def generate(self, sub_graph: AISubGraph, project_desc: ProjectDesc) -> KubernetesPythonJob:
if sub_graph.config.exec_mode == ExecutionMode.BATCH:
run_func = serialize(batch_run_func)
py_context = JobContext(ExecutionMode.BATCH)
else:
run_func = serialize(stream_run_func)
py_context = JobContext(ExecutionMode.STREAM)
py_context.project_config = project_desc.project_config
run_graph: RunGraph = self.build_run_graph(sub_graph, py_context)
job_config: KubernetesPythonJobConfig = sub_graph.config
return KubernetesPythonJob(run_graph=run_graph, run_func=run_func,
job_context=py_context, job_config=job_config)
def generate_job_resource(self, job: LocalFlinkJob) -> None:
"""
Generate flink job resource.
:param job: Local flink job.
"""
# gen config file
project_path = job.job_config.project_path
if project_path is None:
project_path = "/tmp"
project_path_temp = project_path + "/temp"
if not os.path.exists(project_path_temp):
os.mkdir(project_path_temp)
if job.job_config.language_type == LanguageType.JAVA:
execution_config_file = project_path_temp + '/job_execution_config_' + str(
uuid.uuid4()) + "_" + job.instance_id
with open(execution_config_file, 'w') as f:
f.write(dumps(job))
job.config_file = execution_config_file
else:
execution_config_file = '/job_execution_config_' + str(
uuid.uuid4()) + "_" + job.instance_id
real_execution_config_file = project_path_temp + execution_config_file
with open(real_execution_config_file, 'wb') as f:
f.write(serialization_utils.serialize(job))
job.config_file = execution_config_file
def generate_job_resource(self, job: VVPJob) -> None:
"""
Generate flink job resource.
:param job: Local flink job.
"""
project_path = job.job_config.project_path
if project_path is None:
project_path = "/tmp"
project_path_temp = project_path + "/temp"
if not os.path.exists(project_path_temp):
os.mkdir(project_path_temp)
execution_config_file = project_path_temp + '/job_execution_config_' + str(
uuid.uuid4()) + "_" + job.instance_id
job.config_file = execution_config_file
with open(execution_config_file, 'wb') as f:
f.write(serialization_utils.serialize(job))
script_path = version.vvp_job_main_file
entry_module_path = job.job_config.project_desc.project_config[
'entry_module_path']
python3_location = sys.executable
cmd = [
python3_location, script_path, job.job_config.project_path,
execution_config_file, entry_module_path
]
job.exec_cmd = ' '.join(cmd)
logging.info(job.exec_cmd)
def __init__(self, python_func: object, node_id: Text = None) -> None:
"""
Set the user defined python function as member variables.
:param python_func: User defined python function.
:param node_id: Id of node.
"""
super().__init__("python_func", node_id)
self.python_func: bytes = serialize(python_func)
def __init__(self, python_object: object, node_id: Text = None) -> None:
"""
Construct of FlinkPythonExecutor.
:param python_object: Object includes the python executor.
:param node_id: Id of node.
"""
super().__init__(python_object=python_object, node_id=node_id)
self.python_object: bytes = serialize(python_object)
def __init__(self, platform: Text):
super().__init__(platform=platform, engine=FlinkEngine.engine())
self.flink_home = None
self.jm_host_port = 'localhost:8081'
self.class_path = None
self.py_entry_file = None
self.py_files = None
self.py_module = None
self.jar_path = None
self.args = []
self.language_type: LanguageType = LanguageType.PYTHON
self.table_env_create_func: bytes = serialization_utils.serialize(
TableEnvCreator())
def build_run_graph(self, sub_graph: AISubGraph,
context: JobContext) -> RunGraph:
run_graph = RunGraph()
processed_nodes = set()
node_list: List[AINode] = []
for n in sub_graph.nodes.values():
node_list.append(n)
for e in sub_graph.edges:
data_channel_list = []
for c in sub_graph.edges[e]:
cc: DataEdge = c
data_channel_list.append(cc)
run_graph.dependencies[e] = data_channel_list
node_size = len(sub_graph.nodes)
processed_size = len(processed_nodes)
while processed_size != node_size:
p_nodes = []
for i in range(len(node_list)):
if node_list[i].instance_id in sub_graph.edges:
flag = True
for c in sub_graph.edges[node_list[i].instance_id]:
if c.target_node_id in processed_nodes:
pass
else:
flag = False
break
else:
flag = True
if flag:
p_nodes.append(node_list[i])
if 0 == len(p_nodes):
raise Exception("graph has circle!")
