#!/bin/bash
#
# Historical Attack Log Analyzer
# Scans past Apache/Nginx logs for attack patterns using ET Open signatures
#
# Usage: bash analyze-historical-attacks.sh [options]
#
# Options:
#   -d DAYS    Analyze logs from last N days (default: 7)
#   -l LOGFILE Analyze specific log file
#   -o OUTPUT  Output report file (default: /tmp/attack-analysis-TIMESTAMP.txt)
#   -t THRESHOLD  Minimum threat score to report (default: 50)
#   -v         Verbose mode (show all attacks)
#   -h         Show help

# Get script directory
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/.."

# Source required libraries
source "$SCRIPT_DIR/lib/attack-signatures.sh" 2>/dev/null || {
    echo "ERROR: attack-signatures.sh not found"
    exit 1
}
source "$SCRIPT_DIR/lib/http-attack-analyzer.sh" 2>/dev/null || {
    echo "ERROR: http-attack-analyzer.sh not found"
    exit 1
}

# Try to source IP reputation library (optional)
source "$SCRIPT_DIR/lib/ip-reputation.sh" 2>/dev/null

# Colors
RED='\033[0;31m'
YELLOW='\033[1;33m'
GREEN='\033[0;32m'
BLUE='\033[0;34m'
CYAN='\033[0;36m'
BOLD='\033[1m'
NC='\033[0m'

# Default options
DAYS=7
LOG_FILE=""
OUTPUT_FILE="/tmp/attack-analysis-$(date +%Y%m%d_%H%M%S).txt"
THRESHOLD=50
VERBOSE=0

# Parse command line arguments
while getopts "d:l:o:t:vh" opt; do
    case $opt in
        d) DAYS="$OPTARG" ;;
        l) LOG_FILE="$OPTARG" ;;
        o) OUTPUT_FILE="$OPTARG" ;;
        t) THRESHOLD="$OPTARG" ;;
        v) VERBOSE=1 ;;
        h)
            cat << EOF
Historical Attack Log Analyzer
Scans past Apache/Nginx logs for attack patterns using ET Open signatures

Usage: $0 [options]

Options:
  -d DAYS     Analyze logs from last N days (default: 7)
  -l LOGFILE  Analyze specific log file
  -o OUTPUT   Output report file (default: /tmp/attack-analysis-TIMESTAMP.txt)
  -t THRESHOLD Minimum threat score to report (default: 50)
  -v          Verbose mode (show all attacks)
  -h          Show this help

Examples:
  # Analyze last 7 days
  $0

  # Analyze last 30 days
  $0 -d 30

  # Analyze specific log file
  $0 -l /var/log/apache2/access.log

  # Show all attacks (including low severity)
  $0 -t 0 -v

  # Save report to custom location
  $0 -o /root/attack-report.txt
EOF
            exit 0
            ;;
        \?)
            echo "Invalid option: -$OPTARG" >&2
            exit 1
            ;;
    esac
done

echo "================================================================================
"
echo -e "${CYAN}${BOLD}Historical Attack Log Analyzer${NC}"
echo "Powered by Emerging Threats Open Ruleset"
echo "================================================================================
"

# Find log files to analyze
LOG_FILES=()

if [ -n "$LOG_FILE" ]; then
    # Specific log file provided
    if [ ! -f "$LOG_FILE" ]; then
        echo -e "${RED}ERROR: Log file not found: $LOG_FILE${NC}"
        exit 1
    fi
    LOG_FILES=("$LOG_FILE")
    echo -e "${GREEN}✓${NC} Analyzing specific file: $LOG_FILE"
else
    # Auto-detect log files
    echo -e "${BLUE}[*]${NC} Searching for Apache/Nginx log files..."

