fix: Preserve __alias__ when mapping streams with repeated schema messages

[DanilJr]

It relates to this issue

Closes #1521

---

📚 Documentation preview 📚: https://meltano-sdk--1524.org.readthedocs.build/en/1524/

[codecov]

Codecov Report

Merging #1524 (5988c83) into main (f943fa3) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #1524   +/-   ##
=======================================
  Coverage   85.50%   85.51%           
=======================================
  Files          57       57           
  Lines        4692     4693    +1     
  Branches      802      802           
=======================================
+ Hits         4012     4013    +1     
  Misses        489      489           
  Partials      191      191           

Impacted Files	Coverage Δ
singer_sdk/mapper.py	85.84% <100.00%> (+0.06%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[edgarrmondragon]

Hey @DanilJr!

Thanks for the PR. Just one question. Otherwise this looks good 🙂

[edgarrmondragon]

I would like to have automated tests but at least manual testing confirms that the patch works as intended:

1. Singer input

{"type": "SCHEMA", "stream": "example", "schema": {"properties": {"id": {"type": "integer"}}}}
{"type": "RECORD", "stream": "example", "record": {"id": 1}}
{"type": "SCHEMA", "stream": "example", "schema": {"properties": {"id": {"type": "integer"}}}}
{"type": "RECORD", "stream": "example", "record": {"id": 2}}

2. Stream config

{
  "stream_maps": {
    "example": {
      "__alias__": "example_v2"
    }
  }
}

3. Mapper output

{"type": "SCHEMA", "stream": "example_v2", "schema": {"properties": {"id": {"type": "integer"}}}, "key_properties": [], "bookmark_properties": []}
{"type": "RECORD", "stream": "example_v2", "record": {"id": 1}, "time_extracted": "2023-03-24T19:41:37.953424+00:00"}
{"type": "SCHEMA", "stream": "example_v2", "schema": {"properties": {"id": {"type": "integer"}}}, "key_properties": [], "bookmark_properties": []}
{"type": "RECORD", "stream": "example_v2", "record": {"id": 2}, "time_extracted": "2023-03-24T19:41:37.953966+00:00"}

I can also confirm that using a explicit shallow copy (dict.copy()) also works.