Question

我正在制作AWS Lambda功能。现在我想isinstance(context,LambdaContext)，我希望它适用于AWS Lambda。但是我在当地进行单元测试。

那么如何导入LambdaContext。

最佳

Answer 1

您可以尝试使用LocalStack：

LocalStack提供了一个易于使用的测试/模拟框架，用于开发云应用程序。

目前，重点主要是支持AWS云堆栈。

LocalStack在本地计算机上旋转了以下核心Cloud API：


http://localhost:4567
的API网关
Kinesis http://localhost:4568

DynamoDB http://localhost:4569

DynamoDB Streams http://localhost:4570

http://localhost:4571
的弹性搜索
S3 http://localhost:4572

Firehose http://localhost:4573

Lambda在http://localhost:4574

http://localhost:4575
的SNS
http://localhost:4576
的SQS
http://localhost:4577
的红移
ES（Elasticsearch Service）http://localhost:4578

http://localhost:4579
的SES
Route http://localhost:4580

http://localhost:4581
上的CloudFormation
http://localhost:4582
上的CloudWatch

Answer 2

刚发现this package，似乎有人用这个问题中其他答案的概念制作了一个点子包。

遇到同样的问题，希望对以后遇到此问题的人有所帮助。

Answer 3

在/var/runtime/awslambda/bootstrap.py中定义的

LambdaContext类，用于启动用户功能并具有以下结构：

class LambdaContext(object):
    def __init__(self, invokeid, context_objs, client_context, invoked_function_arn=None):
        self.aws_request_id = invokeid
        self.log_group_name = os.environ['AWS_LAMBDA_LOG_GROUP_NAME']
        self.log_stream_name = os.environ['AWS_LAMBDA_LOG_STREAM_NAME']
        self.function_name = os.environ["AWS_LAMBDA_FUNCTION_NAME"]
        self.memory_limit_in_mb = os.environ['AWS_LAMBDA_FUNCTION_MEMORY_SIZE']
        self.function_version = os.environ['AWS_LAMBDA_FUNCTION_VERSION']
        self.invoked_function_arn = invoked_function_arn

        self.client_context = make_obj_from_dict(ClientContext, client_context)
        if self.client_context is not None:
            self.client_context.client = make_obj_from_dict(Client, self.client_context.client)
        self.identity = make_obj_from_dict(CognitoIdentity, context_objs)

    def get_remaining_time_in_millis(self):
        return lambda_runtime.get_remaining_time()

    def log(self, msg):
        str_msg = str(msg)
        lambda_runtime.send_console_message(str_msg, byte_len(str_msg))

如果要在本地环境中模拟它，只需将其添加到脚本中即可：

class ClientContext(object):
    __slots__ = ['custom', 'env', 'client']


def make_obj_from_dict(_class, _dict, fields=None):
    if _dict is None:
        return None
    obj = _class()
    set_obj_from_dict(obj, _dict)
    return obj


def set_obj_from_dict(obj, _dict, fields=None):
    if fields is None:
        fields = obj.__class__.__slots__
    for field in fields:
        setattr(obj, field, _dict.get(field, None))


class LambdaContext(object):
    def __init__(self, invokeid, context_objs, client_context, invoked_function_arn=None):
        self.aws_request_id = invokeid
        self.log_group_name = os.environ['AWS_LAMBDA_LOG_GROUP_NAME']
        self.log_stream_name = os.environ['AWS_LAMBDA_LOG_STREAM_NAME']
        self.function_name = os.environ["AWS_LAMBDA_FUNCTION_NAME"]
        self.memory_limit_in_mb = os.environ['AWS_LAMBDA_FUNCTION_MEMORY_SIZE']
        self.function_version = os.environ['AWS_LAMBDA_FUNCTION_VERSION']
        self.invoked_function_arn = invoked_function_arn

        self.client_context = make_obj_from_dict(ClientContext, client_context)
        if self.client_context is not None:
            self.client_context.client = None
        self.identity = None


    def get_remaining_time_in_millis(self):
        return None

    def log(self, msg):
        str_msg = str(msg)
        print(str_msg)
        # lambda_runtime.send_console_message(str_msg, byte_len(str_msg))

Answer 4

这不是最严格的意义上的答案，如果不合适，请删除。

我已经尝试了@ Dmitry-Masanov的答案，并为pytest绑定了一个python固定装置，可以在测试脚本本身中使用，也可以在conftest.py文件中使用它。

@pytest.fixture
def mock_lambda_context():
    class ClientContext:
        """
        Class for mocking Context

        Has `custom`, `env`, and `client` `__slots__`
        """

        __slots__ = ["custom", "env", "client"]

    def make_obj_from_dict(_class, _dict, fields=None):
        """
        Makes an object of `_class` from a `dict`

        :param _class: A class representing the context
        :type _class: `ContextClass`
        :param _dict: A dictionary of data 
        :type _dict: `dict`
        :param fields: [description], defaults to None
        :type fields: [type], optional
        :return: An object
        :rtype: `ClientContext` class
        """
        if _dict is None:
            return None
        obj = _class()
        set_obj_from_dict(obj, _dict)
        return obj

    def set_obj_from_dict(obj, _dict, fields=None):
        if fields is None:
            fields = obj.__class__.__slots__
        for field in fields:
            setattr(obj, field, _dict.get(field, None))

    class LambdaContext(object):
        """
        Create a Lambda Context Class object
        """

        def __init__(self, invokeid, client_context, invoked_function_arn=None):
            self.aws_request_id = invokeid
            self.log_group_name = "AWS_LAMBDA_LOG_GROUP_NAME"
            self.log_stream_name = "AWS_LAMBDA_LOG_STREAM_NAME"
            self.function_name = "AWS_LAMBDA_FUNCTION_NAME"
            self.memory_limit_in_mb = "AWS_LAMBDA_FUNCTION_MEMORY_SIZE"
            self.function_version = "AWS_LAMBDA_FUNCTION_VERSION"
            self.invoked_function_arn = invoked_function_arn

            self.client_context = make_obj_from_dict(ClientContext, client_context)
            if self.client_context is not None:
                self.client_context.client = None
            self.identity = None

        def get_remaining_time_in_millis(self):
            return None

        def log(self, msg):
            str_msg = str(msg)
            print(str_msg)

    lambda_context = LambdaContext("AWS_ID", {})

    return lambda_context

我几乎可以肯定我的实现可以改进（OO python不是我的强项），但这可以达到我的目的：

模拟lambda上下文对象
与pytest集成

这可以类似于以下方式使用：

@pytest.mark.smoke
def test_lambda_handler(apigw_event, mock_lambda_context):
    ret = app.lambda_handler(apigw_event, mock_lambda_context)
    assert "log_stream_name" in ret["body"]

再次感谢德米特里，您的回答为我节省了很多麻烦。

如何导入LambdaContext？

4 个答案: