def __init__()

in petastorm/hdfs/namenode.py [0:0]

• 23 lines of code
• 5 McCabe index (conditional complexity)

    def __init__(self, hadoop_configuration=None):
        """
        Sets the given HadoopConfiguration object for the resolver; or check for and pull hadoop
        configuration from an environment variable, in below preferred order to check.

        :param hadoop_configuration: an optional ``HadoopConfiguration``
        """
        self._hadoop_env = None
        self._hadoop_path = None
        if hadoop_configuration is None:
            # Pull from environment variable, in this preferred order
            for env in ["HADOOP_HOME", "HADOOP_PREFIX", "HADOOP_INSTALL"]:
                # Use the first available
                if env in os.environ:
                    self._hadoop_env = env
                    self._hadoop_path = os.environ[env]
                    hadoop_configuration = {}
                    self._load_site_xml_into_dict(
                        '{}/etc/hadoop/hdfs-site.xml'.format(self._hadoop_path),
                        hadoop_configuration)
                    self._load_site_xml_into_dict(
                        '{}/etc/hadoop/core-site.xml'.format(self._hadoop_path),
                        hadoop_configuration)
                    break
            if hadoop_configuration is None:
                # ensures at least an empty dict so no further checks required in member functions
                logger.warning('Unable to populate a sensible HadoopConfiguration for namenode resolution!\n'
                               'Path of last environment var (%s) tried [%s]. Please set up your Hadoop and \n'
                               'define environment variable HADOOP_HOME to point to your Hadoop installation path.',
                               self._hadoop_env, self._hadoop_path)
                hadoop_configuration = {}
        self._hadoop_configuration = hadoop_configuration