无法使用 Apache Beam（Python SDK）读取 Pub/Sub 消息

Question

我正在尝试使用 Beam 编程框架 (Python SDK) 从 Pub/Sub 主题流式传输消息并将它们写入控制台。

这是我的代码（apache-beam==2.27.0）：

import apache_beam as beam
from apache_beam.options.pipeline_options import PipelineOptions

TOPIC_PATH = "projects/<project-id>/topics/<topic-id>"

def run(pubsub_topic):
    options = PipelineOptions(
        streaming=True
    )
    runner = 'DirectRunner'

    print("I reached before pipeline")

    with beam.Pipeline(runner, options=options) as pipeline:
        (
            pipeline
            | "Read from Pub/Sub topic" >> beam.io.ReadFromPubSub(topic=pubsub_topic)
            | "Writing to console" >> beam.Map(print)
        )

    print("I reached after pipeline")

    result = pipeline.run()
    result.wait_until_finish()


run(TOPIC_PATH)

然而，当我执行这个管道时，我得到了这个 TypeError：

ERROR:apache_beam.runners.direct.executor:Exception at bundle <apache_beam.runners.direct.bundle_factory._Bundle object at 0x1349763c0>, due to an exception.

TypeError: create_subscription() takes from 1 to 2 positional arguments but 3 were given

最后说：

ERROR:apache_beam.runners.direct.executor:Giving up after 4 attempts.

我不确定，我做错了什么，提前感谢您的帮助。

Answer 1

我不知道错误的确切位置，但您可以考虑使用以下 Beam 示例之一作为模型并从那里开始吗？

https://github.com/apache/beam/blob/master/sdks/python/apache_beam/io/gcp/pubsub_it_pipeline.py

https://github.com/apache/beam/blob/release-2.27.0/sdks/python/apache_beam/examples/snippets/snippets.py#L684

Answer 2

当我通过“pip install apache-beam”安装时遇到了同样的问题。当我切换到“pip install apache-beam [gcp]”时，它对我有用，即使使用 DirectRunner。