Harmonize min-deps script with scanpy

[flying-sheep]

Before this PR, a dependency that only had != constraints broke the script

[ilan-gold]

This looks the same as before, no? The names just changed?

[flying-sheep]

No, the logic of the if statement is different. Before it checked if there were no conditions, now it checks if there are no conditions that we handle below.

The effect is that patsy!=1.0.0 now gets transformed into patsy instead of patsy==0.0.0.a1.* or so

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 84.51%. Comparing base (9f32361) to head (c9a7257).
Report is 8 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1758      +/-   ##
==========================================
- Coverage   87.01%   84.51%   -2.50%     
==========================================
  Files          40       40              
  Lines        6059     6059              
==========================================
- Hits         5272     5121     -151     
- Misses        787      938     +151     

see 8 files with indirect coverage changes

---- 🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[ilan-gold]

I guess what I was saying was that this condition looks identical but renamed

[flying-sheep]

Ah you’re right! Apprently

1. I already did that in Format sparse_dataset code #1524, then
2. I didn’t port it to scanpy, then
3. I fixed it independently in scanpy, finally
4. I blindly “backported” the changes to here by just copying scanpy’s version over

I renamed the PR to match