memefast/bash/check_webm_alpha_frames.sh

#!/bin/bash

# Script to check all WebM files with alpha for empty start and end frames
# Uses libvpx-vp9 decoder to properly handle alpha channels

# Check if directory parameter is provided
if [ $# -eq 0 ]; then
    echo "Usage: sh check_webm_alpha_frames.sh <absolute_path_to_directory>"
    echo "Example: sh check_webm_alpha_frames.sh /Users/charlesteh/Desktop/mag-group-1/"
    exit 1
fi

# Get the target directory from parameter
target_dir="$1"

# Check if directory exists
if [ ! -d "$target_dir" ]; then
    echo "Error: Directory '$target_dir' does not exist"
    exit 1
fi

echo "Checking all WebM files in '$target_dir' for empty start and end frames..."

# Check if ffmpeg is installed
if ! command -v ffmpeg &> /dev/null; then
    echo "Error: ffmpeg is not installed or not in PATH"
    exit 1
fi

# Change to target directory
cd "$target_dir" || {
    echo "Error: Cannot access directory '$target_dir'"
    exit 1
}

# Count total WebM files
webm_count=$(ls *.webm 2>/dev/null | wc -l)

if [ $webm_count -eq 0 ]; then
    echo "No WebM files found in current directory"
    exit 0
fi

echo "Found $webm_count WebM file(s) to check"
echo ""

# Create temporary directory for frame extraction
temp_dir=$(mktemp -d)
trap "rm -rf $temp_dir" EXIT

# Function to check if a frame is empty (fully transparent)
check_frame_empty() {
    local frame_file="$1"
    local frame_name="$2"

    # Check if frame file exists
    if [ ! -f "$frame_file" ]; then
        echo "    ✗ $frame_name: Frame file not found"
        return 2  # unclear/error
    fi

    # Get file size - very small files are likely empty/transparent
    file_size=$(stat -f%z "$frame_file" 2>/dev/null || stat -c%s "$frame_file" 2>/dev/null)

    # If file is very small (less than 2KB), likely empty
    if [ -n "$file_size" ] && [ "$file_size" -lt 2048 ]; then
        echo "    ✗ $frame_name: EMPTY (tiny file: ${file_size} bytes)"
        return 0  # definitely empty
    fi

    # Use simpler alpha analysis: extract alpha channel and get histogram
    alpha_info=$(ffmpeg -i "$frame_file" -vf "extractplanes=a" -f null - 2>&1 | grep "mean\|avg")

    if [ -n "$alpha_info" ]; then
        # Look for mean value in the output
        alpha_mean=$(echo "$alpha_info" | grep -o "mean:[0-9.]*" | cut -d: -f2 | head -1)

        if [ -n "$alpha_mean" ]; then
            # Check if mean is very low (close to 0 = transparent)
            is_empty=$(awk "BEGIN {print ($alpha_mean < 2.0) ? 1 : 0}")

            if [ "$is_empty" = "1" ]; then
                echo "    ✗ $frame_name: EMPTY (alpha mean: $alpha_mean)"
                return 0  # definitely empty
            else
                echo "    ✓ $frame_name: Content detected (alpha mean: $alpha_mean)"
                return 1  # definitely has content
            fi
        fi
    fi

    # If file is reasonably large, assume it has content
    if [ "$file_size" -gt 50000 ]; then
        echo "    ✓ $frame_name: Content detected (large file: ${file_size} bytes)"
        return 1  # definitely has content
    else
        echo "    ? $frame_name: UNCLEAR - needs manual review (size: ${file_size} bytes)"
        return 2  # unclear/unsure
    fi
}

# Function to get video duration and frame count
get_video_info() {
    local video_file="$1"

    # Get duration in seconds
    duration=$(ffprobe -vcodec libvpx-vp9 -v quiet -show_entries format=duration -of csv=p=0 "$video_file" 2>/dev/null)

    # Get total frame count
    frame_count=$(ffprobe -vcodec libvpx-vp9 -v quiet -select_streams v:0 -count_frames -show_entries stream=nb_read_frames -csv=p=0 "$video_file" 2>/dev/null)

    echo "$duration $frame_count"
}

# Counter for progress
count=0
files_with_empty_frames=0
files_with_unclear_frames=0

# Arrays to track files with empty frames
files_with_empty_first=()
files_with_empty_last=()
files_with_both_empty=()

# Arrays to track files with unclear frames
files_with_unclear_first=()
files_with_unclear_last=()
files_with_both_unclear=()
files_with_mixed_issues=()

# Loop through all WebM files in current directory
for webm_file in *.webm; do
    # Check if file actually exists (in case glob doesn't match)
    if [ ! -f "$webm_file" ]; then
        continue
    fi

    # Increment counter
    count=$((count + 1))

    echo "[$count/$webm_count] Checking: $webm_file"

    # Test if file is readable by ffprobe
    if ! ffprobe -v quiet "$webm_file" >/dev/null 2>&1; then
        echo "    ✗ Error: File cannot be read by ffprobe"
        echo ""
        continue
    fi

    # Detect codec type (VP8 or VP9) to use correct decoder - from working script
    CODEC_INFO=$(ffprobe -v quiet -select_streams v:0 -show_entries stream=codec_name -of csv=p=0 "$webm_file" 2>/dev/null)

    # Set appropriate decoder based on codec
    if [ "$CODEC_INFO" = "vp9" ]; then
        DECODER="libvpx-vp9"
        echo "    → Using VP9 decoder"
    else
        DECODER="libvpx"  # Default to VP8 decoder
        echo "    → Using VP8 decoder"
    fi

    # Get video info - use direct commands
    duration=$(ffprobe -v quiet -show_entries format=duration -of csv=p=0 "$webm_file" 2>/dev/null)
    frame_count=$(ffprobe -v quiet -select_streams v:0 -count_frames -show_entries stream=nb_read_frames -of csv=p=0 "$webm_file" 2>/dev/null)

    if [ -z "$duration" ] || [ "$duration" = "N/A" ]; then
        echo "    ✗ Error: Could not get video duration"
        echo ""
        continue
    fi

    echo "    Duration: ${duration}s, Frames: $frame_count"

    # Extract first frame (at 0 seconds) - using proper decoder syntax from working script
    first_frame="$temp_dir/first_frame_${count}.png"
    ffmpeg -c:v "$DECODER" -i "$webm_file" -vf "select=eq(n\,0)" -frames:v 1 -y "$first_frame" -v quiet 2>/dev/null

    # Extract last frame - using proper decoder syntax
    last_frame="$temp_dir/last_frame_${count}.png"
    # Calculate time for last frame (duration - small offset to ensure we get the last frame)
    last_frame_time=$(awk "BEGIN {printf \"%.3f\", $duration - 0.1}")
    # Ensure we don't go before 0
    if [ $(awk "BEGIN {print ($last_frame_time < 0) ? 1 : 0}") = "1" ]; then
        last_frame_time="0"
    fi
    ffmpeg -c:v "$DECODER" -ss "$last_frame_time" -i "$webm_file" -frames:v 1 -y "$last_frame" -v quiet 2>/dev/null

    # Check if frame extraction was successful
    if [ ! -f "$first_frame" ] || [ ! -f "$last_frame" ]; then
        echo "    ✗ Error: Could not extract frames"
        echo ""
        continue
    fi

    # Check if frames are empty
    has_empty_frames=false
    has_unclear_frames=false
    has_empty_first=false
    has_empty_last=false
    has_unclear_first=false
    has_unclear_last=false

    check_frame_empty "$first_frame" "First frame"
    first_result=$?
    if [ $first_result -eq 0 ]; then
        has_empty_first=true
        has_empty_frames=true
    elif [ $first_result -eq 2 ]; then
        has_unclear_first=true
        has_unclear_frames=true
    fi

    check_frame_empty "$last_frame" "Last frame"
    last_result=$?
    if [ $last_result -eq 0 ]; then
        has_empty_last=true
        has_empty_frames=true
    elif [ $last_result -eq 2 ]; then
        has_unclear_last=true
        has_unclear_frames=true
    fi

