#!/bin/bash

################################################################################
# Intelligent Website Error Analyzer
################################################################################
# Purpose: Find REAL website issues affecting REAL users
# Filters out: Bots, plugin warnings, deprecation notices, minor issues
# Focuses on: Critical errors that break sites for actual visitors
################################################################################

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
source "$SCRIPT_DIR/lib/common-functions.sh"
source "$SCRIPT_DIR/lib/system-detect.sh"

# Configuration
APACHE_ERROR_LOG="/var/log/apache2/error_log"
DOMLOGS_DIR="/var/log/apache2/domlogs"
HOURS_TO_ANALYZE=24

print_banner "Intelligent Website Error Analyzer"

echo ""
echo "Finding REAL issues affecting REAL users..."
echo ""
echo "What we filter out:"
echo "  ✗ Bot/scanner activity"
echo "  ✗ Plugin deprecation warnings"
echo "  ✗ PHP notices (non-critical)"
echo "  ✗ Random 404s for missing favicons/assets"
echo "  ✗ Security probe attempts"
echo ""
echo "What we focus on:"
echo "  ✓ 500 errors users actually see"
echo "  ✓ Fatal PHP errors breaking functionality"
echo "  ✓ Database connection failures"
echo "  ✓ Permission issues blocking access"
echo "  ✓ Memory exhaustion affecting performance"
echo ""

# Ask for time range
echo -e "${CYAN}How far back should we analyze?${NC}"
echo "  1) Last 1 hour"
echo "  2) Last 6 hours"
echo "  3) Last 24 hours (default)"
echo "  4) Last 7 days"
echo "  5) Last 30 days"
echo ""
read -p "Select option [3]: " time_choice
time_choice=${time_choice:-3}

case $time_choice in
    1) HOURS_TO_ANALYZE=1 ;;
    2) HOURS_TO_ANALYZE=6 ;;
    3) HOURS_TO_ANALYZE=24 ;;
    4) HOURS_TO_ANALYZE=168 ;;
    5) HOURS_TO_ANALYZE=720 ;;
    *) HOURS_TO_ANALYZE=24 ;;
esac

echo ""
echo "→ Analyzing last $HOURS_TO_ANALYZE hours..."
echo ""

# Temporary files
TEMP_DIR="/tmp/website-error-analysis-$$"
mkdir -p "$TEMP_DIR"
trap "rm -rf $TEMP_DIR" EXIT

CRITICAL_ERRORS="$TEMP_DIR/critical.txt"
USER_IMPACT_ERRORS="$TEMP_DIR/user_impact.txt"
LOG_FILES_LIST="$TEMP_DIR/log_files.txt"

# Discover all log file locations
echo "→ Discovering log file locations..."
> "$LOG_FILES_LIST"

# Apache main error log
for log in /var/log/apache2/error_log /usr/local/apache/logs/error_log /var/log/httpd/error_log; do
    [ -f "$log" ] && echo "$log|apache_main" >> "$LOG_FILES_LIST"
done

# Per-domain PHP error logs in public_html
find /home/*/public_html -name "error_log" -type f 2>/dev/null | while read -r log; do
    domain=$(echo "$log" | grep -oE '/home/[^/]+' | sed 's|/home/||')
    echo "$log|php_$domain" >> "$LOG_FILES_LIST"
done

# Per-domain Apache logs in domlogs
if [ -d "$DOMLOGS_DIR" ]; then
    for log in "$DOMLOGS_DIR"/*; do
        [ -f "$log" ] && echo "$log|domlog_$(basename "$log")" >> "$LOG_FILES_LIST"
    done
fi

log_count=$(wc -l < "$LOG_FILES_LIST" 2>/dev/null || echo "0")
echo "  Found $log_count log files to analyze"
echo ""

################################################################################
# Intelligent Filtering
################################################################################

is_noise() {
    local line="$1"

    # Bot/Scanner patterns
    if echo "$line" | grep -qiE "bot|crawler|spider|scanner|nikto|nmap|masscan|sqlmap|nessus|acunetix|burp|shodan|censys|zgrab|nuclei|semrush|ahrefs|mj12"; then
        return 0
    fi

    # Known scanner IPs
    if echo "$line" | grep -qE "45\.148\.|185\.220\.|89\.248\.165\."; then
        return 0
    fi

    # WordPress plugin deprecation warnings (not user-facing)
    if echo "$line" | grep -qiE "PHP Deprecated|Deprecated.*plugin|call_user_func_array.*deprecated"; then
        return 0
    fi

    # Non-critical PHP notices
    if echo "$line" | grep -qiE "PHP Notice.*Undefined (variable|index|offset)" && \
       ! echo "$line" | grep -qiE "fatal|error 500|white screen"; then
        return 0
    fi

    # Missing favicon/common assets (not real errors)
    if echo "$line" | grep -qiE "favicon\.ico|apple-touch-icon|robots\.txt|sitemap\.xml.*404"; then
        return 0
    fi