    # Common log locations
    SEARCH_PATHS=(
        "/var/log/apache2"
        "/var/log/httpd"
        "/usr/local/apache/logs"
        "/var/log/nginx"
        "/usr/local/apache/domlogs"
    )

    for path in "${SEARCH_PATHS[@]}"; do
        if [ -d "$path" ]; then
            # Find access logs modified in last N days
            while IFS= read -r log; do
                LOG_FILES+=("$log")
            done < <(find "$path" -type f \( -name "access*.log*" -o -name "access_log*" -o -name "*.com" -o -name "*.net" -o -name "*.org" \) -mtime -"$DAYS" 2>/dev/null)
        fi
    done

    if [ ${#LOG_FILES[@]} -eq 0 ]; then
        echo -e "${RED}ERROR: No log files found in last $DAYS days${NC}"
        exit 1
    fi

    echo -e "${GREEN}✓${NC} Found ${#LOG_FILES[@]} log files from last $DAYS days"
fi

# Initialize counters
TOTAL_LINES=0
TOTAL_ATTACKS=0
CRITICAL_ATTACKS=0
HIGH_ATTACKS=0
MEDIUM_ATTACKS=0

declare -A ATTACK_TYPES
declare -A TOP_ATTACKERS
declare -A SIGNATURE_HITS
declare -A IP_ATTACK_DETAILS  # Store detailed attack info per IP
declare -A IP_ATTACK_COUNT    # Count attacks per IP
declare -A IP_SAMPLE_URLS     # Sample URLs per IP

# Progress indicator
show_progress() {
count=$1
total=$2
percent=$((count * 100 / total))
    echo -ne "\r${BLUE}[*]${NC} Processing: $count/$total lines ($percent%) "
}

# Start analysis
echo ""
echo -e "${BLUE}[*]${NC} Starting analysis (Threshold: $THRESHOLD)..."
echo ""

{
    # Write report header
    echo "================================================================================
"
    echo "HISTORICAL ATTACK ANALYSIS REPORT"
    echo "Generated: $(date)"
    echo "Period: Last $DAYS days"
    echo "Threshold: $THRESHOLD"
    echo "================================================================================
"
    echo ""

    # Analyze each log file
    for log_file in "${LOG_FILES[@]}"; do
        echo "[*] Analyzing: $log_file"

        # Handle compressed logs
        if [[ "$log_file" =~ \.gz$ ]]; then
            CAT_CMD="zcat"
        elif [[ "$log_file" =~ \.bz2$ ]]; then
            CAT_CMD="bzcat"
        else
            CAT_CMD="cat"
        fi

file_attacks=0
line_count=0

        while IFS= read -r line; do
            line_count=$((line_count + 1))
            TOTAL_LINES=$((TOTAL_LINES + 1))

            # Show progress every 1000 lines
            if [ $((line_count % 1000)) -eq 0 ]; then
                show_progress "$TOTAL_LINES" "unknown"
            fi

            # Analyze line
result=$(analyze_http_log_line "$line" 2>/dev/null)
threat_score="${result%%||*}"

            if [ "$threat_score" -ge "$THRESHOLD" ]; then
temp="${result#*||}"
attack_types="${temp%%||*}"
                temp="${temp#*||}"
signatures="${temp%%||*}"
                temp="${temp#*||}"
ip="${temp%%||*}"
uri="${temp#*||}"

                # Count attacks
                TOTAL_ATTACKS=$((TOTAL_ATTACKS + 1))
                file_attacks=$((file_attacks + 1))

                # Categorize by severity
                if [ "$threat_score" -ge 85 ]; then
                    CRITICAL_ATTACKS=$((CRITICAL_ATTACKS + 1))
                elif [ "$threat_score" -ge 70 ]; then
                    HIGH_ATTACKS=$((HIGH_ATTACKS + 1))
                elif [ "$threat_score" -ge 50 ]; then
                    MEDIUM_ATTACKS=$((MEDIUM_ATTACKS + 1))
                fi

                # Track attack types
                IFS=',' read -ra types <<< "$attack_types"
                for type in "${types[@]}"; do
                    ATTACK_TYPES["$type"]=$((${ATTACK_TYPES[$type]:-0} + 1))
                done

