pagedet-scripts/make_dataset.sh

#!/bin/bash

usage () {
    echo "$0: [-n] <source directory> <target directory>"
    echo
    echo "Convert filenames to appropriate format for training"
	echo
	echo "-n	Dry run"
    exit 1
}

DRY_RUN=0

while getopts "nh" arg; do
  case $arg in
    h)
      usage
      ;;
    n)
      DRY_RUN=1
      ;;
  esac
done
shift $((OPTIND-1))

if [[ "$#" != "2" ]]
then
    echo "error: Missing directory parameters."
    echo "See -h for help"
    exit 1
fi

IN=$1;shift
OUT=$1;shift

mkdir -p $OUT/{images,labels}

for dir in $(for dir in $IN/*; do basename "$dir"; done)
do
    for file in $IN/$dir/*.orig.jpg
    do
		if [[ "$DRY_RUN" == "1" ]]
        then
            echo cp -a --reflink=auto $file $OUT/images/"$dir"_"$(basename $file|sed 's/\.orig\././')"
        else
            cp -a --reflink=auto $file $OUT/images/"$dir"_"$(basename $file|sed 's/\.orig\././')"
        fi
    done
    for file in $IN/$dir/*.coloured.jpg
    do
		if [[ "$DRY_RUN" == "1" ]]
        then
            echo cp -a --reflink=auto $file $OUT/labels/"$dir"_"$(basename $file|sed 's/\.coloured\././')"
        else
            cp -a --reflink=auto $file $OUT/labels/"$dir"_"$(basename $file|sed 's/\.coloured\././')"
        fi
    done
done