backup_github_full.sh

#!/usr/bin/env bash
# =============================================================================
# GitHub Full/Lite Backup (Bash)
# - 메시지는 영어 출력, 주석은 한국어
# - 토큰 없으면 LITE: 코드 + Wiki 백업만 수행
# - 토큰 있으면 FULL: LITE + 메타데이터(issues/comments, pulls/comments/reviews,
#   labels, milestones, releases+assets, (옵션) discussions)까지 백업
# - 실행 전 모드 안내 및 y/N 확인, 백업 폴더 존재 시 덮어쓰기 y/N 확인
# - 의존성: git, curl (FULL 모드: jq 필요), zip 또는 tar (압축 옵션 사용 시)
# 사용법:
#   ./backup_github_full.sh <RepoUrl> [--out-dir DIR] [--token TOKEN] [--zip] [--include-discussions]
# 예:
#   ./backup_github_full.sh https://github.com/user/repo.git
#   ./backup_github_full.sh https://github.com/user/repo.git --out-dir /data/gh
#   ./backup_github_full.sh https://github.com/user/repo.git --token ghp_xxx --zip --include-discussions
#   GITHUB_TOKEN=ghp_xxx ./backup_github_full.sh https://github.com/user/repo.git
# =============================================================================

set -u  # 정의 안 된 변수를 오류로 처리
# set -e  # 치명적 오류 시 즉시 종료 (프롬프트/분기 흐름을 위해 생략)

# ---------- 공통 유틸 (한국어 주석) -------------------------------------------
say_section() {  # 구분선과 섹션 타이틀 출력
  echo "--------------------------------------------------"
  echo "$1"
  echo "--------------------------------------------------"
}  

need_cmd() {  # 명령 유무 확인
  if ! command -v "$1" >/dev/null 2>&1; then
    echo "[ERROR] '$1' is not installed or not in PATH." >&2
    exit 1
  fi
}

url_basename_no_git() {  # URL에서 .git 제거하고 마지막 이름 추출
  local u="$1"
  u="${u%.git}"
  # 마지막 / 이후 이름
  basename "$u"
}

build_wiki_url() {  # Wiki URL 생성
  local u="$1"
  if [[ "$u" == *.git ]]; then
    echo "${u%.git}.wiki.git"
  else
    echo "${u%.git}.wiki.git"
  fi
}

join_path() {  # 간단한 경로 결합
  local base="$1"; shift
  local leaf="$1"
  if [[ "$base" == */ ]]; then
    echo "${base}${leaf}"
  else
    echo "${base}/${leaf}"
  fi
}

ensure_clean_dir_with_prompt() {  # 폴더 존재 시 덮어쓰기 여부 묻고 정리
  local dir="$1"
  if [[ -d "$dir" ]]; then
    echo "The backup folder already exists: \"$dir\""
    read -r -p "Do you want to OVERWRITE it? [y/N]: " ans
    case "$ans" in
      y|Y|yes|YES) ;;
      *) echo "Aborted. Nothing was changed."; exit 0 ;;
    esac
    echo "Removing existing folder..."
    rm -rf -- "$dir" || { echo "[ERROR] Failed to remove the folder."; exit 1; }
  fi
  mkdir -p -- "$dir" || { echo "[ERROR] Failed to create the backup folder."; exit 1; }
}

