I am writing a apache beam streaming pipeline, where I want to read msg from pubsusb, and want to iterate over all the msg present in window , after my window time, in this case 1 min.

But, although I have given ,window size of 60, it starts printing msg from PubSub, as soon as I publish on the topic.

Here is my code.

import json
import os
import typing
import apache_beam as beam
from apache_beam.options.pipeline_options import PipelineOptions
from apache_beam.options.pipeline_options import GoogleCloudOptions
from apache_beam.io.gcp.internal.clients import bigquery
from apache_beam.transforms.sql import SqlTransform


pipeline_options = PipelineOptions( streaming=True)




def run():
    with beam.Pipeline(options=pipeline_options) as p:


        out= (
            p
            | "Read from Pub/Sub subscription" >> beam.io.ReadFromPubSub(subscription="projects/<PROJ_ID>/subscriptions/Test-sub")
            | "Decode and parse Json" >> beam.Map(lambda element: element.decode("utf-8"))
            
            |"window" >> beam.WindowInto(beam.window.FixedWindows(10))  


            << After the 60sec window gets finished, I want to take each element from window and pass it to a Google Gemini API to classify its sentiments>>

            | beam.Map(print)
        )
    

    # p.run()
if __name__ == '__main__':
  run()

Can anyone tell me whats happening or how to achieve this?

I want to accumulate all the msg, till my window size.
After window gets complete, I want to iterate over all elements, and do some operations.

Thank you.

Above is my code and what I have tried