#!/usr/bin/env bash
#SBATCH -J RainUniq
#SBATCH -N 1
#SBATCH -n 28
#SBATCH -o %j.%N.%a.wrangle-uniq.out.log
#SBATCH -e %j.%N.%a.wrangle-uniq.err.log
#SBATCH -p compute
#SBATCH --time=3-00:00:00

module load utilities/multi
module load readline/7.0
module load gcc/10.2.0

module load python/anaconda/4.6/miniconda/3.7

# TODO: expand this to pull this dynamically from the model
INPUT="${INPUT:-$HOME/rainfallwater_records}";
COUNT_FILE="${COUNT_FILE:4000}"

if [[ -z "${INPUT}" ]]; then
	echo "Error: No input directory specified in the INPUT environment variable.";
	exit 1;
fi
if [[ ! -d "${INPUT}" ]]; then
	echo "Error: That input directory either doesn't exist, isn't a directory, or we don't have permission to access it.";
	exit 3;
fi

export PATH=$HOME/software/bin:$PATH;

echo ">>> Settings";

echo "INPUT $INPUT";

echo ">>> Uniquifying dataset";
/usr/bin/env time -v src/index.mjs uniq --verbose --source "${INPUT}" --count-file "${COUNT_FILE}";
echo ">>> exited with code $?";