rest_paged_to_file() {  # REST 페이지네이션 처리 → 하나의 JSON 배열 파일로 저장 (jq 필요)
  # 인자: $1=url (쿼리 유무 무관), $2=token, $3=outfile
  local base_url="$1" token="$2" out="$3"
  local page=1 tmpdir
  tmpdir="$(mktemp -d)"
  local files=()
  while :; do
    local sep='?'
    [[ "$base_url" == *\?* ]] && sep='&'
    local u="${base_url}${sep}per_page=100&page=${page}"
    # 헤더: Authorization / User-Agent
    if ! curl -fsSL -H "Authorization: token ${token}" -H "User-Agent: backup-script" \
         "$u" -o "${tmpdir}/p${page}.json"; then
      break
    fi
    # 빈 배열/객체인지 판단 (길이가 매우 짧으면 종료 가정)
    if [[ ! -s "${tmpdir}/p${page}.json" ]]; then
      break
    fi
    files+=("${tmpdir}/p${page}.json")
    # Link 헤더 파싱 위해 -I만 별도 호출 → rel="next" 없으면 종료
    local headers
    headers="$(curl -fsSIL -H "Authorization: token ${token}" -H "User-Agent: backup-script" "$u" 2>/dev/null || true)"
    if ! grep -qi 'rel="next"' <<<"$headers"; then
      break
    fi
    page=$((page+1))
  done
  if [[ "${#files[@]}" -eq 0 ]]; then
    echo "[]" >"$out"
  else
    # 여러 페이지 배열을 하나로 합치기
    jq -s 'add' "${files[@]}" >"$out"
  fi
  rm -rf -- "$tmpdir"
}

download_asset() {  # 릴리스 asset 다운로드
  local url="$1" token="$2" dst="$3"
  curl -fsSL -H "Authorization: token ${token}" -H "User-Agent: backup-script" \
    "$url" -o "$dst"
}

zip_folder() {  # 백업 폴더 압축: zip 있으면 zip, 없으면 tar.gz
  local src="$1" zipname="$2"
  if command -v zip >/dev/null 2>&1; then
    (cd "$(dirname "$src")" && zip -r "$zipname" "$(basename "$src")" >/dev/null)
    echo "ZIP created: $(join_path "$(dirname "$src")" "$zipname")"
  else
    (cd "$(dirname "$src")" && tar -czf "${zipname%.zip}.tar.gz" "$(basename "$src")")
    echo "TAR.GZ created: $(join_path "$(dirname "$src")" "${zipname%.zip}.tar.gz")"
  fi
}

# ---------- 인자 파싱 (한국어 주석) --------------------------------------------
REPO_URL=""
OUT_DIR="."
TOKEN="${GITHUB_TOKEN:-}"
ZIP_FLAG="no"
INCLUDE_DISCUSSIONS="no"

while [[ $# -gt 0 ]]; do
  case "$1" in
    --out-dir|-o) OUT_DIR="$2"; shift 2 ;;
    --token|-t)   TOKEN="$2"; shift 2 ;;
    --zip)        ZIP_FLAG="yes"; shift ;;
    --include-discussions) INCLUDE_DISCUSSIONS="yes"; shift ;;
    -h|--help)
      echo "Usage: $0 <RepoUrl> [--out-dir DIR] [--token TOKEN] [--zip] [--include-discussions]"
      exit 0 ;;
    *)
      if [[ -z "$REPO_URL" ]]; then
        REPO_URL="$1"; shift
      else
        echo "[ERROR] Unknown argument: $1" >&2; exit 1
      fi
      ;;
  esac
done

if [[ -z "$REPO_URL" ]]; then
  echo "Usage: $0 <RepoUrl> [--out-dir DIR] [--token TOKEN] [--zip] [--include-discussions]"
  exit 1
fi

# ---------- 사전 점검 (한국어 주석) --------------------------------------------
need_cmd git
need_cmd curl

REPO_NAME="$(url_basename_no_git "$REPO_URL")"
BACKUP_DIR="$(join_path "$OUT_DIR" "${REPO_NAME}_backup")"
CODE_BARE="$(join_path "$BACKUP_DIR" "${REPO_NAME}.git")"
WIKI_BARE="$(join_path "$BACKUP_DIR" "${REPO_NAME}.wiki.git")"

