typing_extensions package isn't installed with apache-airflow-providers-amazon causing an issue for SqlToS3Operator

[frankcash]

Apache Airflow Provider(s)

amazon

Versions of Apache Airflow Providers

3.0.0rc2

Apache Airflow version

2.2.3 (latest released)

Python version

Python 3.9.7 (default, Oct 12 2021, 02:43:43)

Operating System

Debian GNU/Linux 11 (bullseye)

Deployment

Docker-Compose

Deployment details

No response

What happened

I was working on adding this operator to a DAG and it failed to import due to a lack of a required file

What you expected to happen

No response

How to reproduce

Add

from airflow.providers.amazon.aws.transfers.sql_to_s3 import SqlToS3Operator

to a dag

Anything else

This can be resolved by adding typing-extensions==4.1.1 to requirements.txt when building the project (locally)

Are you willing to submit PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct

[boring-cyborg]

Thanks for opening your first issue here! Be sure to follow the issue template!

[frankcash]

It looks like as of Python 3.8 this functionality can be achieved by changing the import to

from typing import Literal

instead of

from typing_extensions import Literal

but utilizing typing_extensions wouldn't break backwards compatibility it would just require a change here

airflow/setup.cfg

Line 140 in 0a6ea57

typing-extensions>=3.7.4;python_version<"3.8"

[frankcash]

Also looks like the SQS Sensor suffers from a similar issue https://github.com/apache/airflow/blob/main/airflow/providers/amazon/aws/sensors/sqs.py#L24

[josh-fell]

FYI - There is typing_compat that could be used for the Literal type import.

[josh-fell]

Might be worth putting in some sort of CI check for these compat-related imports too. @potiuk WDYT?

[potiuk]

Ah yeah. Absolutely :)