ClickHouse/utils/s3tools/s3uploader

#!/usr/bin/env python
# -*- coding: utf-8 -*-
import os
import logging
import argparse
import tarfile
import math

try:
    from boto.s3.connection import S3Connection
except ImportError:
    raise ImportError("You have to install boto package 'pip install boto'")


class S3API(object):
    def __init__(self, access_key, secret_access_key, mds_api, mds_url):
        self.connection = S3Connection(
            host=mds_api,
            aws_access_key_id=access_key,
            aws_secret_access_key=secret_access_key,
        )
        self.mds_url = mds_url

    def upload_file(self, bucket_name, file_path, s3_path):
        logging.info("Start uploading file to bucket %s", bucket_name)
        bucket = self.connection.get_bucket(bucket_name)
        key = bucket.initiate_multipart_upload(s3_path)
        logging.info("Will upload to s3 path %s", s3_path)
        chunksize = 1024 * 1024 * 1024 # 1 GB
        filesize = os.stat(file_path).st_size
        logging.info("File size if %s", filesize)
        chunkcount = int(math.ceil(filesize / chunksize))

        def call_back(x, y):
            print "Uploaded {}/{} bytes".format(x, y)
        try:
            for i in range(chunkcount + 1):
                logging.info("Uploading chunk %s of %s", i, chunkcount + 1)
                offset = chunksize * i
                bytes_size = min(chunksize, filesize - offset)
                with open(file_path, 'r') as fp:
                    fp.seek(offset)
                    key.upload_part_from_file(fp=fp, part_num=i+1,
                                              size=bytes_size, cb=call_back,
                                              num_cb=100)
            key.complete_upload()
        except Exception as ex:
            key.cancel_upload()
            raise ex
        logging.info("Contents were set")
        return "https://{bucket}.{mds_url}/{path}".format(
            bucket=bucket_name, mds_url=self.mds_url, path=s3_path)


def make_tar_file_for_table(clickhouse_data_path, db_name, table_name,
                            tmp_prefix):

    relative_data_path = os.path.join('data', db_name, table_name)
    relative_meta_path = os.path.join('metadata', db_name, table_name + '.sql')
    path_to_data = os.path.join(clickhouse_data_path, relative_data_path)
    path_to_metadata = os.path.join(clickhouse_data_path, relative_meta_path)
    temporary_file_name = tmp_prefix + '/{tname}.tar'.format(tname=table_name)
    with tarfile.open(temporary_file_name, "w") as bundle:
        bundle.add(path_to_data, arcname=relative_data_path)
        bundle.add(path_to_metadata, arcname=relative_meta_path)
    return temporary_file_name


USAGE_EXAMPLES = '''
examples:
\ts3uploader --dataset-name some_ds --access-key-id XXX --secret-access-key YYY --clickhouse-data-path /opt/clickhouse/ --table-name default.some_tbl --bucket-name some-bucket
\ts3uploader --dataset-name some_ds --access-key-id XXX --secret-access-key YYY --file-name some_ds.tsv.xz --bucket-name some-bucket
'''

if __name__ == "__main__":
    logging.basicConfig(level=logging.INFO, format='%(asctime)s %(message)s')

    parser = argparse.ArgumentParser(
        description="Simple tool for uploading datasets to clickhouse S3",
        usage='%(prog)s [options] {}'.format(USAGE_EXAMPLES))
    parser.add_argument('--s3-api-url', default='s3.mds.yandex.net')
    parser.add_argument('--s3-common-url', default='s3.yandex.net')
    parser.add_argument('--bucket-name', default='clickhouse-datasets')
    parser.add_argument('--dataset-name', required=True,
                        help='Name of dataset, will be used in uploaded path')
    parser.add_argument('--access-key-id', required=True)
    parser.add_argument('--secret-access-key', required=True)
    parser.add_argument('--clickhouse-data-path',
                        default='/var/lib/clickhouse/',
                        help='Path to clickhouse database on filesystem')
    parser.add_argument('--s3-path', help='Path in s3, where to upload file')
    parser.add_argument('--tmp-prefix', default='/tmp',
                        help='Prefix to store temporay downloaded file')
    data_group = parser.add_mutually_exclusive_group(required=True)
    data_group.add_argument('--table-name',
                            help='Name of table with database, if you are uploading partitions')
    data_group.add_argument('--file-path',
                            help='Name of file, if you are uploading')
    args = parser.parse_args()

    if args.table_name is not None and args.clickhouse_data_path is None:
        raise argparse.ArgumentError(
            "You should specify --clickhouse-data-path to upload --table")

    s3_conn = S3API(
        args.access_key_id, args.secret_access_key,
        args.s3_api_url, args.s3_common_url)

    if args.table_name is not None:
        if '.' not in args.table_name:
            db_name = 'default'
        else:
            db_name, table_name = args.table_name.split('.')
        file_path = make_tar_file_for_table(
            args.clickhouse_data_path, db_name, table_name, args.tmp_prefix)
    else:
        file_path = args.file_path

    if 'tsv' in file_path:
        s3_path = os.path.join(
            args.dataset_name, 'tsv', os.path.basename(file_path))
    elif args.table_name is not None:
        s3_path = os.path.join(
            args.dataset_name, 'partitions', os.path.basename(file_path))
    elif args.s3_path is not None:
        s3_path = os.path.join(
            args.dataset_name, s3_path, os.path.base_name(file_path))
    else:
        raise Exception("Don't know s3-path to upload")

    url = s3_conn.upload_file(args.bucket_name, file_path, s3_path)
    logging.info("Data uploaded: %s", url)