✨ Add AzureSQL source and Prefect tasks

[adrian-wojcik]

Summary

Adds SQLServer connector

Checklist

This PR:

• follows the guidelines laid out in CONTRIBUTING.md
• links relevant issue(s)
• adds/updates tests (if appropriate)
• adds/updates docstrings (if appropriate)
• adds an entry in CHANGELOG.md

[trymzet]

@adrian-wojcik I think you've mixed up Aselite stuff with AzureSQL stuff. Could you pls create separate PRs for each?

[adrian-wojcik]

@trymzet I did not mixed up. As you can see in viadot 1 Aselite task is using AzureSQL source only.

[trymzet]

@adrian-wojcik ok then can you rename this accordingly? Ie. the tasks should and flows should have the name of the source in them, eg. azure_sql_to_df(), etc.

[adrian-wojcik]

We were discussing it with Rafał and as Aselite is using AzureSQL which inherited from SQLServer, probably good idea is to remove in future this connector and move logic from AzureSQL Source to SQLServer Source, so all aselite deployments would use simply sql_server_to_adls flow. But as for now we believe the name convention can stay as it is, as it is temporary solution.

[trymzet]

@adrian-wojcik

1. I don't think this makes sense - Azure SQL is not the same as SQL Server; it's a layer on top of SQL Server. That's why it inherits from it. It won't work the other way around, and neither will adding AzureSQL functionality into SQLServer (since SQL server - the database - doesn't have this functionality).

2. As for this code, you're adding tasks and flows for Azure SQL, so they should be named accordingly. I don't see how this is related to Aselite - there is no Aselite source in here. Maybe you're confusing the library with how you're using it? I think it's just unnecessarily making things less generic for no reason. These tasks and flows work for any Azure SQL db, not just for Aselite (whatever that is), so there's no reason to use the less generic naming. What if I now want to extract data from Azure SQL? I need to create another task & flow, which will be exactly the same as yours, or use your aselite_to_df() task and aselite_to_adls() flow, which would be quite confusing.

   If you wanted to use the Aselite name in here, you should create an Aselite source - but there's no point since it's literally just Azure SQL. See my point?

[adrian-wojcik]

@trymzet Yes, I see your point. Names are changed. I will let Fabio know to change it in flow deploments as well.

[trymzet]

Added my comments

[trymzet]

@adrian-wojcik I think all conversations are resolved, can you rebase, make sure CI checks pass, and re-request review?

[Rafalz13]

There is only 1 pre-commit error left but couldn't fix it. I added noqa and pragma:.... It didn't help.

[trymzet]

@Rafalz13 I fixed the pre-commit checks for you. Pls fix the tests and rebase and it should be good to go.

[Rafalz13]

Done. There is a risk that bulk_insert will not work because self.credentials['data_source']', is missing in credentials but as I know this option is not used. The task uses only run from the base SQL class.

[trymzet]

Added some last comments