def test_generate_cmd_with_partitioner(self):
_job_name = "test_streaming_job_%s" % uuid.uuid4()
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-D map.output.key.field.separator=| ' \
'-D mapreduce.partition.keypartitioner.options=-k1,2 ' \
'-mapper mapper.py ' \
'-reducer reducer.py ' \
'-numReduceTasks 0 ' \
'-input data ' \
'-output output.txt ' \
'-partitioner org.apache.hadoop.mapred.lib.KeyFieldBasedPartitioner'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'
.format(os.path.dirname(os.path.realpath(__file__))),
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).take('data').process_with(mapper='mapper.py', reducer='reducer.py', reducer_num=0).use(
partitioner='org.apache.hadoop.mapred.lib.KeyFieldBasedPartitioner'
).save('output.txt').with_config_option(
key='map.output.key.field.separator',
value='|').with_config_option(
key='mapreduce.partition.keypartitioner.options',
value='-k1,2').run()
def test_streaming_map_only_job_generation(self):
_config_file = os.path.join(
os.path.dirname(os.path.realpath(__file__)), 'resources',
'mapreduce', 'mapreduce_streaming_job.ini')
metastore = IniFileMetaStore(file=_config_file)
_config = Configuration.load(metastore=metastore)
_job_name = 'streaming_test_job_map_only'
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-D value.delimiter.char=, ' \
'-D partition.to.process=20142010 ' \
'-mapper smapper.py ' \
'-reducer NONE ' \
'-numReduceTasks 0 ' \
'-input /raw/20102014 ' \
'-output /core/20102014'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'.format(
os.path.dirname(os.path.realpath(__file__))),
config=_config,
name=_job_name,
executor=self.assert_generated_command(_expected_command)).run()
def test_generate_job_cmd_with_config_injections(self):
_job_name = "test_streaming_job_%s" % uuid.uuid4()
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-D value.delimiter.char=, ' \
'-D partition.to.process=20142010 ' \
'-mapper mapper.py ' \
'-reducer reducer.py ' \
'-input data ' \
'-output output.txt'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'
.format(os.path.dirname(os.path.realpath(__file__))),
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).take(
'data'
).process_with(
mapper='mapper.py',
reducer='reducer.py'
).save(
'output.txt'
).with_config_option(
key='value.delimiter.char',
value=','
).with_config_option(
key='partition.to.process',
value='20142010'
).run()
def test_streaming_map_only_job_generation(self):
_config_file = os.path.join(
os.path.dirname(os.path.realpath(__file__)),
'resources',
'mapreduce',
'mapreduce_streaming_job.ini')
metastore = IniFileMetaStore(file=_config_file)
_config = Configuration.load(metastore=metastore)
_job_name = 'streaming_test_job_map_only'
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-D value.delimiter.char=, ' \
'-D partition.to.process=20142010 ' \
'-mapper smapper.py ' \
'-reducer NONE ' \
'-numReduceTasks 0 ' \
'-input /raw/20102014 ' \
'-output /core/20102014'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'
.format(os.path.dirname(os.path.realpath(__file__))),
config=_config,
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).run()
def test_streaming_job_with_multiple_inputs(self):
_job_name = 'streaming_test_job_with_multiple_inputs'
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-files dim1.txt ' \
'-libjars mr_001.jar,mr_002.jar ' \
'-mapper smapper.py ' \
'-reducer sreducer.py ' \
'-numReduceTasks 100 ' \
'-input /raw/20102014 ' \
'-input /raw/21102014 ' \
'-input /raw/22102014 ' \
'-output /core/20102014 ' \
'-inputformat \'org.mr.CustomInputFormat\' ' \
'-outputformat \'org.mr.CustomOutputFormat\' ' \
'-cmdenv JAVA_HOME=/java ' \
'-cmdenv tmp.dir=/tmp/streaming_test_job_with_multiple_inputs'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'.format(
os.path.dirname(os.path.realpath(__file__))),
config=self._config,
name=_job_name,
executor=self.assert_generated_command(_expected_command)).run()
def test_generate_cmd_with_partitioner(self):
_job_name = "test_streaming_job_%s" % uuid.uuid4()
