GH-37390: [Python] Expose StreamDecoder API to pyarrow

[kdkavanagh]

Rationale for this change

The stream decoder API allows for more fine-grained control when reading arrow streams that are being concurrently written to. Reads of the stream can be handled manually, with bytes offered incrementally to the stream decoder API.

Are these changes tested?

Yes

• Closes: [Python] Expose StreamDecoder to pyarrow python API #37390

[kou]

This is not a MINOR change: https://github.com/apache/arrow/blob/main/CONTRIBUTING.md#Minor-Fixes

Could you open a new issue for this?

[kdkavanagh]

Done: #37390

[github-actions]

⚠️ GitHub issue #37390 has been automatically assigned in GitHub to PR creator.

[AlenkaF]

Thank you for the contribution @kdkavanagh!
There are some corrections needed to have a green CI.

• One are linter errors in Python files. Yo can use Archery tool for that: https://arrow.apache.org/docs/dev/developers/python.html#coding-style

• A docstring error due to parameter not being documented:

INFO:archery:Running Python docstring linters

pyarrow.ipc.StreamDecoder
PR01: Parameters {'memory_pool'} not documented

• The failure in one of the builds is related:
  https://github.com/apache/arrow/actions/runs/5968833843/job/16226994102?pr=37371#step:6:3931