changemaker.lite/media-manager/scripts/fetch_content.sh

#!/bin/bash

# =================================================
# Fetch Content Script
# Downloads videos using yt-dlp with special handling
# for Redgifs user profiles (auto-organizes by orientation)
# =================================================

# Enable strict error handling
set -e
set -o pipefail  # Catch failures in piped commands

# Environment variables (set by job system)
FETCH_URLS="${FETCH_URLS:-}"
FETCH_MODE="${FETCH_MODE:-single}"
INBOX_PATH="${INBOX_PATH:-/media/inbox}"
GIFS_PATH="${GIFS_PATH:-/media/gifs}"

# Check dependencies
if ! command -v yt-dlp &> /dev/null; then
    echo "ERROR: yt-dlp is not installed"
    exit 1
fi

if ! command -v ffprobe &> /dev/null; then
    echo "ERROR: ffprobe is not installed"
    exit 1
fi

# Function to detect orientation using ffprobe
detect_orientation() {
    local file="$1"
    local width height

    width=$(ffprobe -v error -select_streams v:0 \
        -show_entries stream=width \
        -of default=noprint_wrappers=1:nokey=1 "$file" 2>/dev/null)
    height=$(ffprobe -v error -select_streams v:0 \
        -show_entries stream=height \
        -of default=noprint_wrappers=1:nokey=1 "$file" 2>/dev/null)

    if [[ -z "$width" || -z "$height" ]]; then
        echo "unknown"
        return
    fi

    if [[ $width -ge $height ]]; then
        echo "horizontal"
    else
        echo "vertical"
    fi
}

# Function to extract creator name from Redgifs profile URL
extract_redgifs_creator() {
    local url="$1"
    # Pattern: https://www.redgifs.com/users/username or /users/username/...
    if [[ "$url" =~ redgifs\.com/users/([^/\?]+) ]]; then
        echo "${BASH_REMATCH[1]}"
    else
        echo ""
    fi
}

# Function to check if URL is a Redgifs user profile
is_redgifs_profile() {
    local url="$1"
    [[ "$url" =~ redgifs\.com/users/ ]]
}

# Function to download and process a single URL
process_url() {
    local url="$1"
    local url_num="$2"
    local total_urls="$3"

    echo ""
    echo "========================================"
    echo "Processing URL [$url_num/$total_urls]"
    echo "========================================"
    echo "URL: $url"
    echo ""

    # Check if Redgifs user profile
    if is_redgifs_profile "$url"; then
        local creator
        creator=$(extract_redgifs_creator "$url")

        if [[ -z "$creator" ]]; then
            echo "ERROR: Could not extract creator name from Redgifs URL"
            return 1
        fi

        echo ">>> REDGIFS USER PROFILE DETECTED <<<"
        echo "    Creator: $creator"
        echo "    Destination: $INBOX_PATH/$creator/"
        echo "    (Files will be sorted by orientation)"
        echo ""

        # Create creator directory structure in inbox
        local creator_dir="$INBOX_PATH/$creator"
        mkdir -p "$creator_dir/horizontal"
        mkdir -p "$creator_dir/vertical"

        # Create temp directory for download
        local temp_dir
        temp_dir=$(mktemp -d)

        echo "-> Downloading to temp directory..."

        # Download all videos from user profile
        # Using best format, mp4 container preferred
        if ! yt-dlp \
            -f "bestvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best" \
            --merge-output-format mp4 \
            -o "$temp_dir/%(title)s.%(ext)s" \
            --restrict-filenames \
            --no-overwrites \
            "$url" 2>&1; then
            echo "ERROR: yt-dlp download failed"
            rm -rf "$temp_dir"
            return 1
        fi

        # Process downloaded files
        echo ""
        echo "-> Processing downloaded files..."

        local processed=0
        local h_count=0
        local v_count=0

        for file in "$temp_dir"/*.mp4; do
            [[ -f "$file" ]] || continue

            local filename
            filename=$(basename "$file")
            local orientation
            orientation=$(detect_orientation "$file")

            if [[ "$orientation" == "horizontal" ]]; then
                mv "$file" "$creator_dir/horizontal/$filename"
                echo "   [H] $filename -> horizontal/"
                h_count=$((h_count + 1))
            elif [[ "$orientation" == "vertical" ]]; then
                mv "$file" "$creator_dir/vertical/$filename"
                echo "   [V] $filename -> vertical/"
                v_count=$((v_count + 1))
            else
                # Unknown orientation - put in horizontal as default
                mv "$file" "$creator_dir/horizontal/$filename"
                echo "   [?] $filename -> horizontal/ (orientation unknown)"
                h_count=$((h_count + 1))
            fi
            processed=$((processed + 1))
        done

        # Cleanup temp directory
        rm -rf "$temp_dir"

        echo ""
        echo "Completed: $processed files ($h_count horizontal, $v_count vertical)"

    else
        # Standard download to inbox
        echo "-> Standard download to inbox"
        echo "   Destination: $INBOX_PATH/"
        echo ""

        # Download with best quality
        if ! yt-dlp \
            -f "bestvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best" \
            --merge-output-format mp4 \
            -o "$INBOX_PATH/%(title)s.%(ext)s" \
            --restrict-filenames \
            --no-overwrites \
            "$url" 2>&1; then
            echo "ERROR: yt-dlp download failed"
            return 1
        fi

        echo ""
        echo "Download complete"
    fi
}

# Main execution
echo "========================================="
echo "  Content Fetch Script"
echo "  Mode: $FETCH_MODE"
echo "========================================="

if [[ -z "$FETCH_URLS" ]]; then
    echo "ERROR: No URLs provided"
    exit 1
fi

# Parse URLs (newline-separated)
mapfile -t urls <<< "$FETCH_URLS"

# Filter empty lines and trim whitespace safely
# Using printf '%s' prevents command expansion unlike echo
valid_urls=()
for url in "${urls[@]}"; do
    # Safely trim whitespace without command expansion
    url=$(printf '%s' "$url" | tr -d '[:space:]')
    [[ -n "$url" ]] && valid_urls+=("$url")
done

total_urls=${#valid_urls[@]}

if [[ $total_urls -eq 0 ]]; then
    echo "ERROR: No valid URLs provided"
    exit 1
fi

echo "URLs to process: $total_urls"
echo ""

# Track results
success_count=0
fail_count=0

# Process each URL
for i in "${!valid_urls[@]}"; do
    url="${valid_urls[$i]}"
    url_num=$((i + 1))

    if process_url "$url" "$url_num" "$total_urls"; then
        success_count=$((success_count + 1))
    else
        fail_count=$((fail_count + 1))
        echo ">>> Failed to process: $url"
    fi
done

# Summary
echo ""
echo ""
echo "========================================="
echo "  Fetch Complete!"
echo "========================================="
echo "  Successful: $success_count"
echo "  Failed: $fail_count"
echo "  Total: $total_urls"
echo "========================================="

if [[ $fail_count -gt 0 ]]; then
    exit 1
fi