首先，我们需要使用Beam Kafka IO库来创建流式处理流。以下是一个如何将Beam Kafka IO库添加到您的项目中的示例：

from apache_beam.options.pipeline_options import PipelineOptions
from apache_beam.options.pipeline_options import StandardOptions
import apache_beam as beam
from apache_beam.io.kafka import ReadFromKafka 

def run():
    proj_id = "your-project-here"
    top_id = "your-topic-here"
    pipeline_options = PipelineOptions(streaming=True, project=proj_id, runner="DirectRunner")
    pipeline_options.view_as(StandardOptions).streaming = True

    with beam.Pipeline(options=pipeline_options) as pipeline:
        kafka_messages = ( pipeline 
                          | 'Read From Kafka' >> ReadFromKafka(consumer_config={"bootstrap.servers": "your-bootstrap-servers-here", "group.id": '0'}, topics=[top_id])
        )
    return pipeline.run()

现在，我们需要将Kafka消息转换为可读取的格式并将其输出到控制台或文本文件。以下代码演示了如何执行此操作：

def run():
    proj_id = "your-project-here"
    top_id = "your-topic-here"
    output_file = "output.txt"
    pipeline_options = PipelineOptions(streaming=True, project=proj_id, runner="DirectRunner")
    pipeline_options.view_as(StandardOptions).streaming = True

    with beam.Pipeline(options=pipeline_options) as pipeline:
        kafka_messages = ( pipeline 
                          | 'Read From Kafka' >> ReadFromKafka(consumer_config={"bootstrap.servers": "your-bootstrap-servers-here", "group.id": '0'}, topics=[top_id])
                          | 'Decode' >> beam.Map(lambda message: message.value().decode('utf-8'))
                          | 'Write to Console' >> beam.Map(print)
                          | 'Write to File' >> beam.io.WriteToText(output_file,
                                                                    file_name_suffix='.txt')
        )
    return pipeline.run()

在这个例子中，我们将每个Kafka消息转换为UTF-8格式，然后将其输出到控制台和名为output.txt的文件中。您可以将此示例中的输出方法更改为符合您的要求。

注意：在生产环境中使用此代码时，请将runner从DirectRunner更改为DataflowRunner或其他适合您的用例的选项。