#!/usr/bin/env bash

#########################################################################################
#### log-disk-analyzer.sh — Analyze log directory disk usage and report problems     ####
#### Reports largest files, growth rates, unrotated logs, broken symlinks,           ####
#### empty files, and subdirectory breakdown                                         ####
####                                                                                 ####
#### Author: Phil Connor                                                             ####
#### Contact: contact@mylinux.work                                                   ####
#### License: MIT                                                                    ####
#### Version 1.00                                                                    ####
####                                                                                 ####
#### Usage:                                                                          ####
####   ./log-disk-analyzer.sh                                                        ####
####   ./log-disk-analyzer.sh --path /opt/app/logs --top 10                          ####
####   ./log-disk-analyzer.sh --json                                                 ####
####                                                                                 ####
#### See --help for all options.                                                     ####
#########################################################################################

set -euo pipefail

# ── Defaults ──────────────────────────────────────────────────────────
LOG_PATH="${LOG_PATH:-/var/log}"
TOP_COUNT="${TOP_COUNT:-20}"
COLOR="${COLOR:-auto}"
TEXTFILE_DIR="/var/lib/node_exporter"
PROM_FILE=""
JSON_OUTPUT="false"
VERBOSE="false"

# ── State ─────────────────────────────────────────────────────────────
SCRIPT_NAME="$(basename "$0")"
readonly SCRIPT_NAME
TOTAL_SIZE=0  TOTAL_FILES=0  EMPTY_COUNT=0  BROKEN_COUNT=0  UNROTATED_COUNT=0
FILE_LIST=()  SIZE_LIST=()  RECOMMENDATIONS=()

# ── Colors ────────────────────────────────────────────────────────────
setup_colors() {
    if [[ "$COLOR" == "never" ]]; then
        RED="" GREEN="" YELLOW="" BOLD="" DIM="" RESET=""; return
    fi
    if [[ "$COLOR" == "always" ]] || [[ -t 1 ]]; then
        RED='\033[0;31m'  GREEN='\033[0;32m'  YELLOW='\033[0;33m'
        BOLD='\033[1m'    DIM='\033[2m'        RESET='\033[0m'
    else
        RED="" GREEN="" YELLOW="" BOLD="" DIM="" RESET=""
    fi
}

# ── Logging ───────────────────────────────────────────────────────────
warn()    { echo -e "${YELLOW}[WARN]${RESET} $*" >&2; }
err()     { echo -e "${RED}[ERROR]${RESET} $*" >&2; }

# ── Helpers ───────────────────────────────────────────────────────────
human_size() {
    local b="$1"
    if   [[ "$b" -ge 1073741824 ]]; then printf "%.1f GB" "$(echo "scale=1; $b/1073741824" | bc)"
    elif [[ "$b" -ge 1048576 ]];    then printf "%.1f MB" "$(echo "scale=1; $b/1048576" | bc)"
    elif [[ "$b" -ge 1024 ]];       then printf "%.1f KB" "$(echo "scale=1; $b/1024" | bc)"
    else printf "%d B" "$b"; fi
}

file_age_days() {
    local mtime now age
    mtime=$(stat -c %Y "$1" 2>/dev/null) || return 1
    now=$(date +%s); age=$(( (now - mtime) / 86400 ))
    [[ "$age" -lt 1 ]] && age=1
    echo "$age"
}

separator() { printf "  %s\n" "$(printf '%.0s─' {1..62})"; }

json_escape() { local s="$1"; s="${s//\\/\\\\}"; s="${s//\"/\\\"}"; printf '%s' "$s"; }

# ── Collect all files ─────────────────────────────────────────────────
collect_files() {
    while IFS= read -r line; do
        local size="${line%%	*}" file="${line#*	}"
        [[ -z "$size" || -z "$file" ]] && continue
        FILE_LIST+=("$file"); SIZE_LIST+=("$size")
        TOTAL_SIZE=$((TOTAL_SIZE + size)); TOTAL_FILES=$((TOTAL_FILES + 1))
    done < <(find "$LOG_PATH" -type f -printf '%s\t%p\n' 2>/dev/null | sort -t$'\t' -k1 -rn)
}

# ── Top files by size ─────────────────────────────────────────────────
print_top_files() {
    local limit="$TOP_COUNT"
    [[ "$VERBOSE" == "true" ]] && limit="${#FILE_LIST[@]}"
    [[ "$limit" -gt "${#FILE_LIST[@]}" ]] && limit="${#FILE_LIST[@]}"
    echo ""; echo -e "  ${BOLD}TOP ${limit} FILES BY SIZE${RESET}"; separator
    for (( i = 0; i < limit; i++ )); do
        local size_h color=""
        size_h="$(human_size "${SIZE_LIST[$i]}")"
        [[ "${SIZE_LIST[$i]}" -ge 104857600 ]] && color="$RED"
        [[ "${SIZE_LIST[$i]}" -lt 104857600 && "${SIZE_LIST[$i]}" -ge 52428800 ]] && color="$YELLOW"
        printf "  %b%-10s%b  %s\n" "$color" "$size_h" "$RESET" "${FILE_LIST[$i]}"
    done
}

