fix connection uri parsing when the host includes a scheme

[hussein-awala]

closes: #21638
closes: #31460

This PR introduces a solution to address the parsing problem that occurs when the connection host includes a scheme.

---

^ Add meaningful description above

Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in newsfragments.

[uranusjr]

Instead of trying to guess what the user wants, I’m more inclined to make this fail directly when the user tries to add the connection object in the first place.

[hussein-awala]

@uranusjr IMO there should not be any problem if the user provide this kind of hosts, since it's a valid host but with a protocol, for ex, if we create the same connection from the UI:

Then we use the Airflow CLI to get the connection information:

root@55c0591cd327:/opt/airflow# airflow connections get test_conn
id | conn_id   | conn_type | description | host            | schema | login | password | port | is_encrypted | is_extra_encrypted | extra_dejson       | get_uri                                                   
===+===========+===========+=============+=================+========+=======+==========+======+==============+====================+====================+===========================================================
1  | test_conn | generic   |             | protocol://host |        | user  | pass     | 123  | False        | False              | {'param': 'value'} | generic://user:pass@protocol%3A%2F%2Fhost:123/?param=value

We see that it's successfully parsed and we can use it without any problem, and if I think that the method parse_uri in the connection class should parse the generated uri and create the same connection from it. WDYT?

[hussein-awala]

I think we can add a new column to specify the protocol:

Which is cleaner and easily supported by the uri parsing:

root@cb473c029341:/opt/airflow# airflow connections get spark_on_k8s_conn
id | conn_id           | conn_type | description | host       | schema | login | password | port | is_encrypted | is_extra_encrypted | extra_dejson               | get_uri                                        
===+===================+===========+=============+============+========+=======+==========+======+==============+====================+============================+================================================
1  | spark_on_k8s_conn | spark     |             | 100.68.0.1 |        |       | None     | 443  | False        | False              | {'deploy-mode': 'cluster'} | spark://k8s://100.68.0.1:443/?deploy-mode=clust
   |                   |           |             |            |        |       |          |      |              |                    |                            | er                                             
                                                                                                                                                                                                                   
root@cb473c029341:/opt/airflow# airflow connections get test_conn
id | conn_id   | conn_type | description      | host | schema | login | password | port | is_encrypted | is_extra_encrypted | extra_dejson       | get_uri                                                   
===+===========+===========+==================+======+========+=======+==========+======+==============+====================+====================+===========================================================
2  | test_conn | generic   | test description | host | schema | user  | pass     | 123  | False        | False              | {'param': 'value'} | generic://protocol://user:pass@host:123/schema?param=value 

[hussein-awala]

I will add some changes it to make it backward compatible

[hussein-awala]

It's ready and without a new column, I just updated the methods _parse_from_uri and get_uri and added a new pytest file for the connection model which contain the tests for my changes.

[amoghrajesh]

@hussein-awala the changes mostly look good to me.
Have a few comments

[amoghrajesh]

This variable name is rather confusing to me, can we have something else here?

[hussein-awala]

can you check it now?

[ephraimbuddy]

Hi @hussein-awala , we are about to release 2.6.2, can you address the review comments here so this can be part of 2.6.2

[amoghrajesh]

Thanks for those changes, @hussein-awala
LGTM +1

[hussein-awala]

Hi @hussein-awala , we are about to release 2.6.2, can you address the review comments here so this can be part of 2.6.2

@ephraimbuddy the review is addressed, can your review the PR?

[ephraimbuddy]

Hi @uranusjr , do you want to take another look?