for n in p_nodes:
run_graph.nodes.append(n)
run_graph.executor_bytes.append(
serialize(self.compile_node(n, context=context)))
node_list.remove(n)
processed_nodes.add(n.instance_id)
processed_size = len(processed_nodes)
return run_graph
def generate_job_resource(self, job: LocalPythonJob):
se_func = job.run_func
se_args = serialize(
RunArgs(run_graph=job.run_graph, job_context=job.job_context))
project_path = job.job_config.project_path
if project_path is None:
project_path = "/tmp"
project_path_temp = project_path + "/temp"
if not os.path.exists(project_path_temp):
os.mkdir(project_path_temp)
func_file = project_path_temp + '/tmp_func' + str(
uuid.uuid4()) + job.instance_id
args_file = project_path_temp + '/tmp_args' + str(
uuid.uuid4()) + job.instance_id
with open(func_file, 'wb') as f:
f.write(se_func)
with open(args_file, 'wb') as f:
f.write(se_args)
job.exec_func_file: Text = Path(func_file).name
job.exec_args_file: Text = Path(args_file).name
def generate_job_resource(self, job: KubernetesFlinkJob) -> None:
"""
Generate kubernetes flink job resource.
:param job: Kubernetes flink job.
"""
# gen config file
project_path = job.job_config.project_path
if project_path is None:
project_path = "/tmp"
project_path_temp = project_path + "/temp"
if not os.path.exists(project_path_temp):
os.mkdir(project_path_temp)
execution_config_file = 'job_execution_config_' + str(
uuid.uuid4()) + "_" + job.instance_id
real_execution_config_file = project_path_temp + '/' + execution_config_file
if job.job_config.language_type == LanguageType.JAVA:
with open(real_execution_config_file, 'w') as f:
f.write(dumps(job))
job.config_file = Path(real_execution_config_file).name
else:
with open(real_execution_config_file, 'wb') as f:
f.write(serialization_utils.serialize(job))
job.config_file = execution_config_file
# generate python_codes.zip
python_codes = '{}/python_codes'.format(project_path)
if os.path.exists(python_codes):
zip_dir = '{}/zip'.format(project_path)
if os.path.exists(zip_dir):
shutil.rmtree(zip_dir)
shutil.copytree(python_codes, zip_dir + '/python_codes')
shutil.make_archive(python_codes, 'zip', zip_dir)
def set_table_env_create_func(self, func):
self.table_env_create_func = serialization_utils.serialize(func)
def generate_job_resource(self, job: LocalFlinkJob) -> None:
"""
Generate flink job resource.
:param job: Local flink job.
"""
# gen config file
project_path = job.job_config.project_path
if project_path is None:
project_path = "/tmp"
project_path_temp = project_path + "/temp"
if not os.path.exists(project_path_temp):
os.mkdir(project_path_temp)
if job.job_config.language_type == LanguageType.JAVA:
execution_config_file = project_path_temp + '/job_execution_config_' + str(
uuid.uuid4()) + "_" + job.instance_id
with open(execution_config_file, 'w') as f:
f.write(dumps(job))
job.config_file = execution_config_file
exec_cmd = ['flink', 'run']
exec_cmd.extend(['-m', job.job_config.jm_host_port])
if job.job_config.class_path is not None:
exec_cmd.extend(['-C', job.job_config.class_path])
if job.job_config.project_desc.jar_dependencies is not None:
for jar in job.job_config.project_desc.jar_dependencies:
exec_cmd.extend(['-C', "file://{}".format(jar)])
if job.job_config.main_class is not None:
exec_cmd.extend(['-c', job.job_config.main_class])
exec_cmd.extend([job.job_config.jar_path])
exec_cmd.extend(['--execution-config', execution_config_file])
if job.job_config.args is not None:
exec_cmd.extend(job.job_config.args)
else:
if 'entry_module_path' not in job.job_config.project_desc.project_config:
entry_module_path = (file_path_to_absolute_module(
sys.argv[0])).split('.')[-1]
else:
entry_module_path = job.job_config.project_desc.project_config[
'entry_module_path']
execution_config_file = '/job_execution_config_' + str(
uuid.uuid4()) + "_" + job.instance_id
real_execution_config_file = project_path_temp + execution_config_file
with open(real_execution_config_file, 'wb') as f:
f.write(serialization_utils.serialize(job))
job.config_file = execution_config_file
python3_location = sys.executable
if job.job_config.local_mode == 'python':
exec_cmd = [
python3_location, version.py_main_file,
job.job_config.project_path, execution_config_file,
entry_module_path
]
else:
exec_cmd = [
'flink', 'run', '-pym', version.py_cluster_main_file,
'-pyfs', job.job_config.project_path + ',' +
job.job_config.project_path + '/python_codes/', '-pyexec',
python3_location, '--project-path',
job.job_config.project_path, '--config-file',
execution_config_file, '--entry-module-path',
entry_module_path
]
job.exec_cmd = exec_cmd
logging.info(' '.join(exec_cmd))
