Python SparkFiles.getRootDirectory方法代码示例

# 需要导入模块: from pyspark import SparkFiles [as 别名]
# 或者: from pyspark.SparkFiles import getRootDirectory [as 别名]
def train_partition(idx, iterator):
    port = 50000 + idx % 256
    main = SparkFiles.get("main.py")
    architecture = SparkFiles.get("train_val.prototxt")
    model = SparkFiles.get("deepq16.caffemodel")
    solver = SparkFiles.get("solver.prototxt")
    root = SparkFiles.getRootDirectory()
    dset = os.path.join(root, "dset-%02d.hdf5" % idx)

    flag_file = "flags/__BARISTA_READY__.%d" % port
    if os.path.isfile(flag_file):
        os.remove(flag_file)

    #  out = open(os.path.join(root, "barista.log"), 'w')
    subprocess.Popen(["python", main, architecture, model,
                      "--dataset", dset,
                      "--solver", solver,
                      "--dset-size", "30000",
                      "--initial-replay", "20000",
                      "--debug",
                      "--overwrite",
                      "--port", str(port)])

    while not os.path.isfile(flag_file):
        pass

    for step in iterator:
        dc = DummyClient("127.0.0.1", port)
        dc.send(barista.GRAD_UPDATE)
        response = dc.recv()
        yield response

# 需要导入模块: from pyspark import SparkFiles [as 别名]
# 或者: from pyspark.SparkFiles import getRootDirectory [as 别名]
def spawn_barista(partition):
    main = SparkFiles.get("main.py")
    architecture = SparkFiles.get("train_val.prototxt")
    model = SparkFiles.get("deepq16.caffemodel")
    solver = SparkFiles.get("solver.prototxt")
    root = SparkFiles.getRootDirectory()
    dset = os.path.join(root, "dset.hdf5")
    flag_file = "flags/__BARISTA_READY__"
    if os.path.isfile(flag_file):
        os.remove("flags/__BARISTA_READY__")

    out = open(os.path.join(root, "barista.log"), 'w')
    subprocess.Popen(["python", main, architecture, model,
                      "--dataset", dset,
                      "--solver", solver],
                     stdout=out,
                     stderr=subprocess.STDOUT)

    while not os.path.isfile("flags/__BARISTA_READY__"):
        pass

# 需要导入模块: from pyspark import SparkFiles [as 别名]
# 或者: from pyspark.SparkFiles import getRootDirectory [as 别名]
def start_spark(app_name='my_spark_app', master='local[*]', jar_packages=[],
                files=[], spark_config={}):
    """Start Spark session, get the Spark logger and load config files.
    Start a Spark session on the worker node and register the Spark
    application with the cluster. NOTE - only the app_name argument
    will apply when this is called from a script sent to spark-submit
    (i.e. when __name__ = '__main__'). All other arguments exist solely
    for testing the script from within an interactive Python console.
    This function also looks for a file ending in 'config.json' that
    can be sent with the Spark job. If it is found, it is opened,
    the contents parsed (assuming it contains valid JSON for the ETL job
    configuration), into a dict of ETL job configuration parameters,
    which are returned as the last element in the tuple returned by
    this function. If the file cannot be found then the return tuple
    only contains the Spark session and Spark logger objects.
    :param app_name: Name of Spark app.
    :param master: Cluster connection details (defaults to local[*].
    :param jar_packages: List of Spark JAR package names.
    :param files: List of files to send to Spark cluster (master and
    workers).
    :param spark_config: Dictionary of config key-value pairs.
    :return: A tuple of references to the Spark session, logger and
    config dict (only if available).
    """
    if __name__ == '__main__':
        # get Spark session factory
        spark_builder = (
            SparkSession
            .builder
            .appName(app_name))
    else:
        # get Spark session factory
        spark_builder = (
            SparkSession
            .builder
            .master(master)
            .appName(app_name))

        # create Spark JAR packages string
        spark_jars_packages = ','.join(list(jar_packages))
        spark_builder.config('spark.jars.packages', spark_jars_packages)

        spark_files = ','.join(list(files))
        spark_builder.config('spark.files', spark_files)

        # add other config params
        for key, val in spark_config.items():
            spark_builder.config(key, val)

    # create session and retrieve Spark logger object
    spark_sess = spark_builder.getOrCreate()
    spark_logger = logging.Log4j(spark_sess)

    # get config file if sent to cluster with --files
    spark_files_dir = SparkFiles.getRootDirectory()
    config_files = [filename
                    for filename in listdir(spark_files_dir)
                    if filename.endswith('config.json')]

    if len(config_files) != 0:
        path_to_config_file = path.join(spark_files_dir, config_files[0])
        with open(path_to_config_file, 'r') as config_file:
            config_json = config_file.read().replace('\n', '')
        config_dict = loads(config_json)
        spark_logger.warn('loaded config from ' + config_files[0])
    else:
        config_dict = None

    # build return tuple conditional on presence of config
    if config_dict is not None:
        return_tup = spark_sess, spark_logger, config_dict
    else:
        return_tup = spark_sess, spark_logger
    return return_tup

# 需要导入模块: from pyspark import SparkFiles [as 别名]
# 或者: from pyspark.SparkFiles import getRootDirectory [as 别名]
def __ls(broadcast_vars, iterator):
    """
    Get the list of files in the worker-local directory
    """
    return [__get_hostname(), os.listdir(SparkFiles.getRootDirectory())]