                # Track top attackers (cumulative score)
                TOP_ATTACKERS["$ip"]=$((${TOP_ATTACKERS[$ip]:-0} + threat_score))

                # Track attack count per IP
                IP_ATTACK_COUNT["$ip"]=$((${IP_ATTACK_COUNT[$ip]:-0} + 1))

                # Store attack type details per IP
                current_types="${IP_ATTACK_DETAILS[$ip]}"
                if [ -z "$current_types" ]; then
                    IP_ATTACK_DETAILS["$ip"]="$attack_types"
                else
                    IP_ATTACK_DETAILS["$ip"]="$current_types,$attack_types"
                fi

                # Store sample URL (keep first 3)
                current_urls="${IP_SAMPLE_URLS[$ip]}"
                if [ -z "$current_urls" ]; then
                    # First URL
                    IP_SAMPLE_URLS["$ip"]="${uri:0:100}"
                else
                    # Count existing URLs by counting delimiters + 1
                    url_count=$(echo "$current_urls" | grep -o "||" | wc -l)
                    url_count=$((url_count + 1))
                    if [ "$url_count" -lt 3 ]; then
                        IP_SAMPLE_URLS["$ip"]="$current_urls||${uri:0:100}"
                    fi
                fi

                # Track signatures
                IFS=',' read -ra sigs <<< "$signatures"
                for sig in "${sigs[@]}"; do
                    SIGNATURE_HITS["$sig"]=$((${SIGNATURE_HITS[$sig]:-0} + 1))
                done
            fi
        done < <($CAT_CMD "$log_file" 2>/dev/null)

        echo "  → Found $file_attacks attacks" >> "$OUTPUT_FILE"
    done

    echo "" >> "$OUTPUT_FILE"
    echo "================================================================================
" >> "$OUTPUT_FILE"
    echo "ATTACKING IPs - DETAILED BREAKDOWN" >> "$OUTPUT_FILE"
    echo "================================================================================
" >> "$OUTPUT_FILE"
    echo "" >> "$OUTPUT_FILE"

    # Sort IPs by cumulative threat score and display
    # Create sorted list first to avoid subshell issues
    sorted_ips=$(for ip in "${!TOP_ATTACKERS[@]}"; do
        echo "${TOP_ATTACKERS[$ip]}:$ip"
    done | sort -t: -k1 -nr | head -50)

    ip_count=0
    while IFS=: read -r cumulative_score ip; do
        ip_count=$((ip_count + 1))

        attack_count="${IP_ATTACK_COUNT[$ip]:-0}"
        all_attack_types="${IP_ATTACK_DETAILS[$ip]}"
        sample_urls="${IP_SAMPLE_URLS[$ip]}"

        # Count occurrences of each attack type
        declare -A type_counts
        IFS=',' read -ra attacks <<< "$all_attack_types"
        for attack in "${attacks[@]}"; do
            [ -n "$attack" ] && type_counts["$attack"]=$((${type_counts[$attack]:-0} + 1))
        done

        # Format attack summary
        attack_summary=""
        for type in "${!type_counts[@]}"; do
            if [ -z "$attack_summary" ]; then
                attack_summary="$type(${type_counts[$type]})"
            else
                attack_summary="$attack_summary, $type(${type_counts[$type]})"
            fi
        done
        unset type_counts

        # Determine threat level
        avg_score=$((cumulative_score / attack_count))
        if [ "$avg_score" -ge 85 ]; then
            level="CRITICAL"
        elif [ "$avg_score" -ge 70 ]; then
            level="HIGH"
        else
            level="MEDIUM"
        fi

        # Print IP summary
        echo "[$ip_count] $ip" >> "$OUTPUT_FILE"
        printf "    Attacks: %d | Avg Score: %d | Threat Level: %s\n" "$attack_count" "$avg_score" "$level" >> "$OUTPUT_FILE"
        echo "    Attack Types: $attack_summary" >> "$OUTPUT_FILE"