    # Security probes
    if echo "$line" | grep -qiE "wp-admin|wp-login|phpMyAdmin|phpmyadmin|\.env|\.git|config\.php|xmlrpc|eval-stdin|shell\.php|c99\.php|r57\.php|adminer\.php" && \
       echo "$line" | grep -qiE "404|403|not found"; then
        return 0
    fi

    # WordPress auto-updates and cron (normal operations)
    if echo "$line" | grep -qiE "wp-cron\.php|doing_cron|auto.*update"; then
        return 0
    fi

    # Common plugin update checks (not errors)
    if echo "$line" | grep -qiE "update-check|version-check|api\.wordpress\.org"; then
        return 0
    fi

    return 1
}

is_critical_user_facing() {
    local line="$1"

    # 500 Internal Server Error (users see white page)
    if echo "$line" | grep -qiE " 500 |Internal Server Error"; then
        return 0
    fi

    # PHP Fatal Errors (breaks functionality)
    if echo "$line" | grep -qiE "PHP Fatal error|Fatal error:.*in /"; then
        return 0
    fi

    # PHP Parse Errors (site completely broken)
    if echo "$line" | grep -qiE "PHP Parse error|syntax error"; then
        return 0
    fi

    # Database connection failures (site down)
    if echo "$line" | grep -qiE "Error establishing.*database|Can't connect.*MySQL|Access denied for user.*database|Too many connections|MySQL server has gone away"; then
        return 0
    fi

    # Memory exhaustion (white screen/incomplete pages)
    if echo "$line" | grep -qiE "Allowed memory size.*exhausted|Out of memory|Fatal.*memory"; then
        return 0
    fi

    # Segmentation fault (complete crash)
    if echo "$line" | grep -qiE "Segmentation fault|signal 11"; then
        return 0
    fi

    # Permission denied on critical files
    if echo "$line" | grep -qiE "Permission denied" && \
       echo "$line" | grep -qE "index\.php|wp-config\.php|\.htaccess|config\.php"; then
        return 0
    fi

    # File not found for MAIN page (404 on homepage/index)
    if echo "$line" | grep -qiE "File does not exist.*index\.(php|html)" || \
       echo "$line" | grep -qE " 404 .*\" \"GET / "; then
        return 0
    fi

    return 1
}

extract_useful_info() {
    local line="$1"

    # Extract domain
    domain=$(echo "$line" | grep -oE '\[vhost [^:]+' | sed 's/\[vhost //' || \
             echo "$line" | grep -oE '[a-zA-Z0-9.-]+\.(com|net|org|io|co|uk|us|dev)' | head -1 || \
             echo "$line" | grep -oE '/home/[^/]+' | sed 's|/home/||' || echo "unknown")

    # Extract file path if PHP error
    file_path=$(echo "$line" | grep -oE "in /[^ ]+\.php" | sed 's/in //' || echo "")

    # Extract error message (simplified)
    error_msg=$(echo "$line" | sed 's/^\[.*\] //' | sed 's/\[client [^]]*\] //' | cut -c1-100)

    echo "$domain|$file_path|$error_msg"
}

################################################################################
# Main Analysis
################################################################################

echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo " SCANNING FOR USER-IMPACTING ERRORS"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""

total_lines=0
filtered_out=0
critical_found=0

# Analyze all discovered log files
while IFS='|' read -r log_path log_type; do
    echo "  Analyzing: $(basename "$log_path")..."

    # Determine if it's an access log (contains status codes) or error log
    is_access_log=false
    [[ "$log_type" == domlog_* ]] && ! [[ "$log_path" =~ error ]] && is_access_log=true

    if $is_access_log; then
        # Access log - look for 5xx status codes
        while IFS= read -r line; do
            ((total_lines++))

            # Skip if bot/scanner
            if is_noise "$line"; then
                ((filtered_out++))
                continue
            fi

            # Extract status code and URL
            if echo "$line" | grep -qE '" 5[0-9]{2} '; then
                ((critical_found++))
                status=$(echo "$line" | grep -oE '" 5[0-9]{2} ' | tr -d '" ')
                url=$(echo "$line" | awk '{print $7}' | cut -c1-80)
                ip=$(echo "$line" | awk '{print $1}')
                domain=$(basename "$log_path" | sed 's/-.*//')

                echo "$domain||HTTP $status - $url (from $ip)" >> "$CRITICAL_ERRORS"
            fi

        done < <(tail -n 5000 "$log_path" 2>/dev/null)
    else
        # Error log - look for critical errors
        while IFS= read -r line; do
            ((total_lines++))

            # Skip noise
            if is_noise "$line"; then
                ((filtered_out++))
                continue
            fi

            # Check if it's critical and user-facing
            if is_critical_user_facing "$line"; then
                ((critical_found++))
                extract_useful_info "$line" >> "$CRITICAL_ERRORS"
            fi

        done < <(tail -n 10000 "$log_path" 2>/dev/null)
    fi

done < "$LOG_FILES_LIST"

echo ""
echo "Scanned: $total_lines log entries"
echo "Filtered: $filtered_out noise entries (bots, warnings, etc.)"
echo "Found: $critical_found critical user-facing errors"
echo ""

