Class CassandraStorage

join/LSH/datasketch/storage.py:701–798

Storage implementation using Cassandra. Note: like other implementations, each storage has its own client. Unlike other implementations, all storage instances share one session and can potentially share the same buffer.

    class CassandraStorage(object):
        """
        Storage implementation using Cassandra.

        Note: like other implementations, each storage has its own client. Unlike other
            implementations, all storage instances share one session and can potentially share the
            same buffer.
        """

        DEFAULT_BUFFER_SIZE = 5000

        def __init__(self, config, name=None, buffer_size=None):
            """
            Constructor.

            :param dict[str, any] config: configuration following the following format:
                {
                    'basename': b'test',
                    'type': 'cassandra',
                    'cassandra': {
                        'seeds': ['127.0.0.1'],
                        'keyspace': 'lsh_test',
                        'replication': {
                            'class': 'SimpleStrategy',
                            'replication_factor': '1'
                        },
                        'drop_keyspace': True,
                        'drop_tables': True,
                        'shared_buffer': False,
                    }
                }
            :param bytes name: the name
            :param int buffer_size: the buffer size
            """
            self._config = config
            if buffer_size is None:
                buffer_size = CassandraStorage.DEFAULT_BUFFER_SIZE
            cassandra_param = self._parse_config(self._config['cassandra'])
            self._name = name if name else _random_name(11).decode('utf-8')
            self._buffer_size = buffer_size
            self._client = CassandraClient(cassandra_param, name, self._buffer_size)

        @staticmethod
        def _parse_config(config):
            """
            Parse a configuration dictionary, optionally fetching data from env variables.

            :param dict[str, any] config: the configuration
            :rtype: dict[str, str]
            :return: the parse configuration
            """
            cfg = {}
            for key, value in config.items():
                if isinstance(value, dict):
                    if 'env' in value:
                        value = os.getenv(value['env'], value.get('default', None))
                cfg[key] = value
            return cfg