使用尚未实现方法的模拟补丁增强 moto
Augmenting moto with mock patch where method is not yet implemented
我正在编写一个 lambda 函数，它获取 CW 日志组列表并在每个日志组上运行“导出到 s3”任务。
我正在使用 pytest 编写自动化测试，我正在使用 moto.mock_logs（以及其他），但是 create_export_tasks() 尚未实现（NotImplementedError）。
为了继续对所有其他方法使用 moto.mock_logs，我尝试使用 mock.patch 仅修补单个 create_export_task() 方法，但它无法找到要修补的正确对象（ ImportError).
我成功地使用 mock.Mock() 为我提供了我需要的功能，但我想知道我是否可以用 mock.patch() 做同样的事情？
工作代码：lambda.py
# lambda.py
"""Export CloudWatch Logs to S3 every 24 hours."""
import logging
import os
from time import time
import boto3
from botocore.config import Config
from botocore.exceptions import ClientError
logger = logging.getLogger(__name__)
logging.basicConfig(level=logging.INFO)
class CloudWatchLogsS3Archive:
botocore_config = Config(retries={"max_attempts": 10, "mode": "adaptive"})
def __init__(self, s3_bucket, account_id) -> None:
self.s3_bucket = s3_bucket
self.account_id = account_id
self.extra_args = {}
self.log_groups = []
self.log_groups_to_export = []
self.logs = boto3.client("logs", config=self.botocore_config)
self.ssm = boto3.client("ssm", config=self.botocore_config)
def check_valid_inputs(self):
"""Check that required inputs are present and valid"""
if len(self.account_id) != 12:
logging.error("Account Id must be valid 12-digit AWS account id")
raise ValueError("Account Id must be valid 12-digit AWS account id")
def collect_log_groups(self):
"""Capture the names of all of the CloudWatch Log Groups"""
paginator = self.logs.get_paginator("describe_log_groups")
page_it = paginator.paginate()
for p in page_it:
for lg in p["logGroups"]:
yield lg["logGroupName"]
def get_last_export_time(self, Name) -> str:
"""Get time of the last export from SSM Parameter Store"""
try:
return self.ssm.get_parameter(Name=Name)["Parameter"]["Value"]
except (self.ssm.exceptions.ParameterNotFound, ClientError) as exc:
logger.warning(*exc.args)
if exc.response["Error"]["Code"] == "ParameterNotFound": # type: ignore
return "0"
else:
raise
def set_export_time(self):
"""Set current export time"""
return round(time() * 1000)
def put_export_time(self, put_time, Name):
"""Put current export time to SSM Parameter Store"""
self.ssm.put_parameter(Name=Name, Value=str(put_time), Overwrite=True)
def create_export_tasks(
self, log_group_name, fromTime, toTime, s3_bucket, account_id
):
"""Create new CloudWatchLogs Export Tasks"""
try:
response = self.logs.create_export_task(
logGroupName=log_group_name,
fromTime=int(fromTime),
to=toTime,
destination=s3_bucket,
destinationPrefix="{}/{}".format(account_id, log_group_name.strip("/")),
)
logger.info("✔ Task created: %s" % response["taskId"])
except self.logs.exceptions.LimitExceededException:
"""The Boto3 standard retry mode will catch throttling errors and
exceptions, and will back off and retry them for you."""
logger.warning(
"⚠ Too many concurrently running export tasks "
"(LimitExceededException); backing off and retrying..."
)
# return False
except Exception as e:
logger.exception(
"✖ Error exporting %s: %s",
log_group_name,
getattr(e, "message", repr(e)),
)
def lambda_handler(event, context):
s3_bucket = os.environ["S3_BUCKET"]
account_id = os.environ["ACCOUNT_ID"]
c = CloudWatchLogsS3Archive(s3_bucket, account_id)
c.check_valid_inputs()
log_groups = c.collect_log_groups()
for log_group_name in log_groups:
fromTime = c.get_last_export_time(log_group_name)
toTime = c.set_export_time()
c.create_export_tasks(log_group_name, fromTime, toTime, s3_bucket, account_id)
c.put_export_time(log_group_name, toTime)
测试代码（pytest）：test_lambda.py
# test_lambda.py
"""Test Lambda Function"""
import os
from unittest import mock
import boto3
import moto
import pytest
@pytest.fixture
def f_aws_credentials(autouse=True):
"""Mocked AWS Credentials for moto.
