✨ add support for psycopg3 async

[idumancic]

We have an architecture problem here aiopg implements cursor as coroutine but psycopg does not. I don't know how you want to handle this decision but that's why the fmt check fails because of the base class difference.

Let me know what you decide so I can implement, I don't want to do any major changes before you see it because it affects both CommandsAsync and AsynConnectionType public API.

Tests pass only because I am overriding the cursor method in Psycopg3CommandsAsync, but the typecheck fails as it should.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (12f88f4) to head (3cfe43c).

❗ Current head 3cfe43c differs from pull request most recent head c2a2707. Consider uploading reports for the commit c2a2707 to get more accurate results

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #218   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           21        21           
  Lines          582       595   +13     
  Branches       100       102    +2     
=========================================
+ Hits           582       595   +13     

Flag	Coverage Δ
3.10-bigquery	71.26% <69.23%> (-0.05%)	⬇️
3.10-core	86.72% <69.23%> (-0.40%)	⬇️
3.10-mssql	72.10% <69.23%> (-0.07%)	⬇️
3.10-mysql	71.26% <69.23%> (-0.05%)	⬇️
3.10-oracle	72.10% <69.23%> (-0.07%)	⬇️
3.10-postgresql	87.22% <100.00%> (+0.97%)	⬆️
3.10-sqlite	70.92% <69.23%> (-0.04%)	⬇️
3.11-bigquery	71.26% <69.23%> (-0.05%)	⬇️
3.11-core	86.72% <69.23%> (-0.40%)	⬇️
3.11-mssql	72.10% <69.23%> (-0.07%)	⬇️
3.11-mysql	71.26% <69.23%> (-0.05%)	⬇️
3.11-oracle	72.10% <69.23%> (-0.07%)	⬇️
3.11-postgresql	87.22% <100.00%> (+0.97%)	⬆️
3.11-sqlite	70.92% <69.23%> (-0.04%)	⬇️
3.12-bigquery	71.26% <69.23%> (-0.05%)	⬇️
3.12-core	86.89% <69.23%> (-0.40%)	⬇️
3.12-mssql	72.10% <69.23%> (-0.07%)	⬇️
3.12-mysql	71.26% <69.23%> (-0.05%)	⬇️
3.12-oracle	72.10% <69.23%> (-0.07%)	⬇️
3.12-postgresql	87.39% <100.00%> (+0.96%)	⬆️
3.12-sqlite	70.92% <69.23%> (-0.04%)	⬇️
3.8-bigquery	71.33% <69.23%> (-0.05%)	⬇️
3.8-core	87.18% <69.23%> (-0.41%)	⬇️
3.8-mssql	72.17% <69.23%> (-0.07%)	⬇️
3.8-mysql	71.33% <69.23%> (-0.05%)	⬇️
3.8-oracle	72.17% <69.23%> (-0.07%)	⬇️
3.8-postgresql	87.35% <100.00%> (+0.97%)	⬆️
3.8-sqlite	70.99% <69.23%> (-0.04%)	⬇️
3.9-bigquery	71.33% <69.23%> (-0.05%)	⬇️
3.9-core	87.18% <69.23%> (-0.41%)	⬇️
3.9-mssql	72.17% <69.23%> (-0.07%)	⬇️
3.9-mysql	71.33% <69.23%> (-0.05%)	⬇️
3.9-oracle	72.17% <69.23%> (-0.07%)	⬇️
3.9-postgresql	87.35% <100.00%> (+0.97%)	⬆️
3.9-sqlite	70.99% <69.23%> (-0.04%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[zschumacher]

Seems like the cleanest thing might be wrapping the cursor in an awaitable for backwards compat, but I'm open to suggestions

[idumancic]

Seems like the cleanest thing might be wrapping the cursor in an awaitable for backwards compat, but I'm open to suggestions

That was my thought also, to keep the API as it is and just wrap the cursor in awaitable since aiopg implements it by default. Would that have any consequences to psycopg if we go with that solution, can you think of anything?

[zschumacher]

hey sorry - been super busy. I think if we wrap it in an awaitable it will be completely fine. Is this still a feature you want to work on?