        # Get reputation (if available)
        if type get_threat_intelligence &>/dev/null; then
            threat_intel=$(get_threat_intelligence "$ip" 2>/dev/null)
            if [ -n "$threat_intel" ]; then
                IFS='|' read -r abuse_conf abuse_rpts country isp geo timing whitelisted <<< "$threat_intel"
                if [ "${abuse_conf:-0}" -gt 0 ]; then
                    printf "    Reputation: AbuseIPDB %d%% confidence (%d reports) | %s\n" "${abuse_conf:-0}" "${abuse_rpts:-0}" "${country:-Unknown}" >> "$OUTPUT_FILE"
                fi
            fi
        fi

        # Show sample URLs
        if [ -n "$sample_urls" ]; then
            echo "    Sample Targets:" >> "$OUTPUT_FILE"
            IFS='||' read -ra urls <<< "$sample_urls"
            for url in "${urls[@]}"; do
                echo "      - $url" >> "$OUTPUT_FILE"
            done
        fi

        echo "" >> "$OUTPUT_FILE"
    done <<< "$sorted_ips"

    echo "================================================================================
" >> "$OUTPUT_FILE"
    echo "SUMMARY STATISTICS" >> "$OUTPUT_FILE"
    echo "================================================================================
" >> "$OUTPUT_FILE"
    echo "" >> "$OUTPUT_FILE"
    echo "Total lines processed: $TOTAL_LINES" >> "$OUTPUT_FILE"
    echo "Total attacks detected: $TOTAL_ATTACKS" >> "$OUTPUT_FILE"
    echo "Unique attacking IPs: ${#TOP_ATTACKERS[@]}" >> "$OUTPUT_FILE"
    echo "" >> "$OUTPUT_FILE"
    echo "Attack Severity:" >> "$OUTPUT_FILE"
    echo "  - Critical (≥85): $CRITICAL_ATTACKS" >> "$OUTPUT_FILE"
    echo "  - High (70-84): $HIGH_ATTACKS" >> "$OUTPUT_FILE"
    echo "  - Medium (50-69): $MEDIUM_ATTACKS" >> "$OUTPUT_FILE"
    echo "" >> "$OUTPUT_FILE"

    # Top Attack Types
    echo "Top Attack Types:" >> "$OUTPUT_FILE"
    for type in "${!ATTACK_TYPES[@]}"; do
        echo "$type:${ATTACK_TYPES[$type]}"
    done | sort -t: -k2 -nr | head -10 | while IFS=: read -r type count; do
        printf "  %-20s %5d attacks\n" "$type" "$count" >> "$OUTPUT_FILE"
    done
    echo "" >> "$OUTPUT_FILE"

    echo "================================================================================
" >> "$OUTPUT_FILE"
    echo "END OF REPORT" >> "$OUTPUT_FILE"
    echo "================================================================================
" >> "$OUTPUT_FILE"

} > "$OUTPUT_FILE"

# Clear progress line
echo -ne "\r\033[K"

# Display summary to terminal
echo ""
echo -e "${GREEN}✓${NC} Analysis complete!"
echo ""
echo "Summary:"
echo "  Lines processed:    $TOTAL_LINES"
echo "  Attacks detected:   $TOTAL_ATTACKS"
echo "    - Critical (≥85): $CRITICAL_ATTACKS"
echo "    - High (70-84):   $HIGH_ATTACKS"
echo "    - Medium (50-69): $MEDIUM_ATTACKS"
echo ""
echo -e "${GREEN}✓${NC} Full report saved to: $OUTPUT_FILE"
echo ""

# Offer to view report
read -p "View report now? [y/N]: " view_report
if [[ "$view_report" =~ ^[Yy]$ ]]; then
    less "$OUTPUT_FILE"
fi