#!/bin/bash
################################################################################
# SUBTITLE SCANNER UTILITY
################################################################################
# Scans a directory tree and reports which files have subtitle streams
# Useful for:
#   - Planning which files need processing
#   - Verifying results after re-encoding
#   - Estimating workload
################################################################################

set -euo pipefail

SCAN_DIR="${1:-/mnt/user/movies}"
OUTPUT_MODE="${OUTPUT_MODE:-summary}"  # summary, detailed, or csv

################################################################################
# Functions
################################################################################

check_file() {
    local file="$1"
    local subtitle_count=$(ffprobe -v error -select_streams s -show_entries stream=codec_name -of default=noprint_wrappers=1:nokey=1 "$file" 2>/dev/null | wc -l)
    echo "$subtitle_count"
}

human_readable_size() {
    local size=$1
    if [[ $size -lt 1024 ]]; then
        echo "${size}B"
    elif [[ $size -lt 1048576 ]]; then
        echo "$(( size / 1024 ))KB"
    elif [[ $size -lt 1073741824 ]]; then
        echo "$(( size / 1048576 ))MB"
    else
        echo "$(( size / 1073741824 ))GB"
    fi
}

################################################################################
# Main
################################################################################

if [[ ! -d "$SCAN_DIR" ]]; then
    echo "ERROR: Directory not found: $SCAN_DIR" >&2
    exit 1
fi

echo "=========================================="
echo "Scanning: $SCAN_DIR"
echo "=========================================="
echo ""

total_files=0
files_with_subs=0
files_without_subs=0
total_size=0
size_with_subs=0

# CSV header if needed
if [[ "$OUTPUT_MODE" == "csv" ]]; then
    echo "filepath,subtitle_count,size_bytes"
fi

while IFS= read -r file; do
    ((total_files++)) || true
    
    subtitle_count=$(check_file "$file")
    file_size=$(stat -c%s "$file" 2>/dev/null || stat -f%z "$file" 2>/dev/null)
    total_size=$((total_size + file_size))
    
    if [[ $subtitle_count -gt 0 ]]; then
        ((files_with_subs++)) || true
        size_with_subs=$((size_with_subs + file_size))
        
        case "$OUTPUT_MODE" in
            detailed)
                echo "[HAS SUBS] $file ($subtitle_count subtitle stream(s), $(human_readable_size $file_size))"
                ;;
            csv)
                echo "$file,$subtitle_count,$file_size"
                ;;
        esac
    else
        ((files_without_subs++)) || true
        
        case "$OUTPUT_MODE" in
            detailed)
                echo "[NO SUBS]  $file ($(human_readable_size $file_size))"
                ;;
            csv)
                echo "$file,0,$file_size"
                ;;
        esac
    fi
    
    # Progress indicator for summary mode
    if [[ "$OUTPUT_MODE" == "summary" ]] && [[ $((total_files % 10)) -eq 0 ]]; then
        echo -ne "Scanned $total_files files...\r"
    fi
    
done < <(find "$SCAN_DIR" -type f \( -name "*.mkv" -o -name "*.mp4" -o -name "*.avi" -o -name "*.m4v" \) | sort)

# Summary
if [[ "$OUTPUT_MODE" != "csv" ]]; then
    echo ""
    echo "=========================================="
    echo "SUMMARY"
    echo "=========================================="
    echo "Total files scanned:      $total_files"
    echo "Files with subtitles:     $files_with_subs"
    echo "Files without subtitles:  $files_without_subs"
    echo ""
    echo "Total library size:       $(human_readable_size $total_size)"
    echo "Size needing processing:  $(human_readable_size $size_with_subs)"
    
    if [[ $files_with_subs -gt 0 ]]; then
        avg_size=$((size_with_subs / files_with_subs))
        echo "Average file size:        $(human_readable_size $avg_size)"
        
        # Rough time estimate (assuming 2-hour movies at 1 fps average)
        estimated_hours=$((files_with_subs * 2))
        echo ""
        echo "Estimated processing time: ~$estimated_hours hours (at 1 fps average)"
        echo "  (Actual time varies greatly based on hardware and settings)"
    fi
fi

echo ""
echo "=========================================="

# Usage examples
if [[ "$OUTPUT_MODE" == "summary" ]]; then
    echo ""
    echo "TIP: For more detail, run with:"
    echo "  OUTPUT_MODE=detailed $0 $SCAN_DIR"
    echo ""
    echo "Or generate a CSV:"
    echo "  OUTPUT_MODE=csv $0 $SCAN_DIR > subtitles.csv"
fi