dedup/get_dups.sh

#!/bin/bash
# Copyright (c) 2025 by Robert Strutts
# License: MIT
source myenv/bin/activate

# Check if the first argument is a directory
if [ ! -d "$1" ]; then
    echo "Error: '$1' is not a directory."
    exit 1
fi

# Get list of images
shopt -s nullglob
pushd "$1" || exit 1
images=(*.jpg *.png)
popd || exit 1

# Check if we found any images
if [ ${#images[@]} -eq 0 ]; then
    echo "No images found."
    exit 1
fi

# Outer loop
for ((i = 0; i < ${#images[@]}; i++)); do
    outer_image="${images[$i]}"

    # Inner loop (only later images to avoid double-checks)
    for ((j = i + 1; j < ${#images[@]}; j++)); do
        inner_image="${images[$j]}"
		
		echo -e "\nCompairing files: $outer_image TO $inner_image"
        python3 dedup.py "$1/$outer_image" "$1/$inner_image" "$2"
        exit_code=$?
        if [ $exit_code -eq 1 ]; then
            echo "$1/$outer_image # $inner_image" >> dups.txt
            break   # No need to check more once found duplicate
        fi
        if [ $exit_code -eq 2 ]; then
            echo "$1/$outer_image # $inner_image" >> alike.txt
            break   # No need to check more once found close match to duplicate
        fi
        if [ $exit_code -eq 4 ]; then
            break   # Skip Invaild inner Image
        fi
        if [ $exit_code -eq 5 ]; then
            echo "$1/$outer_image # $inner_image" >> sameGPS.txt
            break   # No need to check more once found matching GPS image
        fi
        if [ $exit_code -eq 6 ]; then
            echo "$1/$outer_image # $inner_image" >> sameGPSmile.txt
            break   # No need to check more once found matching GPS image
        fi
        if [ $exit_code -eq 8 ]; then
            echo "$1/$outer_image" >> invalid.txt
            break   # No need to check more once found bad image
        fi
        if [ $exit_code -eq 9 ]; then
            echo "$1/$outer_image" >> size.txt
            break   # No need to check more once found image too Small or Large
        fi
        
    done
done