#!/usr/bin/env bash

#########################################################################################
#### disk-usage-reporter.sh — Find what's consuming disk space                       ####
#### Scans filesystems, ranks largest directories and files, flags old data          ####
####                                                                                 ####
#### Author: Phil Connor                                                             ####
#### Contact: contact@mylinux.work                                                   ####
#### License: MIT                                                                    ####
#### Version 1.00                                                                    ####
####                                                                                 ####
#### Usage:                                                                          ####
####   ./disk-usage-reporter.sh                                                      ####
####   ./disk-usage-reporter.sh --path /var                                          ####
####   ./disk-usage-reporter.sh --top 50 --min-size 100M                             ####
####   ./disk-usage-reporter.sh --json                                               ####
####                                                                                 ####
#### See --help for all options.                                                     ####
#########################################################################################

set -euo pipefail

# ============================================================================
# DEFAULTS
# ============================================================================

SCAN_PATH="/"
TOP_N=20
MIN_SIZE="1M"
MAX_DEPTH=3
AGE_WARN=90
JSON_MODE=false
NO_COLOR=false
VERSION="1.00"

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

SCRIPT_NAME="$(basename "$0")"

# ============================================================================
# USAGE & ARGUMENT PARSING
# ============================================================================

show_usage() {
    cat <<EOF
Usage: ${SCRIPT_NAME} [OPTIONS]

Scan filesystems to find what's consuming disk space (v${VERSION}).

OPTIONS:
  --path PATH          Path to scan (default: /)
  --top N              Show top N results per section (default: ${TOP_N})
  --min-size SIZE      Minimum file size to report (default: ${MIN_SIZE})
  --max-depth N        Directory depth limit for du scan (default: ${MAX_DEPTH})
  --age-warn DAYS      Flag files older than N days (default: ${AGE_WARN})
  --json               Output results in JSON format
  --no-color           Disable colored output
  -h, --help           Show this help

EXAMPLES:
  ${SCRIPT_NAME}                                  # Scan / with defaults
  ${SCRIPT_NAME} --path /var                      # Scan /var only
  ${SCRIPT_NAME} --top 50 --min-size 100M         # Top 50, files >= 100M
  ${SCRIPT_NAME} --path /home --age-warn 365      # Flag files older than 1 year
  ${SCRIPT_NAME} --json                           # JSON output for scripting

EOF
    exit 0
}

parse_args() {
    while [[ $# -gt 0 ]]; do
        case $1 in
            -h|--help)      show_usage ;;
            --path)         SCAN_PATH="$2"; shift 2 ;;
            --top)          TOP_N="$2"; shift 2 ;;
            --min-size)     MIN_SIZE="$2"; shift 2 ;;
            --max-depth)    MAX_DEPTH="$2"; shift 2 ;;
            --age-warn)     AGE_WARN="$2"; shift 2 ;;
            --json)         JSON_MODE=true; shift ;;
            --no-color)     NO_COLOR=true; shift ;;
            *)              echo "Unknown option: $1" >&2; exit 1 ;;
        esac
    done
    if [[ "$NO_COLOR" == true ]]; then
        RED="" YELLOW="" GREEN="" CYAN="" BOLD="" NC=""
    fi
}

# ============================================================================
# HELPERS
# ============================================================================

header() {
    echo ""
    echo -e "${CYAN}====================================================${NC}"
    echo -e "${CYAN}  ${BOLD}${1}${NC}"
    echo -e "${CYAN}====================================================${NC}"
    echo ""
}

format_bytes() {
    local b="$1"
    if [[ "$b" -ge 1073741824 ]]; then
        awk "BEGIN {printf \"%.2f GB\", $b/1073741824}"
    elif [[ "$b" -ge 1048576 ]]; then
        awk "BEGIN {printf \"%.1f MB\", $b/1048576}"
    elif [[ "$b" -ge 1024 ]]; then
        awk "BEGIN {printf \"%.1f KB\", $b/1024}"
    else
        echo "${b} B"
    fi
}

