -d '%' | datamash min 1 max 1 median 1
}

# Sample size.
N=300

# Random seed.
SEED=$RANDOM

# Select a random sample.
seqtk sample -s $SEED refs/sars2.fa $N > subset.fa

# Cluster the sequences.
cd-hit -i subset.fa -o out >> log

# Print the seed (so we know it in case we want to repeat selection).
printf "$SEED\t"

# Print the average similarities.
cat out.clstr | stats