Watch a folder, split fast5 into 150-file chunks and run guppy basecalling from a chunk list.

watch_and_basecalling.sh

#!/usr/bin/env bash
# Author: Thanh Le Viet
# This script will split every consecutive fast5 files into a batch file list  for basecalling with guppy
# It is used for "live" basecalling while the sequencing still running on another machine.
# Command: bash ./run_basecalling.sh
# summary_file="sequencing_summary_FAO15487_23198198.txt"
# Usage: run watch_and_basecalling.sh sequencing_summary_FAO15487_23198198.txt
# Note: each run has a different summary_file name.

summary_file=$1

#ROOT ONT 
ONT=/home/prom/ont-software/ont-guppy-4.0.15/ont-guppy

while true; do
ls -tr  $PWD/fast5/*.fast5 | split --lines=150 --numeric-suffixes=1 - rec_batch_

echo "Generated...."

wc -l rec_batch_*
echo "================================================="
batches=`ls -v rec_batch_*`

for b in $batches;
do

#echo "$b"
batch_folder="./fastq/${b}"

if [[ -d $batch_folder ]]  && [[ -f "$batch_folder/sequencing_telemetry.js"  ]]
then 

echo "$b has been already basecalled"

else 
	items="x$(cat $b | wc -l)"

if [[ $items != "x150" ]] && [[ ! -f $summary_file  ]]
then
	 echo "$b is not basecalled and fast5 files smaller than 150"
	 date
else

if [[ ! -f .lockbc ]] 
then
	echo "Start basecalling on $b"
	echo "$b" > .lockbc
	#Start guppy	
	"$ONT/bin/guppy_basecaller" \
	--qscore_filtering \
	--input_file_list $b  \
	-s "./fastq/${b}" \
	-c "$ONT/data/dna_r9.4.1_450bps_hac.cfg" \
	-x 'cuda:1,2,3,0' > "log_${b}.txt"

if [[ $? -eq 0 ]]
then
	echo "Finished sucessfully $b" 
	rm .lockbc
else
	echo "Basecalling for $b is failed"
fi
else
	echo "Cannot start basecalling for $b. Please check .lockbc file"
fi


fi

fi

done
sleep 600
done