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-D map.output.key.field.separator=| ' \
'-D mapreduce.partition.keypartitioner.options=-k1,2 ' \
'-mapper mapper.py ' \
'-reducer reducer.py ' \
'-numReduceTasks 0 ' \
'-input data ' \
'-output output.txt ' \
'-partitioner org.apache.hadoop.mapred.lib.KeyFieldBasedPartitioner'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'.format(
os.path.dirname(os.path.realpath(__file__))),
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).take('data').process_with(
mapper='mapper.py', reducer='reducer.py', reducer_num=0
).use(
partitioner='org.apache.hadoop.mapred.lib.KeyFieldBasedPartitioner'
).save('output.txt').with_config_option(
key='map.output.key.field.separator',
value='|').with_config_option(
key='mapreduce.partition.keypartitioner.options',
value='-k1,2').run()
def test_streaming_job_with_multiple_inputs(self):
_job_name = 'streaming_test_job_with_multiple_inputs'
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-files dim1.txt ' \
'-libjars mr_001.jar,mr_002.jar ' \
'-mapper smapper.py ' \
'-reducer sreducer.py ' \
'-numReduceTasks 100 ' \
'-input /raw/20102014 ' \
'-input /raw/21102014 ' \
'-input /raw/22102014 ' \
'-output /core/20102014 ' \
'-inputformat \'org.mr.CustomInputFormat\' ' \
'-outputformat \'org.mr.CustomOutputFormat\' ' \
'-cmdenv JAVA_HOME=/java ' \
'-cmdenv tmp.dir=/tmp/streaming_test_job_with_multiple_inputs'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'
.format(os.path.dirname(os.path.realpath(__file__))),
config=self._config,
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).run()
def test_streaming_job_generation(self):
_job_name = 'streaming_test_job'
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-mapper smapper.py ' \
'-reducer sreducer.py ' \
'-input /raw/20102014 ' \
'-output /core/20102014'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'.format(
os.path.dirname(os.path.realpath(__file__))),
config=self._config,
name=_job_name,
executor=self.assert_generated_command(_expected_command)).run()
def test_streaming_job_generation(self):
_job_name = 'streaming_test_job'
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-mapper smapper.py ' \
'-reducer sreducer.py ' \
'-input /raw/20102014 ' \
'-output /core/20102014'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'
.format(os.path.dirname(os.path.realpath(__file__))),
config=self._config,
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).run()
def test_streaming_job_without_reducers(self):
_job_name = "test_streaming_job_%s" % uuid.uuid4()
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-mapper mapper.py ' \
'-reducer NONE ' \
'-numReduceTasks 0 ' \
'-input data ' \
'-output output.txt'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'.format(
os.path.dirname(os.path.realpath(__file__))),
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).take("data").map_with(
mapper="mapper.py").disable_reducers().save("output.txt").run()
def test_load_streaming_job_with_config_injections(self):
_job_name = 'streaming_test_job_with_custom_configurations'
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-D value.delimiter.char=, ' \
'-D partition.to.process=20142010 ' \
'-mapper smapper.py ' \
'-reducer sreducer.py ' \
'-input /raw/20102014 ' \
'-output /core/20102014'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'.format(
os.path.dirname(os.path.realpath(__file__))),
config=self._config,
name=_job_name,
executor=self.assert_generated_command(_expected_command)).run()
def test_load_streaming_job_with_config_injections(self):
_job_name = 'streaming_test_job_with_custom_configurations'
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-D value.delimiter.char=, ' \
'-D partition.to.process=20142010 ' \
'-mapper smapper.py ' \
'-reducer sreducer.py ' \
'-input /raw/20102014 ' \
'-output /core/20102014'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'
.format(os.path.dirname(os.path.realpath(__file__))),
config=self._config,
name=_job_name,