fmt_num() {
    printf "%'d" "$1" 2>/dev/null || echo "$1"
}

# Convert human-readable size (1M, 500K, 2G) to find -size argument
parse_min_size() {
    echo "${MIN_SIZE}"
}

# Convert human-readable size to bytes for comparison
size_to_bytes() {
    local size="$1"
    local num unit
    num="$(echo "$size" | sed 's/[^0-9.]//g')"
    unit="$(echo "$size" | sed 's/[0-9.]//g' | tr '[:lower:]' '[:upper:]')"
    case "$unit" in
        K) awk "BEGIN {printf \"%d\", $num * 1024}" ;;
        M) awk "BEGIN {printf \"%d\", $num * 1048576}" ;;
        G) awk "BEGIN {printf \"%d\", $num * 1073741824}" ;;
        T) awk "BEGIN {printf \"%d\", $num * 1099511627776}" ;;
        *) echo "$num" ;;
    esac
}

# ============================================================================
# FILESYSTEM OVERVIEW
# ============================================================================

filesystem_overview() {
    header "Filesystem Overview"

    printf "  ${BOLD}%-30s %6s %6s %6s %5s  %-20s${NC}\n" \
        "Filesystem" "Size" "Used" "Avail" "Use%" "Mounted on"
    echo "  ────────────────────────────────────────────────────────────────────────────────────"

    df -hP -x tmpfs -x devtmpfs -x squashfs 2>/dev/null | tail -n +2 | while IFS= read -r line; do
        local fs size used avail pct mount
        fs="$(echo "$line" | awk '{print $1}')"
        size="$(echo "$line" | awk '{print $2}')"
        used="$(echo "$line" | awk '{print $3}')"
        avail="$(echo "$line" | awk '{print $4}')"
        pct="$(echo "$line" | awk '{print $5}')"
        mount="$(echo "$line" | awk '{print $6}')"

        local pct_num="${pct%\%}"
        local color=""
        if [[ "$pct_num" -ge 90 ]]; then
            color="${RED}"
        elif [[ "$pct_num" -ge 80 ]]; then
            color="${YELLOW}"
        else
            color="${GREEN}"
        fi

        printf "  ${color}%-30s %6s %6s %6s %5s  %-20s${NC}\n" \
            "$fs" "$size" "$used" "$avail" "$pct" "$mount"
    done
}

# ============================================================================
# TOP DIRECTORIES BY SIZE
# ============================================================================

top_directories() {
    header "Top ${TOP_N} Directories by Size"

    printf "  ${BOLD}%4s  %-60s %10s${NC}\n" "#" "Directory" "Size"
    echo "  ────────────────────────────────────────────────────────────────────────────────────"

    du -x --max-depth="${MAX_DEPTH}" "${SCAN_PATH}" 2>/dev/null \
        | sort -rn \
        | head -n "${TOP_N}" \
        | while IFS=$'\t' read -r size_kb dir; do
            local num
            num=$((COUNTER + 1))
            COUNTER=$num
            local size_bytes=$((size_kb * 1024))
            local hsize
            hsize="$(format_bytes "$size_bytes")"

            local color="${NC}"
            if [[ "$size_bytes" -ge 10737418240 ]]; then
                color="${RED}"
            elif [[ "$size_bytes" -ge 1073741824 ]]; then
                color="${YELLOW}"
            fi

            printf "  ${color}%4d  %-60s %10s${NC}\n" "$num" "$dir" "$hsize"
        done
}

# ============================================================================
# TOP FILES BY SIZE
# ============================================================================

top_files() {
    header "Top ${TOP_N} Files by Size"

    printf "  ${BOLD}%4s  %-60s %10s${NC}\n" "#" "File" "Size"
    echo "  ────────────────────────────────────────────────────────────────────────────────────"

    find "${SCAN_PATH}" -xdev -type f -size +"$(parse_min_size)" -printf '%s\t%p\n' 2>/dev/null \
        | sort -rn \
        | head -n "${TOP_N}" \
        | awk -v idx=0 '{idx++; print idx"\t"$1"\t"$2}' \
        | while IFS=$'\t' read -r num size_bytes filepath; do
            local hsize
            hsize="$(format_bytes "$size_bytes")"

            local color="${NC}"
            if [[ "$size_bytes" -ge 1073741824 ]]; then
                color="${RED}"
            elif [[ "$size_bytes" -ge 104857600 ]]; then
                color="${YELLOW}"
            fi

            printf "  ${color}%4d  %-60s %10s${NC}\n" "$num" "$filepath" "$hsize"
        done
}

