fastqcClassify

#!/bin/gawk -f
#
# Copyright (c) 2018 German Cancer Research Center (DKFZ).
#
# Distributed under the MIT License (license terms are at https://github.com/DKFZ-ODCF/AlignmentAndQCWorkflows).
#
# Author: Xavier Pastor Hostench

# Variables
# sum: float; cumulated value of the firs quantile
# n: integer; total number of positions evaluated
# down: boolean; Was any median below 21?
# status: 

BEGIN{FS="\t";OFS="\t";sum=0;n=0;down=1==2} # Set the field delimiters and initialize 'sum' and 'n' to 0, and 'down' to false
/Median/{flag=1;next}/^>>END/{flag=0} # Extracts the text block that goes from the line matching 'Median' to the first line starting with '>>END', excluding these lines
flag{sum+=$4;n++;down=down || $3<=20} # Processes the lines of interest: accumulates the value of the first quantile in 'sum', updates the number of positions evaluated in 'n' and checks if the median was above 20
END{status="PASS"; # Once all the positions have been read makes the final calculations: sets 'status' to 'PASS' by default
avg=sum/n; # calculates the average of the first quantile
if (avg<=28){if (avg<=20) {status="FAIL"} else {status="WARN"}}; # updates 'status' to 'WARN' if it's lower or equal to 28, and to 'FAIL' if it's lower or equal to 20
if (down && status == "PASS"){status="WARN"}; # downgrades 'status' to 'WARN' if it had been left as 'PASS' but one of the medians was lower or equal to 20
print status} # Finally prints 'status' ('PASS', 'WARN' or 'FAIL')