1. 使用举例

您必须替换代码中的image_id，access_key_id，access_key_secret等用户信息以及oss相关的路径。

import time
from batchcompute import Cient, ClientError
from batchcompute import CN_SHENZHEN as REGION
from batchcompute.resources import (
    JobDescription, TaskDescription, DAG,
    GroupDescription, ClusterDescription,
)
# some other codes here
access_key_id = ... # your access key id
access_key_secret = ... # your access key secret
image_id = ... # the id of a image created before
instance_type = ... # instance type
client = Client(REGION, access_key_id, access_key_secret)
try: 
    # Create cluster.
    cluster_desc = ClusterDescription()
    group_desc = GroupDescription()
    group_desc.DesiredVMCount = 1
    group_desc.InstanceType = instance_type
    cluster_desc.add_group('group1', group_desc)
    cluster_desc.Name = "BatchcomputeCluster" 
    cluster_desc.ImageId = image_id 
    cluster_desc.Description = "Python SDK test" 
    cluster_id = client.create_cluster(cluster_desc).Id
    # Create job.
    job_desc = JobDescription()
    echo_task = TaskDescription()
    # Create map task.
    echo_task.Parameters.Command.CommandLine = "echo Batchcompute service"
    echo_task.Parameters.Command.PackagePath = ""
    echo_task.Parameters.StdoutRedirectPath = "oss://xxx/xxx/"  # Better replace this path
    echo_task.Parameters.StderrRedirectPath = "oss://xxx/xxx/"  # Better replace this path 
    echo_task.InstanceCount = 3
    echo_task.ClusterId = cluster_id
    # Create task dag.
    task_dag = DAG()
    task_dag.add_task(task_name="Echo", task=echo_task)
    # Create job description.
    job_desc.DAG = task_dag
    job_desc.Priority = 99 # 0-1000
    job_desc.Name = "PythonSDKDemo" 
    job_desc.JobFailOnInstanceFail = True
    job_id = client.create_job(job_desc).Id
    # Wait for job finished.
    errs = client.poll(job_id)
    if errs: print ("Some errors occur: %s" % '\n'.join(errs))
    # Delete cluster
    client.delete_cluster(cluster_id)
except ClientError, e:
    print (e.get_status_code(), e.get_code(), e.get_requestid(), e.get_msg())

2、 类和常量

2.1 接口类型

接口类型提供了BatchCompute服务所有API接口的Python实现以及其他一些有用的辅助接口。

2.2 描述类型

描述类型主要作为创建资源时的参数类型或获取资源状态信息时由服务的返回类型。

关于可序列化

描述类型均为可序列化的类型，SDK中所有可序列化的类均从内部类型 Jsonizable 继承而来。以下是关于 Jsonizable 类型及其子类的描述；

参数说明：

Jsonizable 及其子类对象均可通过字典，Jsonizable 对象或者描述字典的JSON串初始化。注意，在初始化 Jsonizable 对象及其子类时，会丢弃字典或者JSON串中所有不合法的属性描述信息。

• 通过字典初始化 Jsonizable 类:

e.g.

from batchcompute.resources import JobDescription
# A dict object.
properties = {
    "Name": "PythonSDKDemo",
    "Description": "Batchcompute"
}
jsonizable = JobDescription(properties)
print (jsonizable.Name)
print (jsonizable.Description)

• 通过JSON字符串初始化 Jsonizable 类

e.g.

from batchcompute.resources import JobDescription
# A string jsonized from a dict object.
properties = '''{
    "Name": "PythonSDKDemo",
    "Description": "Batchcompute"
}'''
jsonizable = JobDescription(properties)
print (jsonizable.Name)
print (jsonizable.Description)

• 通过相同类的对象初始化 Jsonizable 类。

e.g.

form batchcompute.resources import JobDescription
# A JobDescription object.
jsonizable1 = JobDescription()
jsonizable1.Name = "PythonSDKDemo"
jsonizable1.Description = "Batchcompute"
jsonizable2 = JobDescription(jsonizable1)
print(jsonizable2.Name)
print(jsonizable2.Description)

方法说明：

关于类属性

可序列化的类型均具有各种属性。属性均可以通过其名称直接读取，例如，你可以通过如下代码获取作业的ID:另外属性名与Python规范PEP8中类的命名方式保持一致(区别于类方法的命名规则)，遵循CamelCase的拼写规则.

# job is a Job object.
job = ...
job_id = job.Id
print (job_id)

• 另外，可以通过字典取值的方式获取属性，例如：

# job is a Job object.
job = ...
job_id = job["Id"]
print (job_id)

• 对于类 JobDescription, DAG, TaskDescription, 可以通过赋值的方式更改某个属性的值,例如:

from batchcompute.resources import JobDescription
job_desc = JobDescription()
job_desc.Name = "PythonSDKDemo"

• 对于类 JobDescription, DAG, TaskDescription, 可以通过字典的方式对类的属性进行赋值, 例如:

from batchcompute.resources import JobDescription
job_desc = JobDescription()
job_desc["Id"] = "PythonSDKDemo"

2.3 响应类型

关于响应类

所有的响应类型 (CreateResponse, GetResponse, ActionRespnse, ListResponse)均继承自内部类型RawResponse.以下描述适用于所有RawResponse的子类。

属性说明：

e.g.

...
response = client.create_job(job_desc)
print (response.RequestId)
print (response.StatusCode)

2.4 异常类型

非法的参数或者非法请求时会抛出异常

2.5 常量