################################################################################
# Generate Report
################################################################################

if [ ! -f "$CRITICAL_ERRORS" ] || [ ! -s "$CRITICAL_ERRORS" ]; then
    echo ""
    echo -e "${GREEN}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}"
    echo -e "${GREEN} ✓ NO CRITICAL USER-FACING ERRORS FOUND!${NC}"
    echo -e "${GREEN}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}"
    echo ""
    echo "Your websites appear to be functioning normally for real users."
    echo "All errors found were from bots, scanners, or non-critical warnings."
    echo ""
    exit 0
fi

echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo " ⚠️  CRITICAL ERRORS AFFECTING REAL USERS"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""

# Group by domain and count
echo -e "${RED}${BOLD}AFFECTED WEBSITES:${NC}"
echo ""

cat "$CRITICAL_ERRORS" | cut -d'|' -f1 | sort | uniq -c | sort -rn | head -10 | while read -r count domain; do
    printf "  ${RED}●${NC} %-40s ${BOLD}%4d errors${NC}\n" "$domain" "$count"
done

echo ""
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo " ERROR DETAILS"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""

# Show errors grouped by domain
current_domain=""
domain_count=0

sort "$CRITICAL_ERRORS" | while IFS='|' read -r domain file_path error_msg; do
    if [ "$domain" != "$current_domain" ]; then
        if [ -n "$current_domain" ]; then
            echo ""
        fi

        ((domain_count++))
        [ $domain_count -gt 5 ] && break

        echo -e "${YELLOW}▼ $domain${NC}"
        current_domain="$domain"
    fi

    if [ -n "$file_path" ]; then
        echo "    File: $file_path"
    fi

    echo "    Error: $error_msg"
    echo ""
done

################################################################################
# Intelligent Recommendations
################################################################################

echo ""
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo " 🔧 RECOMMENDED ACTIONS"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""

# Detect issue types and provide specific recommendations
if grep -qiE "500|Internal Server" "$CRITICAL_ERRORS"; then
    echo -e "${RED}● 500 Internal Server Errors detected${NC}"
    echo "  Priority: URGENT - Users see broken pages"
    echo ""
    echo "  Actions:"
    echo "    1. Check PHP error logs for affected domains:"
    for domain in $(grep -iE "500|Internal Server" "$CRITICAL_ERRORS" | cut -d'|' -f1 | sort -u | head -3); do
        echo "       tail -100 /home/$domain/public_html/error_log"
    done
    echo ""
    echo "    2. Common causes:"
    echo "       - PHP syntax errors (check recent file changes)"
    echo "       - .htaccess syntax errors"
    echo "       - Incorrect file permissions"
    echo "       - Missing PHP modules"
    echo ""
fi

if grep -qiE "Fatal error|Parse error" "$CRITICAL_ERRORS"; then
    echo -e "${RED}● PHP Fatal/Parse Errors detected${NC}"
    echo "  Priority: URGENT - Site functionality broken"
    echo ""
    echo "  Actions:"
    echo "    1. Review recent plugin/theme updates"
    echo "    2. Check PHP version compatibility"
    echo "    3. Look for syntax errors in custom code"
    echo "    4. Temporarily disable recently added plugins"
    echo ""
fi

if grep -qiE "database|MySQL|Connection" "$CRITICAL_ERRORS"; then
    echo -e "${RED}● Database Connection Errors detected${NC}"
    echo "  Priority: CRITICAL - Site may be completely down"
    echo ""
    echo "  Actions:"
    echo "    1. Check MySQL service: systemctl status mysql"
    echo "    2. Verify database credentials in config files"
    echo "    3. Check max_connections: mysql -e 'SHOW VARIABLES LIKE \"max_connections\"'"
    echo "    4. Review current connections: mysql -e 'SHOW PROCESSLIST'"
    echo ""
fi

if grep -qiE "memory.*exhausted|Out of memory" "$CRITICAL_ERRORS"; then
    echo -e "${RED}● Memory Exhaustion detected${NC}"
    echo "  Priority: HIGH - Pages loading slowly or incompletely"
    echo ""
    echo "  Current PHP memory limit: $(php -r 'echo ini_get("memory_limit");' 2>/dev/null)"
    echo ""
    echo "  Actions:"
    echo "    1. Increase memory_limit in php.ini or .user.ini"
    echo "    2. Recommended: 256M minimum, 512M for WooCommerce"
    echo "    3. Check for memory leaks in plugins"
    echo "    4. Optimize database queries"
    echo ""
fi

echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""
echo "Full analysis saved to: $CRITICAL_ERRORS"
echo ""
echo "Next steps:"
echo "  1. Address critical errors first (500s, database, fatal errors)"
echo "  2. Check error_log files for affected domains"
echo "  3. Test affected pages after fixes"
echo "  4. Re-run this tool to verify issues are resolved"
echo ""

press_enter