# ============================================================================
# OLD LARGE FILES
# ============================================================================

old_large_files() {
    header "Old Large Files (> ${MIN_SIZE}, older than ${AGE_WARN} days)"

    printf "  ${BOLD}%4s  %-50s %10s %12s${NC}\n" "#" "File" "Size" "Last Modified"
    echo "  ────────────────────────────────────────────────────────────────────────────────────"

    OLD_FILES_DATA="$(find "${SCAN_PATH}" -xdev -type f -size +"$(parse_min_size)" -mtime +"${AGE_WARN}" \
        -printf '%s\t%T+\t%p\n' 2>/dev/null \
        | sort -rn \
        | head -n "${TOP_N}")"

    OLD_FILE_COUNT=0
    OLD_FILE_BYTES=0

    if [[ -z "$OLD_FILES_DATA" ]]; then
        echo "  No files found matching criteria."
        return
    fi

    echo "$OLD_FILES_DATA" | awk -v idx=0 '{idx++; print idx"\t"$0}' \
        | while IFS=$'\t' read -r num size_bytes mtime filepath; do
            OLD_FILE_COUNT=$((OLD_FILE_COUNT + 1))
            OLD_FILE_BYTES=$((OLD_FILE_BYTES + size_bytes))

            local hsize mdate
            hsize="$(format_bytes "$size_bytes")"
            mdate="$(echo "$mtime" | cut -d'+' -f1)"

            printf "  ${YELLOW}%4d  %-50s %10s %12s${NC}\n" "$num" "$filepath" "$hsize" "$mdate"
        done
}

# ============================================================================
# SUMMARY
# ============================================================================

compute_summary() {
    local total_scanned old_count old_bytes

    total_scanned="$(du -sx "${SCAN_PATH}" 2>/dev/null | awk '{print $1}')"
    total_scanned=$((total_scanned * 1024))

    old_bytes="$(find "${SCAN_PATH}" -xdev -type f -size +"$(parse_min_size)" -mtime +"${AGE_WARN}" \
        -printf '%s\n' 2>/dev/null | awk '{s+=$1} END {print s+0}')"
    old_count="$(find "${SCAN_PATH}" -xdev -type f -size +"$(parse_min_size)" -mtime +"${AGE_WARN}" \
        2>/dev/null | wc -l)"

    echo "$total_scanned" "$old_count" "$old_bytes"
}

print_summary() {
    header "Summary"

    local data total_scanned old_count old_bytes
    data="$(compute_summary)"
    total_scanned="$(echo "$data" | awk '{print $1}')"
    old_count="$(echo "$data" | awk '{print $2}')"
    old_bytes="$(echo "$data" | awk '{print $3}')"

    echo -e "  ${BOLD}Scan path:${NC}              ${SCAN_PATH}"
    echo -e "  ${BOLD}Total scanned:${NC}          $(format_bytes "$total_scanned")"
    echo -e "  ${BOLD}Min file size:${NC}          ${MIN_SIZE}"
    echo -e "  ${BOLD}Age threshold:${NC}          ${AGE_WARN} days"
    echo ""
    echo -e "  ${BOLD}Old large files:${NC}        $(fmt_num "$old_count") files"
    echo -e "  ${BOLD}Reclaimable space:${NC}      ${YELLOW}$(format_bytes "$old_bytes")${NC}"
    echo ""

    if [[ "$old_bytes" -gt 0 ]]; then
        echo -e "  ${YELLOW}→ Review old files above — candidates for cleanup or archival${NC}"
    else
        echo -e "  ${GREEN}✓ No old large files found${NC}"
    fi
    echo ""
}

# ============================================================================
# JSON OUTPUT
# ============================================================================

json_output() {
    local total_scanned old_count old_bytes
    local data
    data="$(compute_summary)"
    total_scanned="$(echo "$data" | awk '{print $1}')"
    old_count="$(echo "$data" | awk '{print $2}')"
    old_bytes="$(echo "$data" | awk '{print $3}')"

    echo "{"
    echo "  \"scan_path\": \"${SCAN_PATH}\","
    echo "  \"timestamp\": \"$(date -u +%Y-%m-%dT%H:%M:%SZ)\","
    echo "  \"min_size\": \"${MIN_SIZE}\","
    echo "  \"age_warn_days\": ${AGE_WARN},"
    echo "  \"max_depth\": ${MAX_DEPTH},"

