add qlv script

[WolfgangFahl]

https://wiki.bitplan.com/index.php/Wikidata_Import_2024-10-17#Using_qlv_script

#!/bin/bash
#
# QLever Wikidata Backup and Indexing Script
# Author: WF
# Date: 2024-10-17
# Version: 1.0
#
# Purpose:
# This script manages daily Wikidata backups and indexing using the QLever SPARQL engine
# in a rotating fashion across multiple disks. It is designed to be run once per day by cron.
#
# Key Functions:
# 1. Rotates between available disks (/hd/*) for daily Wikidata backups
# 2. Ensures disk space and RAM are not overutilized
# 3. Manages Docker containers for indexing and serving, preventing interference
# 4. Sets up multiple SPARQL endpoints with appropriate naming
#
# Color definitions
blue='\033[0;34m'
red='\033[0;31m'
green='\033[0;32m'
endColor='\033[0m'

# Function to display colored messages
color_msg() {
  local l_color="$1"
  local l_msg="$2"
  echo -e "${l_color}$l_msg${endColor}"
}

# Function to display errors
error() {
  local l_msg="$1"
  color_msg $red "Error:" 1>&2
  color_msg $red "\t$l_msg" 1>&2
  exit 1
}

# Function to display negative messages
negative() {
  local l_msg="$1"
  color_msg $red "❌:$l_msg"
}

# Function to display positive messages
positive() {
  local l_msg="$1"
  color_msg $green "✅:$l_msg"
}

# Function to display usage information
usage() {
  echo "Usage: $0 [OPTIONS]"
  echo "Options:"
  echo "  -h, --help             Show this help message"
  echo "  -c, --current          Show the disk currently used by QLever"
  echo "  -d, --debug            Enable debug output"
  echo "  -ir, --index-run       Run QLever wikidata indexing on today's disk"
  echo "  -p, --pull             Pull QLever Docker images"
  echo "  -qc, --qlever-control  setup qlever-control"
  echo "  -s, --space            Show free disk space"
  echo "  -t, --today            Show disk to be used today"
  echo "  -v, --version          Show version information"
  exit 1
}

#
# setup qlever control
#
setup_qlever_control() {
  # Check if /opt/qlever-control already exists
  if [ ! -d "/opt/qlever-control" ]; then
    color_msg $blue "Setting up QLever control in /opt/qlever-control..."
    
    # Create the directory with the right permissions
    sudo mkdir -p /opt/qlever-control
    sudo chown -R root:users /opt/qlever-control
    sudo chmod -R 775 /opt/qlever-control
    
    # Clone the qlever-control repository
    sudo git clone https://github.com/ad-freiburg/qlever-control.git /opt/qlever-control
    cd /opt/qlever-control
    
    # Checkout the correct branch
    sudo git checkout python-qlever
    
    # Install qlever-control with pip
    sudo pip install .
    
    positive "QLever control setup and installed successfully."
  else
    positive "QLever control is already set up in /opt/qlever-control."
  fi

  # Verify if the qlever command is installed
  if which qlever > /dev/null 2>&1; then
    positive "QLever is installed and available at $(which qlever)"
  else
    error "QLever installation failed or is not in the system PATH."
  fi
}


# Function to pull Docker images
pull_docker_images() {
  color_msg $blue "Pulling QLever Docker images..."
  
  if docker pull adfreiburg/qlever; then
    positive "Successfully pulled adfreiburg/qlever"
  else
    negative "Failed to pull adfreiburg/qlever"
  fi
  
  if docker pull adfreiburg/qlever-ui; then
    positive "Successfully pulled adfreiburg/qlever-ui"
  else
    negative "Failed to pull adfreiburg/qlever-ui"
  fi
}

#
# get the disk that is currently in use
#
check_current_disk() {
    # Get the current disk used by qlever.server.wikidata
    current_disk=$(docker inspect qlever.server.wikidata --format '{{ range .Mounts }}{{ .Source }}:{{ .Destination }}{{ printf "\n" }}{{ end }}' | grep "/hd/" | awk -F':' '{print $1}')
    