executor=self.assert_generated_command(_expected_command)).run()
def test_streaming_job(self):
_job_name = "test_streaming_job_%s" % uuid.uuid4()
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-mapper mapper.py ' \
'-reducer reducer.py ' \
'-input data ' \
'-output output.txt'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'
.format(os.path.dirname(os.path.realpath(__file__))),
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).take(
"data"
).process_with(
mapper="mapper.py",
reducer="reducer.py").save("output.txt").run()
def test_generate_cmd_with_input_output_format(self):
_job_name = "test_streaming_job_%s" % uuid.uuid4()
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-mapper mapper.py ' \
'-reducer reducer.py ' \
'-numReduceTasks 10 ' \
'-input data ' \
'-output output.txt ' \
'-inputformat org.apache.hadoop.mapred.KeyValueTextInputFormat ' \
'-outputformat org.apache.hadoop.mapred.SequenceFileOutputFormat'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'
.format(os.path.dirname(os.path.realpath(__file__))),
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).process_with(mapper='mapper.py', reducer='reducer.py', reducer_num=10).use(
inputformat='org.apache.hadoop.mapred.KeyValueTextInputFormat',
outputformat='org.apache.hadoop.mapred.SequenceFileOutputFormat'
).take('data').save('output.txt').run()
def test_generate_cmd_with_input_output_format(self):
_job_name = "test_streaming_job_%s" % uuid.uuid4()
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-mapper mapper.py ' \
'-reducer reducer.py ' \
'-numReduceTasks 10 ' \
'-input data ' \
'-output output.txt ' \
'-inputformat org.apache.hadoop.mapred.KeyValueTextInputFormat ' \
'-outputformat org.apache.hadoop.mapred.SequenceFileOutputFormat'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'.format(
os.path.dirname(os.path.realpath(__file__))),
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).process_with(
mapper='mapper.py', reducer='reducer.py', reducer_num=10
).use(inputformat='org.apache.hadoop.mapred.KeyValueTextInputFormat',
outputformat='org.apache.hadoop.mapred.SequenceFileOutputFormat'
).take('data').save('output.txt').run()
def test_generate_job_cmd_with_config_injections(self):
_job_name = "test_streaming_job_%s" % uuid.uuid4()
_expected_command = 'hadoop jar ' \
'{0}/resources/mapreduce/hadoop-streaming.jar ' \
'-D mapreduce.job.name={1} ' \
'-D value.delimiter.char=, ' \
'-D partition.to.process=20142010 ' \
'-mapper mapper.py ' \
'-reducer reducer.py ' \
'-input data ' \
'-output output.txt'\
.format(os.path.dirname(os.path.realpath(__file__)),
_job_name)
MapReduce.prepare_streaming_job(
jar='{0}/resources/mapreduce/hadoop-streaming.jar'.format(
os.path.dirname(os.path.realpath(__file__))),
name=_job_name,
executor=self.assert_generated_command(_expected_command)
).take('data').process_with(
mapper='mapper.py',
reducer='reducer.py').save('output.txt').with_config_option(
key='value.delimiter.char',
value=',').with_config_option(key='partition.to.process',
value='20142010').run()
def _template_streaming_job_(self, base_dir="/tmp", map_only_job=False):
if not os.path.exists(HADOOP_STREAMING_JAR):
self.skip("Cannot allocate %s" % HADOOP_STREAMING_JAR)
_hdfs_basdir = HDFS(base_dir)
if not _hdfs_basdir.exists():
_hdfs_basdir.create_directory()
_job_input = HDFS(os.path.join(_hdfs_basdir.path, "input"))
_job_input.create_directory()
_job_output = HDFS(os.path.join(_hdfs_basdir.path, "output"))
home = os.path.dirname(__file__)
_mapper = os.path.join(home, 'resources', 'mapreduce', 'mapper.py')
_reducer = os.path.join(home, 'resources', 'mapreduce', 'reducer.py')
LocalFS(
os.path.join(os.path.dirname(__file__), 'resources', 'mapreduce', 'raw-data.txt')
).copy_to_hdfs(
_job_input.path
)
return MapReduce.prepare_streaming_job(name="test-mr-streaming-job{}".format(str(uuid.uuid4())), jar=HADOOP_STREAMING_JAR) \
.take(_job_input.path) \
.process_with(mapper=_mapper, reducer=None if map_only_job else _reducer) \
.save(_job_output.path)