pdf-concat

#!/bin/bash

# Script: pdf-concat
# Description: Combines multiple PDF files into a single PDF
# Author: Konnor Kooi
# Date: 2024-11-11

# Colors for output
GREEN='\033[0;32m'
RED='\033[0;31m'
BLUE='\033[0;34m'
CYAN='\033[0;36m'
BOLD='\033[1m'
NC='\033[0m' # No Color

# Function to print status messages
print_status() {
    echo -e "${CYAN}=>${NC} $1"
}

# Function to print success messages
print_success() {
    echo -e "${GREEN}✓${NC} $1"
}

# Function to print error messages
print_error() {
    echo -e "${RED}✗${NC} $1" >&2
}

# Function to show help message
show_help() {
    echo -e "
    ${GREEN}PDF Concatenator${NC}
    A tool to combine multiple PDF files into a single PDF

    ${BOLD}USAGE${NC}
      $(basename "$0") [options] [files...]

    ${BOLD}OPTIONS${NC}
      ${CYAN}-o${NC}, ${CYAN}--output${NC}    NAME      Output filename (default: combined.pdf)
      ${CYAN}-d${NC}, ${CYAN}--debug${NC}              Enable debug output
      ${CYAN}-h${NC}, ${CYAN}--help${NC}               Show this help message

    ${BOLD}EXAMPLES${NC}
      • Combine all PDFs in current directory:
        $(basename "$0") .
      
      • Combine specific PDF files:
        $(basename "$0") file1.pdf file2.pdf file3.pdf
      
      • Combine with custom output name:
        $(basename "$0") -o merged.pdf file1.pdf file2.pdf
      
      • Combine all PDFs in specific directory:
        $(basename "$0") /path/to/directory

    ${BOLD}NOTES${NC}
      • Requires pdftk to be installed
      • Files are combined in alphabetical order when using directory mode
      • Output file is created in the current directory"
}

# Default values
OUTPUT_FILE="combined.pdf"
DEBUG=false
PDF_FILES=()
DIRECTORY_MODE=false

# Check if pdftk is installed
if ! command -v pdftk &> /dev/null; then
    print_error "pdftk is not installed. Please install it first:"
    echo -e "${BLUE}For Debian/Ubuntu:${NC}"
    echo "    sudo apt-get install pdftk"
    exit 1
fi

# Function to process directory
process_directory() {
    local dir="$1"
    if [ ! -d "$dir" ]; then
        print_error "Directory does not exist: $dir"
        exit 1
    fi

    local dir_path="$(cd "$dir" && pwd)"
    [ "$DEBUG" = true ] && print_status "Scanning directory: $dir_path"

    # Get all PDF files in directory
    while IFS= read -r -d '' file; do
        PDF_FILES+=("$file")
    done < <(find "$dir_path" -maxdepth 1 -type f -name "*.pdf" -print0 | sort -z)

    if [ ${#PDF_FILES[@]} -eq 0 ]; then
        print_error "No PDF files found in directory: $dir_path"
        exit 1
    fi
}

# Parse command line arguments
while [[ $# -gt 0 ]]; do
    case $1 in
        -o|--output)
            OUTPUT_FILE="$2"
            # Add .pdf extension if not present
            [[ "$OUTPUT_FILE" != *.pdf ]] && OUTPUT_FILE="${OUTPUT_FILE}.pdf"
            shift 2
            ;;
        -d|--debug)
            DEBUG=true
            shift
            ;;
        -h|--help)
            show_help
            exit 0
            ;;
        *)
            # Check if it's a directory
            if [ -d "$1" ]; then
                DIRECTORY_MODE=true
                process_directory "$1"
                shift
            # Check if it's a PDF file
            elif [[ "$1" == *.pdf ]]; then
                if [ -f "$1" ]; then
                    PDF_FILES+=("$1")
                else
                    print_error "File does not exist: $1"
                    exit 1
                fi
                shift
            else
                print_error "Invalid input: $1"
                show_help
                exit 1
            fi
            ;;
    esac
done

# Check if we have any files to process
if [ ${#PDF_FILES[@]} -eq 0 ]; then
    # If no files specified, try current directory
    if [ "$DIRECTORY_MODE" = false ]; then
        process_directory "."
    else
        print_error "No PDF files specified"
        show_help
        exit 1
    fi
fi

# Check if output file exists
if [ -f "$OUTPUT_FILE" ]; then
    print_error "Output file already exists: $OUTPUT_FILE"
    exit 1
fi

# Debug output
if [ "$DEBUG" = true ]; then
    print_status "Files to be combined:"
    for file in "${PDF_FILES[@]}"; do
        echo "  - $(basename "$file")"
    done
    echo
fi

# Combine PDFs
print_status "Combining ${#PDF_FILES[@]} PDF files..."
if pdftk "${PDF_FILES[@]}" cat output "$OUTPUT_FILE" 2>/dev/null; then
    print_success "PDFs combined successfully into: $OUTPUT_FILE"
    print_status "Total files combined: ${#PDF_FILES[@]}"
else
    print_error "Failed to combine PDFs"
    [ -f "$OUTPUT_FILE" ] && rm "$OUTPUT_FILE"
    exit 1
fi