本文属于机器翻译版本。若本译文内容与英语原文存在差异,则一律以英文原文为准。
SQS使用 Amazon 向 Amazon 发送和接收批量消息 Amazon SDK
以下代码示例展示了如何:
创建 Amazon SQS 队列。
将成批消息发送到队列。
从队列中接收成批消息。
从队列中删除成批消息。
- Python
-
- SDK适用于 Python (Boto3)
-
注意
还有更多相关信息 GitHub。查找完整示例,学习如何在 Amazon 代码示例存储库
中进行设置和运行。 创建用于封装 Amazon SQS 消息函数的函数。
import logging import sys import boto3 from botocore.exceptions import ClientError import queue_wrapper logger = logging.getLogger(__name__) sqs = boto3.resource("sqs") def send_messages(queue, messages): """ Send a batch of messages in a single request to an SQS queue. This request may return overall success even when some messages were not sent. The caller must inspect the Successful and Failed lists in the response and resend any failed messages. :param queue: The queue to receive the messages. :param messages: The messages to send to the queue. These are simplified to contain only the message body and attributes. :return: The response from SQS that contains the list of successful and failed messages. """ try: entries = [ { "Id": str(ind), "MessageBody": msg["body"], "MessageAttributes": msg["attributes"], } for ind, msg in enumerate(messages) ] response = queue.send_messages(Entries=entries) if "Successful" in response: for msg_meta in response["Successful"]: logger.info( "Message sent: %s: %s", msg_meta["MessageId"], messages[int(msg_meta["Id"])]["body"], ) if "Failed" in response: for msg_meta in response["Failed"]: logger.warning( "Failed to send: %s: %s", msg_meta["MessageId"], messages[int(msg_meta["Id"])]["body"], ) except ClientError as error: logger.exception("Send messages failed to queue: %s", queue) raise error else: return response def receive_messages(queue, max_number, wait_time): """ Receive a batch of messages in a single request from an SQS queue. :param queue: The queue from which to receive messages. :param max_number: The maximum number of messages to receive. The actual number of messages received might be less. :param wait_time: The maximum time to wait (in seconds) before returning. When this number is greater than zero, long polling is used. This can result in reduced costs and fewer false empty responses. :return: The list of Message objects received. These each contain the body of the message and metadata and custom attributes. """ try: messages = queue.receive_messages( MessageAttributeNames=["All"], MaxNumberOfMessages=max_number, WaitTimeSeconds=wait_time, ) for msg in messages: logger.info("Received message: %s: %s", msg.message_id, msg.body) except ClientError as error: logger.exception("Couldn't receive messages from queue: %s", queue) raise error else: return messages def delete_messages(queue, messages): """ Delete a batch of messages from a queue in a single request. :param queue: The queue from which to delete the messages. :param messages: The list of messages to delete. :return: The response from SQS that contains the list of successful and failed message deletions. """ try: entries = [ {"Id": str(ind), "ReceiptHandle": msg.receipt_handle} for ind, msg in enumerate(messages) ] response = queue.delete_messages(Entries=entries) if "Successful" in response: for msg_meta in response["Successful"]: logger.info("Deleted %s", messages[int(msg_meta["Id"])].receipt_handle) if "Failed" in response: for msg_meta in response["Failed"]: logger.warning( "Could not delete %s", messages[int(msg_meta["Id"])].receipt_handle ) except ClientError: logger.exception("Couldn't delete messages from queue %s", queue) else: return response
使用包装器函数批量发送和接收消息。
def usage_demo(): """ Shows how to: * Read the lines from this Python file and send the lines in batches of 10 as messages to a queue. * Receive the messages in batches until the queue is empty. * Reassemble the lines of the file and verify they match the original file. """ def pack_message(msg_path, msg_body, msg_line): return { "body": msg_body, "attributes": { "path": {"StringValue": msg_path, "DataType": "String"}, "line": {"StringValue": str(msg_line), "DataType": "String"}, }, } def unpack_message(msg): return ( msg.message_attributes["path"]["StringValue"], msg.body, int(msg.message_attributes["line"]["StringValue"]), ) print("-" * 88) print("Welcome to the Amazon Simple Queue Service (Amazon SQS) demo!") print("-" * 88) queue = queue_wrapper.create_queue("sqs-usage-demo-message-wrapper") with open(__file__) as file: lines = file.readlines() line = 0 batch_size = 10 received_lines = [None] * len(lines) print(f"Sending file lines in batches of {batch_size} as messages.") while line < len(lines): messages = [ pack_message(__file__, lines[index], index) for index in range(line, min(line + batch_size, len(lines))) ] line = line + batch_size send_messages(queue, messages) print(".", end="") sys.stdout.flush() print(f"Done. Sent {len(lines) - 1} messages.") print(f"Receiving, handling, and deleting messages in batches of {batch_size}.") more_messages = True while more_messages: received_messages = receive_messages(queue, batch_size, 2) print(".", end="") sys.stdout.flush() for message in received_messages: path, body, line = unpack_message(message) received_lines[line] = body if received_messages: delete_messages(queue, received_messages) else: more_messages = False print("Done.") if all([lines[index] == received_lines[index] for index in range(len(lines))]): print(f"Successfully reassembled all file lines!") else: print(f"Uh oh, some lines were missed!") queue.delete() print("Thanks for watching!") print("-" * 88)
-
有关API详细信息,请参阅 Python (Boto3) API 参考中的Amazon SDK以下主题。
-
有关 Amazon SDK开发者指南和代码示例的完整列表,请参阅将 Amazon SQS 与 Amazon SDK 结合使用。本主题还包括有关入门的信息以及有关先前SDK版本的详细信息。
将消息发布到队列
将 Amazon 消息处理框架用于. NET与亚马逊合作 SQS