Add unpaginated versions of multi-row get methods (w/tests)

[slaarti]

Description

Per #59, this PR is to add new methods to FastCRUD to do multi-row gets (without and with joins) that return all rows in a single list without pagination.

Changes

Added new methods get_multi_all and get_multi_joined_all that are basically copies minus the offset, limit, and return_total_count args as well as without wrapping the list of result rows in a dict (since that happens in the existing methods to help with pagination).

Tests

test_get_multi.py and test_get_multi_joined.py for both sqlalchemy and sqlmodel were copied into new files and adjusted to test the new methods (including removing tests that only made sense in the context of pagination).

Checklist

• I have read the CONTRIBUTING document.
• My code follows the code style of this project.
• I have added necessary documentation (if appropriate).
• I have added tests that cover my changes (if applicable).
• All new and existing tests passed.

[igorbenav]

@slaarti could you please refactor? I think doing it this way would be bettet indeed

[slaarti]

Okay, sure. Not tonight, might take a couple days, but I'm fine with the alternate approach. 👍

[igorbenav]

Nice, thanks!

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (5e45aee) to head (661357a).

Additional details and impacted files

@@            Coverage Diff            @@
##              main       #62   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           64        64           
  Lines         4282      4336   +54     
=========================================
+ Hits          4282      4336   +54     

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

[slaarti]

Okay, I rolled back the original branch (well, renamed it and did a fresh checkout) and implemented the alternate method, including checks on the endpoint creator and crud router (with a flag to bypass the checks). The force push is for the new branch.

[igorbenav]

Sorry, could you refactor it to use None instead of 0? I think someone using 0 to retrieve no record might make sense (think you have a limit for someone to retrieve x objects, he might get 10 objects, then 5 more, you subtract it from x until it becomes zero, then the person can't retrieve).

It's a dumb example, but I think it makes more sense with limit=None.

[slaarti]

I'd opted for 0 as much to avoid having to deal with changes in typing and looking for None vs. 0 as feeling like saying you want zero rows made no sense... but yeah, okay, you make a good point. Will do. 👍

[slaarti]

...Actually, come to think of it, an interesting aspect of changing it to None is that I'm pretty sure there's no good way to pass None as a path/query parameter, which eliminates the need for the changes made for sanitizing the automatically generated endpoints.

Anyone who wants to provide an unlimited get in their (presumably private, as in my cases) API will have to create an endpoint specifically for it, but I'm okay with that.

[igorbenav]

I guess everything worked out even better, great!
Thanks @slaarti and welcome to fastcrud contributors🎉