MODE="LITE"
if [[ -n "$TOKEN" ]]; then
  MODE="FULL"
fi

# 모드 안내 및 실행 확인
say_section "Detected mode: ${MODE}"
if [[ "$MODE" == "FULL" ]]; then
  echo "This will back up: Code+Wiki+Metadata (issues, issue comments, PRs, PR comments/reviews, labels, milestones, releases+assets)."
  # FULL 모드에는 jq 필요
  if ! command -v jq >/dev/null 2>&1; then
    echo "[ERROR] 'jq' is required in FULL mode. Please install jq (e.g., apt install jq, brew install jq)."
    exit 1
  fi
else
  echo "This will back up: Code+Wiki only (no metadata). Set GITHUB_TOKEN or --token to enable FULL mode."
fi
read -r -p "Proceed with this mode? [y/N]: " CONFIRM
case "$CONFIRM" in
  y|Y|yes|YES) ;;
  *) echo "Aborted. Nothing was changed."; exit 0 ;;
esac

# 백업 폴더 준비
say_section "[1/6] Preparing backup folder: ${BACKUP_DIR}"
ensure_clean_dir_with_prompt "$BACKUP_DIR"

# 코드 미러
say_section "[2/6] Backing up main repository (mirror: all branches/tags)"
if ! git clone --mirror "$REPO_URL" "$CODE_BARE"; then
  echo "[ERROR] Failed to back up the main repository."
  exit 1
fi

# 위키 확인/백업
say_section "[3/6] Checking Wiki"
WIKI_URL="$(build_wiki_url "$REPO_URL")"
echo "Wiki URL: ${WIKI_URL}"
if git ls-remote "$WIKI_URL" >/dev/null 2>&1; then
  echo "Wiki repository found. Backing up Wiki..."
  if ! git clone --mirror "$WIKI_URL" "$WIKI_BARE"; then
    echo "[WARNING] Failed to back up Wiki repository. Main repository backup succeeded."
  fi
else
  echo "Wiki repository not found or inaccessible. Skipping..."
fi

# 메타데이터 (FULL 모드)
if [[ "$MODE" == "FULL" ]]; then
  META_DIR="$(join_path "$BACKUP_DIR" "meta")"
  ISSUES_DIR="$(join_path "$META_DIR" "issues")"
  PULLS_DIR="$(join_path "$META_DIR" "pulls")"
  RELS_DIR="$(join_path "$META_DIR" "releases")"
  DISC_DIR="$(join_path "$META_DIR" "discussions")"

  mkdir -p "$META_DIR" "$ISSUES_DIR/comments" "$PULLS_DIR/comments" "$PULLS_DIR/reviews" "$RELS_DIR/assets"

  # 소유자/저장소 파싱
  # 예: https://github.com/owner/repo(.git)
  TMP="${REPO_URL%.git}"
  OWNER="$(echo "$TMP" | awk -F'github.com/' '{print $2}' | awk -F'/' '{print $1}')"
  REPO="$(echo "$TMP" | awk -F'github.com/' '{print $2}' | awk -F'/' '{print $2}')"

  say_section "[4/6] Exporting repository metadata (basic, labels, milestones)"
  curl -fsSL -H "Authorization: token ${TOKEN}" -H "User-Agent: backup-script" \
    "https://api.github.com/repos/${OWNER}/${REPO}" -o "${META_DIR}/repository.json"

  rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/labels" "$TOKEN" "${META_DIR}/labels.json"
  rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/milestones?state=all" "$TOKEN" "${META_DIR}/milestones.json"

  say_section "[5/6] Exporting issues and pull requests (with comments/reviews)"
  rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/issues?state=all&filter=all" "$TOKEN" "${ISSUES_DIR}/issues.json"

  # 이슈 번호 중 PR이 아닌 것만 추출하여 코멘트 저장
  mapfile -t ISSUE_NUMS < <(jq -r '.[] | select(.pull_request|not) | .number' "${ISSUES_DIR}/issues.json")
  for num in "${ISSUE_NUMS[@]}"; do
    rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/issues/${num}/comments" "$TOKEN" "${ISSUES_DIR}/comments/${num}.comments.json"
  done

