AWSのS3　Glacierストレージクラスのファイルをダウンロードするスクリプトを書きました

pip install boto3

# -*- coding:utf-8 -*-

import boto3
import os
import re
import time
from datetime import datetime

TARGET_BUCKET_NAME = '[type your bucket name]'
TARGET_TYPE        = 'DIR' # or FILE
TARGET_DIR_NAME    = ''
TARGET_FILE_NAME   = ''
DOWNLOADED_DIR     = './downloaded/'
POLLING_TIME       = 7200 # 2.0h

s3     = boto3.resource('s3')
bucket = s3.Bucket(TARGET_BUCKET_NAME)

def main():
    print(datetime.now().strftime("%Y/%m/%d %H:%M:%S"))

    # DOWNLOADED_DIRの有無を確認し、なければ作成
    print('check and create download dir.')
    if os.path.exists(DOWNLOADED_DIR) is False:
        os.mkdir(DOWNLOADED_DIR)

    # グローバル変数チェック
    print('variables check.')
    if TARGET_BUCKET_NAME == '':
        raise ValueError('TARGET_BUCKET_NAMEが空です')

    if TARGET_TYPE not in ['DIR', 'FILE']:
        raise ValueError('TARGET_TYPEが正しくありません')

    if TARGET_TYPE == 'DIR' and (TARGET_DIR_NAME == '' or TARGET_DIR_NAME is None):
        raise ValueError('TARGET_DIR_NAMEが正しくありません')

    if TARGET_TYPE == 'FILE' and (TARGET_FILE_NAME == '' or TARGET_FILE_NAME is None):
        raise ValueError('TARGET_FILE_NAMEが正しくありません')


    # 指定条件に合致するオブジェクト（ファイル）の情報を取得
    print('search target files.')
    target_objs = {}
    for obj_sum in bucket.objects.all():
        obj           = bucket.Object(obj_sum.key)
        obj_dir_path  = re.sub(r'(.*)\/.*', r'\1', obj_sum.key)
        obj_first_dir = re.sub(r'\/.*', '', obj_sum.key)
        obj_file_name = re.sub(r'.*\/', '', obj_sum.key)

        is_target = False
        if TARGET_TYPE == 'DIR':
            if obj_first_dir == TARGET_DIR_NAME and obj_file_name is not '':
                is_target = True

        elif TARGET_TYPE == 'FILE':
            if obj_file_name == TARGET_FILE_NAME:
                is_target = True

        if is_target:
            target_objs[obj_sum.key] = {
                  'dir_path'      : obj_dir_path
                , 'storage_class' : obj.storage_class
                , 'status'        : ''
            }


    # 各ファイルのkeyに含まれるディレクトリ情報を抽出してDOWNLOADED_DIR内に作成
    dirs = [v['dir_path'] for k, v in target_objs.items() ]
    dirs = list(set(dirs))
    for d in dirs:
        os.makedirs(DOWNLOADED_DIR + d, exist_ok=True)


    # ダウンロードリクエスト & ポーリングでダウンロード
    print('start polling.')
    while is_complete(target_objs) is False:
        print('='*80)
        print(datetime.now().strftime("%Y/%m/%d %H:%M:%S"))

        for key, attrs in target_objs.items():
            if attrs['storage_class'] == 'GLACIER':
                
                obj = bucket.Object(key)
                if obj.restore is None:

                    # ここがrestoreのリクエスト部分
                    obj.restore_object(
                        RequestPayer   = 'requester',
                        RestoreRequest = {
                            'Days' : 1,
                            'GlacierJobParameters': {
                                'Tier' : 'Standard'
                            }
                        }
                    )

                elif 'ongoing-request="true"' in obj.restore:
                    attrs['status'] = 'in_progress'

                elif 'ongoing-request="false"' in obj.restore:
                    obj.download_file(DOWNLOADED_DIR + key)
                    attrs['status'] =  'complete'
            
            print('{:<150}{:<10}{:<10}'.format(key[:120], attrs['storage_class'], attrs['status']))

        if is_complete(target_objs) is False:
            print('sleep.')
            time.sleep(POLLING_TIME)
        else:
            continue

    print("END")
    print(datetime.now().strftime("%Y/%m/%d %H:%M:%S"))


def is_complete(objs):
    ret = True
    for k, v in objs.items():
        if v['status'] is not 'complete':
            ret = False
    return ret

if __name__ == '__main__':
    main()