This is a "side effect" function and None is returned because we are
modifying the environment in which other downstream functions are excuting
"""
os.environ["AWS_ACCESS_KEY_ID"] = "testing"
os.environ["AWS_SECRET_ACCESS_KEY"] = "testing"
os.environ["AWS_SECURITY_TOKEN"] = "testing"
os.environ["AWS_SESSION_TOKEN"] = "testing"
os.environ["AWS_DEFAULT_REGION"] = "us-east-1"
@moto.mock_logs
@moto.mock_ssm
def test_create_export_tasks():
from cloudwatch_logs_s3_archive import CloudWatchLogsS3Archive
c = CloudWatchLogsS3Archive("bucket", "123412341234")
# ssm = boto3.client("ssm")
logs = boto3.client("logs")
logs.create_log_group(logGroupName="/log-exporter-last-export/first")
logs.create_log_group(logGroupName="/log-exporter-last-export/second")
logs.create_log_group(logGroupName="/log-exporter-last-export/third")
log_group_name = "/log-exporter-last-export/first"
s3_bucket = "s3_bucket"
account_id = 123412341234
toTime = c.set_export_time()
fromTime = c.get_last_export_time(Name="/log-exporter-last-export/first")
c.logs.create_export_task = mock.Mock(
return_value={"taskId": "I am mocked via mock.Mock"}
)
c.create_export_tasks(
"/log-exporter-last-export/first", fromTime, toTime, "s3_bucket", 123412341234
)
assert c.logs.create_export_task.called
c.logs.create_export_task.assert_called
c.logs.create_export_task.assert_called_with(
logGroupName=log_group_name,
fromTime=int(fromTime),
to=toTime,
destination=s3_bucket,
destinationPrefix="{}/{}".format(account_id, log_group_name.strip("/")),
)
I'm wondering if I can do the same with mock.patch()?
当然可以，使用 mock.patch.object():
with mock.patch.object(
c.logs,
'create_export_task',
return_value={"taskId": "I am mocked via mock.Mock"}
):
c.create_export_tasks(
"/log-exporter-last-export/first", fromTime, toTime, "s3_bucket", 123412341234
)
assert c.logs.create_export_task.called
如果您不喜欢使用上下文管理器，我建议您安装 pytest-mock 插件和 pytest 插件，它提供了一个方便的 mocker 固定装置。你的测试看起来像
def test_create_export_tasks(mocker):
...
mocker.patch.object(
c.logs,
'create_export_task',
return_value={"taskId": "I am mocked via mock.Mock"}
)
c.create_export_tasks(
"/log-exporter-last-export/first", fromTime, toTime, "s3_bucket", 123412341234
)
assert c.logs.create_export_task.called
mocker 基本上是 unittest.mock 模块的代理，提供相同的功能和方法，除了它会在测试结束时自动清除所有补丁，所以少了一件需要关心的事情关于。
我正在编写一个 lambda 函数，它获取 CW 日志组列表并在每个日志组上运行“导出到 s3”任务。
我正在使用 pytest 编写自动化测试，我正在使用 moto.mock_logs（以及其他），但是 create_export_tasks() 尚未实现（NotImplementedError）。
为了继续对所有其他方法使用 moto.mock_logs，我尝试使用 mock.patch 仅修补单个 create_export_task() 方法，但它无法找到要修补的正确对象（ ImportError).
我成功地使用 mock.Mock() 为我提供了我需要的功能，但我想知道我是否可以用 mock.patch() 做同样的事情？
工作代码：lambda.py
# lambda.py
"""Export CloudWatch Logs to S3 every 24 hours."""
import logging
import os
from time import time
import boto3
from botocore.config import Config
from botocore.exceptions import ClientError
logger = logging.getLogger(__name__)
logging.basicConfig(level=logging.INFO)
class CloudWatchLogsS3Archive:
botocore_config = Config(retries={"max_attempts": 10, "mode": "adaptive"})
def __init__(self, s3_bucket, account_id) -> None:
self.s3_bucket = s3_bucket
self.account_id = account_id
self.extra_args = {}
self.log_groups = []
self.log_groups_to_export = []
self.logs = boto3.client("logs", config=self.botocore_config)
self.ssm = boto3.client("ssm", config=self.botocore_config)
def check_valid_inputs(self):
"""Check that required inputs are present and valid"""
if len(self.account_id) != 12:
logging.error("Account Id must be valid 12-digit AWS account id")
raise ValueError("Account Id must be valid 12-digit AWS account id")
def collect_log_groups(self):
"""Capture the names of all of the CloudWatch Log Groups"""
paginator = self.logs.get_paginator("describe_log_groups")
page_it = paginator.paginate()
for p in page_it:
for lg in p["logGroups"]:
yield lg["logGroupName"]
def get_last_export_time(self, Name) -> str:
"""Get time of the last export from SSM Parameter Store"""
try:
return self.ssm.get_parameter(Name=Name)["Parameter"]["Value"]
except (self.ssm.exceptions.ParameterNotFound, ClientError) as exc:
logger.warning(*exc.args)
if exc.response["Error"]["Code"] == "ParameterNotFound": # type: ignore
return "0"
else:
raise
def set_export_time(self):
"""Set current export time"""
return round(time() * 1000)
def put_export_time(self, put_time, Name):
"""Put current export time to SSM Parameter Store"""
self.ssm.put_parameter(Name=Name, Value=str(put_time), Overwrite=True)
def create_export_tasks(
self, log_group_name, fromTime, toTime, s3_bucket, account_id
):
"""Create new CloudWatchLogs Export Tasks"""
try:
response = self.logs.create_export_task(
logGroupName=log_group_name,
fromTime=int(fromTime),
to=toTime,
destination=s3_bucket,
destinationPrefix="{}/{}".format(account_id, log_group_name.strip("/")),
)
logger.info("✔ Task created: %s" % response["taskId"])
except self.logs.exceptions.LimitExceededException:
"""The Boto3 standard retry mode will catch throttling errors and
exceptions, and will back off and retry them for you."""