    # Filesystems
    echo "  \"filesystems\": ["
    local fs_first=true
    df -hP -x tmpfs -x devtmpfs -x squashfs 2>/dev/null | tail -n +2 | while IFS= read -r line; do
        local fs size used avail pct mount
        fs="$(echo "$line" | awk '{print $1}')"
        size="$(echo "$line" | awk '{print $2}')"
        used="$(echo "$line" | awk '{print $3}')"
        avail="$(echo "$line" | awk '{print $4}')"
        pct="$(echo "$line" | awk '{print $5}')"
        mount="$(echo "$line" | awk '{print $6}')"
        if [[ "$fs_first" == true ]]; then
            fs_first=false
        else
            echo ","
        fi
        printf '    {"filesystem":"%s","size":"%s","used":"%s","avail":"%s","use_pct":"%s","mount":"%s"}' \
            "$fs" "$size" "$used" "$avail" "$pct" "$mount"
    done
    echo ""
    echo "  ],"

    # Top directories
    echo "  \"top_directories\": ["
    local dir_first=true
    du -x --max-depth="${MAX_DEPTH}" "${SCAN_PATH}" 2>/dev/null \
        | sort -rn | head -n "${TOP_N}" \
        | while IFS=$'\t' read -r size_kb dir; do
            if [[ "$dir_first" == true ]]; then
                dir_first=false
            else
                echo ","
            fi
            printf '    {"path":"%s","size_bytes":%d}' "$dir" "$((size_kb * 1024))"
        done
    echo ""
    echo "  ],"

    # Top files
    echo "  \"top_files\": ["
    local file_first=true
    find "${SCAN_PATH}" -xdev -type f -size +"$(parse_min_size)" -printf '%s\t%p\n' 2>/dev/null \
        | sort -rn | head -n "${TOP_N}" \
        | while IFS=$'\t' read -r size_bytes filepath; do
            if [[ "$file_first" == true ]]; then
                file_first=false
            else
                echo ","
            fi
            printf '    {"path":"%s","size_bytes":%d}' "$filepath" "$size_bytes"
        done
    echo ""
    echo "  ],"

    # Old files
    echo "  \"old_large_files\": ["
    local old_first=true
    find "${SCAN_PATH}" -xdev -type f -size +"$(parse_min_size)" -mtime +"${AGE_WARN}" \
        -printf '%s\t%T+\t%p\n' 2>/dev/null \
        | sort -rn | head -n "${TOP_N}" \
        | while IFS=$'\t' read -r size_bytes mtime filepath; do
            local mdate
            mdate="$(echo "$mtime" | cut -d'+' -f1)"
            if [[ "$old_first" == true ]]; then
                old_first=false
            else
                echo ","
            fi
            printf '    {"path":"%s","size_bytes":%d,"last_modified":"%s"}' "$filepath" "$size_bytes" "$mdate"
        done
    echo ""
    echo "  ],"

    # Summary
    echo "  \"summary\": {"
    echo "    \"total_scanned_bytes\": ${total_scanned},"
    echo "    \"old_file_count\": ${old_count},"
    echo "    \"reclaimable_bytes\": ${old_bytes}"
    echo "  }"
    echo "}"
}

# ============================================================================
# MAIN
# ============================================================================

main() {
    parse_args "$@"

    if [[ ! -d "$SCAN_PATH" ]]; then
        echo -e "${RED}[ERROR]${NC} Path does not exist: ${SCAN_PATH}" >&2
        exit 1
    fi

    if [[ "$JSON_MODE" == true ]]; then
        json_output
        exit 0
    fi

    echo ""
    echo -e "${BOLD}Disk Usage Report${NC}"
    echo -e "$(date '+%Y-%m-%d %H:%M:%S %Z')  —  Scanning: ${SCAN_PATH}"

    COUNTER=0
    filesystem_overview
    top_directories
    top_files
    old_large_files
    print_summary
}

main "$@"