# ── Growth rates ──────────────────────────────────────────────────────
print_growth_rates() {
    local limit="$TOP_COUNT"
    [[ "$VERBOSE" == "true" ]] && limit="${#FILE_LIST[@]}"
    [[ "$limit" -gt "${#FILE_LIST[@]}" ]] && limit="${#FILE_LIST[@]}"
    local -a rates=() rate_files=() rate_ages=()
    for (( i = 0; i < ${#FILE_LIST[@]}; i++ )); do
        local age; age=$(file_age_days "${FILE_LIST[$i]}") || continue
        rates+=("$(( SIZE_LIST[i] / age ))"); rate_files+=("${FILE_LIST[$i]}"); rate_ages+=("$age")
    done
    [[ ${#rates[@]} -eq 0 ]] && return
    local -a sorted_idx=()
    mapfile -t sorted_idx < <(for i in "${!rates[@]}"; do echo "$i ${rates[$i]}"; done | sort -k2 -rn | head -n "$limit" | awk '{print $1}')
    echo ""; echo -e "  ${BOLD}GROWTH RATE (SIZE / AGE)${RESET}"; separator
    for idx in "${sorted_idx[@]}"; do
        printf "  %-16s %s  (%d days old)\n" "$(human_size "${rates[$idx]}")/day" "${rate_files[$idx]}" "${rate_ages[$idx]}"
    done
}

# ── Unrotated logs ────────────────────────────────────────────────────
print_unrotated() {
    echo ""; echo -e "  ${BOLD}UNROTATED LOGS (>100 MB, >7 days old)${RESET}"; separator
    local found=0
    while IFS= read -r file; do
        [[ -z "$file" ]] && continue
        local size age; size=$(stat -c %s "$file" 2>/dev/null) || continue
        [[ "$size" -lt 104857600 ]] && continue
        age=$(file_age_days "$file") || continue; [[ "$age" -lt 7 ]] && continue
        found=1; UNROTATED_COUNT=$((UNROTATED_COUNT + 1))
        local size_h; size_h="$(human_size "$size")"
        printf "  %b%-10s%b  %3d days   %s\n" "$RED" "$size_h" "$RESET" "$age" "$file"
        RECOMMENDATIONS+=("Rotate $(basename "$file") — ${size_h} and ${age} days old")
    done < <(find "$LOG_PATH" -type f -name '*.log' 2>/dev/null)
    [[ "$found" -eq 0 ]] && echo -e "  ${GREEN}None found${RESET}"
}

# ── Empty files ───────────────────────────────────────────────────────
print_empty_files() {
    echo ""; echo -e "  ${BOLD}EMPTY FILES${RESET}"; separator
    local found=0
    while IFS= read -r file; do
        [[ -z "$file" ]] && continue; found=1; EMPTY_COUNT=$((EMPTY_COUNT + 1))
        echo "  $file"
    done < <(find "$LOG_PATH" -type f -empty 2>/dev/null)
    [[ "$found" -eq 0 ]] && echo -e "  ${GREEN}None found${RESET}"
    if [[ "$EMPTY_COUNT" -gt 0 ]]; then
        RECOMMENDATIONS+=("Remove ${EMPTY_COUNT} empty log files to reclaim inodes")
    fi
}

# ── Broken symlinks ──────────────────────────────────────────────────
print_broken_symlinks() {
    echo ""; echo -e "  ${BOLD}BROKEN SYMLINKS${RESET}"; separator
    local found=0
    while IFS= read -r link; do
        [[ -z "$link" ]] && continue; found=1; BROKEN_COUNT=$((BROKEN_COUNT + 1))
        local target; target=$(readlink "$link" 2>/dev/null || echo "unknown")
        printf "  %b%s%b -> %s\n" "$YELLOW" "$link" "$RESET" "$target"
    done < <(find "$LOG_PATH" -xtype l 2>/dev/null)
    [[ "$found" -eq 0 ]] && echo -e "  ${GREEN}None found${RESET}"
    if [[ "$BROKEN_COUNT" -gt 0 ]]; then
        RECOMMENDATIONS+=("Fix ${BROKEN_COUNT} broken symlinks")
    fi
}