logger.warning(
"⚠ Too many concurrently running export tasks "
"(LimitExceededException); backing off and retrying..."
)
# return False
except Exception as e:
logger.exception(
"✖ Error exporting %s: %s",
log_group_name,
getattr(e, "message", repr(e)),
)
def lambda_handler(event, context):
s3_bucket = os.environ["S3_BUCKET"]
account_id = os.environ["ACCOUNT_ID"]
c = CloudWatchLogsS3Archive(s3_bucket, account_id)
c.check_valid_inputs()
log_groups = c.collect_log_groups()
for log_group_name in log_groups:
fromTime = c.get_last_export_time(log_group_name)
toTime = c.set_export_time()
c.create_export_tasks(log_group_name, fromTime, toTime, s3_bucket, account_id)
c.put_export_time(log_group_name, toTime)
测试代码（pytest）：test_lambda.py
# test_lambda.py
"""Test Lambda Function"""
import os
from unittest import mock
import boto3
import moto
import pytest
@pytest.fixture
def f_aws_credentials(autouse=True):
"""Mocked AWS Credentials for moto.
This is a "side effect" function and None is returned because we are
modifying the environment in which other downstream functions are excuting
"""
os.environ["AWS_ACCESS_KEY_ID"] = "testing"
os.environ["AWS_SECRET_ACCESS_KEY"] = "testing"
os.environ["AWS_SECURITY_TOKEN"] = "testing"
os.environ["AWS_SESSION_TOKEN"] = "testing"
os.environ["AWS_DEFAULT_REGION"] = "us-east-1"
@moto.mock_logs
@moto.mock_ssm
def test_create_export_tasks():
from cloudwatch_logs_s3_archive import CloudWatchLogsS3Archive
c = CloudWatchLogsS3Archive("bucket", "123412341234")
# ssm = boto3.client("ssm")
logs = boto3.client("logs")
logs.create_log_group(logGroupName="/log-exporter-last-export/first")
logs.create_log_group(logGroupName="/log-exporter-last-export/second")
logs.create_log_group(logGroupName="/log-exporter-last-export/third")
log_group_name = "/log-exporter-last-export/first"
s3_bucket = "s3_bucket"
account_id = 123412341234
toTime = c.set_export_time()
fromTime = c.get_last_export_time(Name="/log-exporter-last-export/first")
c.logs.create_export_task = mock.Mock(
return_value={"taskId": "I am mocked via mock.Mock"}
)
c.create_export_tasks(
"/log-exporter-last-export/first", fromTime, toTime, "s3_bucket", 123412341234
)
assert c.logs.create_export_task.called
c.logs.create_export_task.assert_called
c.logs.create_export_task.assert_called_with(
logGroupName=log_group_name,
fromTime=int(fromTime),
to=toTime,
destination=s3_bucket,
destinationPrefix="{}/{}".format(account_id, log_group_name.strip("/")),
)
I'm wondering if I can do the same with
mock.patch()?
当然可以，使用 mock.patch.object():
with mock.patch.object(
c.logs,
'create_export_task',
return_value={"taskId": "I am mocked via mock.Mock"}
):
c.create_export_tasks(
"/log-exporter-last-export/first", fromTime, toTime, "s3_bucket", 123412341234
)
assert c.logs.create_export_task.called
如果您不喜欢使用上下文管理器，我建议您安装 pytest-mock 插件和 pytest 插件，它提供了一个方便的 mocker 固定装置。你的测试看起来像
def test_create_export_tasks(mocker):
...
mocker.patch.object(
c.logs,
'create_export_task',
return_value={"taskId": "I am mocked via mock.Mock"}
)
c.create_export_tasks(
"/log-exporter-last-export/first", fromTime, toTime, "s3_bucket", 123412341234
)
assert c.logs.create_export_task.called
mocker 基本上是 unittest.mock 模块的代理，提供相同的功能和方法，除了它会在测试结束时自动清除所有补丁，所以少了一件需要关心的事情关于。