    # Add to tracking arrays
    if [ "$has_empty_first" = true ] && [ "$has_empty_last" = true ]; then
        files_with_both_empty+=("$webm_file")
    elif [ "$has_unclear_first" = true ] && [ "$has_unclear_last" = true ]; then
        files_with_both_unclear+=("$webm_file")
    elif ([ "$has_empty_first" = true ] && [ "$has_unclear_last" = true ]) || ([ "$has_unclear_first" = true ] && [ "$has_empty_last" = true ]); then
        files_with_mixed_issues+=("$webm_file")
    elif [ "$has_empty_first" = true ]; then
        files_with_empty_first+=("$webm_file")
    elif [ "$has_empty_last" = true ]; then
        files_with_empty_last+=("$webm_file")
    elif [ "$has_unclear_first" = true ]; then
        files_with_unclear_first+=("$webm_file")
    elif [ "$has_unclear_last" = true ]; then
        files_with_unclear_last+=("$webm_file")
    fi

    if [ "$has_empty_frames" = true ]; then
        files_with_empty_frames=$((files_with_empty_frames + 1))
    fi

    if [ "$has_unclear_frames" = true ]; then
        files_with_unclear_frames=$((files_with_unclear_frames + 1))
    fi

    # Clean up temporary frames
    rm -f "$first_frame" "$last_frame"

    if [ "$has_empty_frames" = true ]; then
        echo "    ⚠️  WARNING: This file has empty frames!"
    elif [ "$has_unclear_frames" = true ]; then
        echo "    ❓ NOTE: This file has unclear frames that need manual review"
    fi

    echo ""
done

echo "==================================="
echo "Analysis complete!"
echo "Total files checked: $count"
echo "Files with empty start/end frames: $files_with_empty_frames"
echo "Files with unclear frames (need review): $files_with_unclear_frames"

total_issues=$((files_with_empty_frames + files_with_unclear_frames))

if [ $total_issues -gt 0 ]; then
    echo ""
    echo "==================================="
    echo "DETAILED SUMMARY - FILES NEEDING ATTENTION:"
    echo "==================================="

    # Files with both first and last frames empty
    if [ ${#files_with_both_empty[@]} -gt 0 ]; then
        echo ""
        echo "🚨 Files with BOTH first AND last frames EMPTY (${#files_with_both_empty[@]} files):"
        for file in "${files_with_both_empty[@]}"; do
            echo "   • $(basename "$file")"
        done
    fi

    # Files with only first frame empty
    if [ ${#files_with_empty_first[@]} -gt 0 ]; then
        echo ""
        echo "⚠️  Files with FIRST frame EMPTY (${#files_with_empty_first[@]} files):"
        for file in "${files_with_empty_first[@]}"; do
            echo "   • $(basename "$file")"
        done
    fi

    # Files with only last frame empty
    if [ ${#files_with_empty_last[@]} -gt 0 ]; then
        echo ""
        echo "⚠️  Files with LAST frame EMPTY (${#files_with_empty_last[@]} files):"
        for file in "${files_with_empty_last[@]}"; do
            echo "   • $(basename "$file")"
        done
    fi

    # Files with both first and last frames unclear
    if [ ${#files_with_both_unclear[@]} -gt 0 ]; then
        echo ""
        echo "❓ Files with BOTH first AND last frames UNCLEAR (${#files_with_both_unclear[@]} files):"
        for file in "${files_with_both_unclear[@]}"; do
            echo "   • $(basename "$file")"
        done
    fi

    # Files with only first frame unclear
    if [ ${#files_with_unclear_first[@]} -gt 0 ]; then
        echo ""
        echo "❓ Files with FIRST frame UNCLEAR (${#files_with_unclear_first[@]} files):"
        for file in "${files_with_unclear_first[@]}"; do
            echo "   • $(basename "$file")"
        done
    fi

    # Files with only last frame unclear
    if [ ${#files_with_unclear_last[@]} -gt 0 ]; then
        echo ""
        echo "❓ Files with LAST frame UNCLEAR (${#files_with_unclear_last[@]} files):"
        for file in "${files_with_unclear_last[@]}"; do
            echo "   • $(basename "$file")"
        done
    fi

    # Files with mixed issues
    if [ ${#files_with_mixed_issues[@]} -gt 0 ]; then
        echo ""
        echo "🔀 Files with MIXED issues (empty + unclear frames) (${#files_with_mixed_issues[@]} files):"
        for file in "${files_with_mixed_issues[@]}"; do
            echo "   • $(basename "$file")"
        done
    fi

    echo ""
    echo "==================================="
    echo "RECOMMENDATIONS:"
    echo "🚨 EMPTY frames: Definitely need trimming with ffmpeg"
    echo "❓ UNCLEAR frames: Manually review these files"
    echo "• Files with both empty first/last frames may be completely problematic"
    echo "• Consider using ffmpeg to trim empty frames where confirmed"
else
    echo ""
    echo "✅ All files appear to have clear content in start and end frames."
fi