# ── Subdirectory breakdown ────────────────────────────────────────────
print_subdir_breakdown() {
    echo ""; echo -e "  ${BOLD}DISK USAGE BY SUBDIRECTORY${RESET}"; separator
    du -d 1 "$LOG_PATH" 2>/dev/null | sort -rn | while IFS=$'\t' read -r kb path; do
        [[ "$path" == "$LOG_PATH" ]] && continue
        printf "  %-10s  %s\n" "$(human_size $((kb * 1024)))" "$path"
    done || true
}

# ── Summary ───────────────────────────────────────────────────────────
print_summary() {
    echo ""; echo -e "  ${BOLD}Summary${RESET}"
    printf "  %-20s %s\n" "Total size:" "$(human_size "$TOTAL_SIZE")"
    printf "  %-20s %d\n" "Total files:" "$TOTAL_FILES"
    printf "  %-20s %d\n" "Empty files:" "$EMPTY_COUNT"
    printf "  %-20s %d\n" "Broken symlinks:" "$BROKEN_COUNT"
    printf "  %-20s %d\n" "Unrotated logs:" "$UNROTATED_COUNT"
    if [[ ${#RECOMMENDATIONS[@]} -gt 0 ]]; then
        echo ""; echo -e "  ${BOLD}Recommendations:${RESET}"
        for rec in "${RECOMMENDATIONS[@]}"; do echo -e "  ${YELLOW}•${RESET} $rec"; done
    fi
    echo ""
}

# ── JSON output ───────────────────────────────────────────────────────
print_json() {
    local limit="$TOP_COUNT" i first
    [[ "$VERBOSE" == "true" ]] && limit="${#FILE_LIST[@]}"
    [[ "$limit" -gt "${#FILE_LIST[@]}" ]] && limit="${#FILE_LIST[@]}"
    printf '{"path":"%s","top_files":[' "$(json_escape "$LOG_PATH")"
    for (( i = 0; i < limit; i++ )); do
        [[ "$i" -gt 0 ]] && printf ','
        printf '{"file":"%s","size":%s}' "$(json_escape "${FILE_LIST[$i]}")" "${SIZE_LIST[$i]}"
    done
    printf '],"growth_rates":['
    first=1
    for (( i = 0; i < limit && i < ${#FILE_LIST[@]}; i++ )); do
        local age; age=$(file_age_days "${FILE_LIST[$i]}") || continue
        [[ "$first" -eq 0 ]] && printf ','; first=0
        printf '{"file":"%s","bytes_per_day":%d,"age_days":%d}' \
            "$(json_escape "${FILE_LIST[$i]}")" "$(( SIZE_LIST[i] / age ))" "$age"
    done
    printf '],"unrotated_logs":['
    first=1
    while IFS= read -r file; do
        [[ -z "$file" ]] && continue
        local size age; size=$(stat -c %s "$file" 2>/dev/null) || continue
        [[ "$size" -lt 104857600 ]] && continue
        age=$(file_age_days "$file") || continue; [[ "$age" -lt 7 ]] && continue
        [[ "$first" -eq 0 ]] && printf ','; first=0
        printf '{"file":"%s","size":%d,"age_days":%d}' "$(json_escape "$file")" "$size" "$age"
    done < <(find "$LOG_PATH" -type f -name '*.log' 2>/dev/null)
    printf '],"empty_files":['
    first=1
    while IFS= read -r file; do
        [[ -z "$file" ]] && continue; [[ "$first" -eq 0 ]] && printf ','; first=0
        printf '"%s"' "$(json_escape "$file")"
    done < <(find "$LOG_PATH" -type f -empty 2>/dev/null)
    printf '],"broken_symlinks":['
    first=1
    while IFS= read -r link; do
        [[ -z "$link" ]] && continue; [[ "$first" -eq 0 ]] && printf ','; first=0
        printf '{"link":"%s","target":"%s"}' "$(json_escape "$link")" "$(json_escape "$(readlink "$link" 2>/dev/null || echo unknown)")"
    done < <(find "$LOG_PATH" -xtype l 2>/dev/null)
    printf '],"summary":{"total_size":%d,"total_files":%d,"empty_files":%d,"broken_symlinks":%d,"unrotated_logs":%d}}\n' \
        "$TOTAL_SIZE" "$TOTAL_FILES" "$EMPTY_COUNT" "$BROKEN_COUNT" "$UNROTATED_COUNT"
}