    # Display the current disk
    if [ -n "$current_disk" ]; then
        echo "$current_disk"
    else
        error "No disk found for qlever.server.wikidata"
    fi
}


#
# get the disk of the day
#
disk_of_the_day() {
    # Define the available disks
    disks=(/hd/alpha /hd/beta /hd/gamma /hd/delta)

    # Get the current day of the week as an index
    day=$(date +%u) # 1 (Monday) to 7 (Sunday)

    # Select the disk based on the day
    disk_index=$(( (day - 1) % ${#disks[@]} ))
    selected_disk=${disks[$disk_index]}

    # Return the selected disk
    echo "$selected_disk"
}

#
# show the available disk space
#
show_disk_space() {
    # Print the header with proper formatting
    printf "%-10s %-15s %10s %10s %4s\n" "Directory" "Device" "Available" "Total" "Type"

    for dir in /hd/*; do
        # Get the device mounted to this directory
        device=$(df $dir | tail -1 | awk '{print $1}')
        
        # Get available and total space using df
        available=$(df -h $dir | tail -1 | awk '{print $4}')
        total=$(df -h $dir | tail -1 | awk '{print $2}')
        
        # Determine if the device is SSD or HDD
        if [[ "$device" == *nvme* ]]; then
            type="SSD"
        else
            base_device=$(echo "$device" | sed 's/[0-9]*$//') # Strip partition number
            rotational=$(cat /sys/block/$(basename $base_device)/queue/rotational)
            if [ "$rotational" -eq 0 ]; then
                type="SSD"
            else
                type="HDD"
            fi
        fi
        
        # Print the results with proper formatting
        printf "%-10s %-15s %10s %10s %4s\n" "$(basename $dir)" "$device" "$available" "$total" "$type"
    done
}

# Function to prepare the directory for today's QLever indexing
prepare_dir() {
    local disk=$(disk_of_the_day)
    local isodate=$(date +%Y%m%d)
    dir="$disk/qlever/wikidata_$isodate"

    sudo mkdir -p "$dir"
    sudo chown -R $(whoami):$(whoami) "$dir"
    sudo chmod -R 775 "$dir"
    echo $dir
}

# Function to execute the QLever indexing inside a screen session with logging
execute_index() {
    local isodate=$(date +%Y%m%d)
    local dir="$1"
    local session="qlever_wikidata_$isodate"
    local scriptfile="$dir/qlever_index.sh"
    local logfile="$dir/screen.log"

    # Create the script that will be run inside the screen session
    cat <<EOF > "$scriptfile"
#!/bin/bash
cd $dir
exec > >(tee -a "$logfile") 2>&1
echo "Starting QLever indexing process at \$(date)"
qlever setup-config wikidata
qlever get-data
qlever index
qlever start
echo "QLever indexing process completed at \$(date)"
EOF
    chmod +x "$scriptfile"
    screen -dmS $session bash -c "$scriptfile"

    # Verify that the screen session started successfully
    if screen -ls | grep -q "$session"; then
        positive "Started screen session $session."
        positive "Logging to $logfile"
    else
        error "Failed to start screen session $session."
    fi
}

# run qlever wikidata indexing with screen on today's disk
create_and_run_index() {
    # Prepare the directory for today's QLever indexing and capture the directory
    dir=$(prepare_dir)
    positive "Created directory $dir"
    
    # Execute the QLever indexing inside a screen session with logging
    execute_index "$dir"
}

# Parse command line options
while [[ "$#" -gt 0 ]]; do
  case $1 in
    -h|--help) usage ;;
    -c|--current) check_current_disk; exit 0 ;;
    -d|--debug) DEBUG=1 ;;
    -ir|--index-run) create_and_run_index; exit 0 ;;
    -p|--pull) pull_docker_images; exit 0 ;;
    -qc|--qlever-control) setup_qlever_control; exit 0 ;;
    -s|--space) show_disk_space; exit 0 ;;
    -t|--today) disk_of_the_day; exit 0 ;;
    -v|--version) echo "Version 1.0"; exit 0 ;;
    *) error "Unknown parameter passed: $1" ;;
  esac
  shift
done

exit 0