  # PR 본문/코멘트/리뷰
  rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/pulls?state=all" "$TOKEN" "${PULLS_DIR}/pulls.json"
  mapfile -t PR_NUMS < <(jq -r '.[].number' "${PULLS_DIR}/pulls.json")
  for num in "${PR_NUMS[@]}"; do
    rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/pulls/${num}/comments" "$TOKEN" "${PULLS_DIR}/comments/${num}.comments.json"
    rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/issues/${num}/comments" "$TOKEN" "${PULLS_DIR}/comments/${num}.issue-comments.json"
    rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/pulls/${num}/reviews" "$TOKEN" "${PULLS_DIR}/reviews/${num}.reviews.json"
  done

  say_section "[6/6] Exporting releases (and assets)"
  rest_paged_to_file "https://api.github.com/repos/${OWNER}/${REPO}/releases" "$TOKEN" "${RELS_DIR}/releases.json"
  # 릴리스 자산 다운로드
  COUNT_ASSETS=$(jq 'map(.assets // []) | flatten | length' "${RELS_DIR}/releases.json")
  if [[ "$COUNT_ASSETS" -gt 0 ]]; then
    # 릴리스별 자산 저장 폴더 생성 후 다운로드
    mapfile -t REL_IDS < <(jq -r '.[].id' "${RELS_DIR}/releases.json")
    for rid in "${REL_IDS[@]}"; do
      mkdir -p "${RELS_DIR}/assets/${rid}"
      # 해당 릴리스의 자산만 추출
      mapfile -t ASSET_NAMES < <(jq -r ".[] | select(.id==${rid}) | .assets[]?.name" "${RELS_DIR}/releases.json")
      mapfile -t ASSET_URLS  < <(jq -r ".[] | select(.id==${rid}) | .assets[]?.browser_download_url" "${RELS_DIR}/releases.json")
      for ((i=0; i<${#ASSET_URLS[@]}; i++)); do
        an="${ASSET_NAMES[$i]}"; du="${ASSET_URLS[$i]}"
        [[ -z "$du" ]] && continue
        echo "Downloading asset: $an"
        download_asset "$du" "$TOKEN" "${RELS_DIR}/assets/${rid}/${an}"
      done
    done
  fi

  # Discussions (옵션)
  if [[ "$INCLUDE_DISCUSSIONS" == "yes" ]]; then
    say_section "[Extra] Exporting discussions (experimental)"
    mkdir -p "$DISC_DIR"
    read -r -d '' GQL <<'EOF'
{
  "query": "query($owner:String!,$repo:String!){ repository(owner:$owner,name:$repo){ discussions(first:100, orderBy:{field:CREATED_AT, direction:ASC}){ nodes{ id number title body createdAt author { login } url } pageInfo{ hasNextPage endCursor }}}}",
  "variables": { "owner": "__OWNER__", "repo": "__REPO__" }
}
EOF
    # 변수 치환
    GQL_PAYLOAD="${GQL/__OWNER__/${OWNER}}"
    GQL_PAYLOAD="${GQL_PAYLOAD/__REPO__/${REPO}}"
    curl -fsSL -H "Authorization: bearer ${TOKEN}" -H "User-Agent: backup-script" \
         -H "Content-Type: application/json" \
         -d "$GQL_PAYLOAD" \
         https://api.github.com/graphql > "${DISC_DIR}/discussions.json"
  fi
else
  say_section "[Info] Running in LITE mode (Code+Wiki only). Metadata is skipped."
fi

# 압축 옵션
if [[ "$ZIP_FLAG" == "yes" ]]; then
  zip_folder "$BACKUP_DIR" "$(basename "$BACKUP_DIR").zip"
fi

say_section "[Done] Backup finished (${MODE} mode)"
echo "Backup location: ${BACKUP_DIR}"
echo " - Repository (mirror): ${CODE_BARE}"
if [[ -d "$WIKI_BARE" ]]; then
  echo " - Wiki (mirror): ${WIKI_BARE}"
else
  echo " - Wiki: (none)"
fi
if [[ "$MODE" == "FULL" ]]; then
  echo " - Metadata: ${BACKUP_DIR}/meta"
else
  echo " - Metadata: (skipped)"
fi