# ── Prometheus output ─────────────────────────────────────────────────
write_prometheus() {
    local file="$1"
    local output_dir
    output_dir="$(dirname "$file")"
    mkdir -p "$output_dir"
    local tmp
    tmp=$(mktemp "${output_dir}/.log_disk.XXXXXX")
    {
        echo "# HELP log_disk_total_bytes Total size of log directory in bytes"
        echo "# TYPE log_disk_total_bytes gauge"
        printf 'log_disk_total_bytes{path="%s"} %d\n' "$LOG_PATH" "$TOTAL_SIZE"
        echo "# HELP log_disk_total_files Total number of files in log directory"
        echo "# TYPE log_disk_total_files gauge"
        printf 'log_disk_total_files{path="%s"} %d\n' "$LOG_PATH" "$TOTAL_FILES"
        echo "# HELP log_disk_empty_files Number of empty files in log directory"
        echo "# TYPE log_disk_empty_files gauge"
        printf 'log_disk_empty_files{path="%s"} %d\n' "$LOG_PATH" "$EMPTY_COUNT"
        echo "# HELP log_disk_broken_symlinks Number of broken symlinks in log directory"
        echo "# TYPE log_disk_broken_symlinks gauge"
        printf 'log_disk_broken_symlinks{path="%s"} %d\n' "$LOG_PATH" "$BROKEN_COUNT"
        echo "# HELP log_disk_unrotated_logs Number of unrotated log files"
        echo "# TYPE log_disk_unrotated_logs gauge"
        printf 'log_disk_unrotated_logs{path="%s"} %d\n' "$LOG_PATH" "$UNROTATED_COUNT"
    } > "$tmp"
    chmod 644 "$tmp"
    mv -f "$tmp" "$file"
    verbose "Metrics written to ${file}"
}

# ══════════════════════════════════════════════════════════════════════
#                       USAGE
# ══════════════════════════════════════════════════════════════════════

usage() {
    cat <<EOF
${SCRIPT_NAME} — Analyze log directory disk usage and report problems

USAGE:
  ${SCRIPT_NAME} [OPTIONS]

OPTIONS:
  --path DIR            Directory to scan (default: ${LOG_PATH})
  --top N               Show top N largest files (default: ${TOP_COUNT})
  --json                Output results in JSON format
  --no-color            Disable colored output
  --textfile            Write metrics to node_exporter textfile collector
  -o, --output PATH     Write metrics to custom file path
  --verbose             Show all files, not just top N
  --help                Show this help

ENVIRONMENT VARIABLES:
  LOG_PATH              Directory to scan (default: /var/log)
  TOP_COUNT             Number of top files to show (default: 20)
  COLOR                 Color mode: auto, always, never (default: auto)

EXAMPLES:
  sudo ./log-disk-analyzer.sh
  ./log-disk-analyzer.sh --path /opt/app/logs
  ./log-disk-analyzer.sh --json
  ./log-disk-analyzer.sh --textfile
  ./log-disk-analyzer.sh -o /tmp/log_disk.prom
EOF
}

# ══════════════════════════════════════════════════════════════════════
#                       ARGUMENT PARSING
# ══════════════════════════════════════════════════════════════════════

parse_args() {
    while [[ $# -gt 0 ]]; do
        case "$1" in
            --path)      LOG_PATH="$2"; shift 2 ;;
            --top)       TOP_COUNT="$2"; shift 2 ;;
            --json)      JSON_OUTPUT="true"; shift ;;
            --verbose)   VERBOSE="true"; shift ;;
            --no-color)  COLOR="never"; shift ;;
            --textfile)  PROM_FILE="$TEXTFILE_DIR/log_disk.prom"; shift ;;
            -o|--output) PROM_FILE="$2"; shift 2 ;;
            --help|-h)   setup_colors; usage; exit 0 ;;
            -*)          err "Unknown option: $1"; echo "Run ${SCRIPT_NAME} --help for usage" >&2; exit 1 ;;
            *)           err "Unexpected argument: $1"; echo "Run ${SCRIPT_NAME} --help for usage" >&2; exit 1 ;;
        esac
    done
}

# ══════════════════════════════════════════════════════════════════════
#                       MAIN
# ══════════════════════════════════════════════════════════════════════

main() {
    parse_args "$@"
    setup_colors

    if [[ ! -d "$LOG_PATH" ]]; then
        err "Directory not found: $LOG_PATH"
        exit 1
    fi

    collect_files

    if [[ "$JSON_OUTPUT" == "true" ]]; then
        print_json
    else
        echo ""
        echo -e "${BOLD}Log Disk Analyzer${RESET}"
        echo -e "${DIM}Path: ${LOG_PATH}${RESET}"
        print_top_files
        print_growth_rates
        print_unrotated
        print_empty_files
        print_broken_symlinks
        print_subdir_breakdown
        print_summary
    fi

    if [[ -n "$PROM_FILE" ]]; then
        write_prometheus "$PROM_FILE"
    fi
}

main "$@"