Fix ignore functionality (especially for 'version' rules)

[CasperWA]

Fixes #130

By extending the SemanticVersion class with rich comparison magic methods, the ignore rules for versions can be more easily checked.

An extended test ensures some more specific ignore rules are respected as intended.

Dependency version ranges are still not properly respected.
For example if >=1.0.0,<2 is specified, but the latest version is 2.3.1, then the result (with no ignore rules) will be >=2.3.1,<2, which makes no sense.
This should be remedied in another issue/PR.
Edit: This issue has now been reported in #141

[codecov]

Codecov Report

Merging #140 (0cb5694) into main (7907901) will increase coverage by 5.93%.
The diff coverage is 96.66%.

@@            Coverage Diff             @@
##             main     #140      +/-   ##
==========================================
+ Coverage   65.67%   71.61%   +5.93%     
==========================================
  Files           9        9              
  Lines         472      546      +74     
==========================================
+ Hits          310      391      +81     
+ Misses        162      155       -7     

Impacted Files	Coverage Δ
ci_cd/utils.py	94.16% <96.20%> (+24.39%)	⬆️
ci_cd/tasks/update_deps.py	81.31% <100.00%> (+0.21%)	⬆️

[ajeklund]

This looks good although update_deps.py is (in my opinion) too long to understand and review efficiently. Perhaps it could be worth to clarify its central operational loop and move the logic functions into one or more separate files?

Also, Codecov generates a number of warnings about non-covered lines. Do these need to be addressed before merging?

[CasperWA]

This looks good although update_deps.py is (in my opinion) too long to understand and review efficiently. Perhaps it could be worth to clarify its central operational loop and move the logic functions into one or more separate files?

Completely fair.
I'll create an issue to follow-up on this, as I think it's out-of-scope for the current PR.

Edit: Opened #148

Also, Codecov generates a number of warnings about non-covered lines. Do these need to be addressed before merging?

Right. I'll have a look at the lot of them and see how easy it would be to check some of the non-covered lines.
The one that immediately sprung out at me was that the __hash__ function was not checked. Perhaps it can be removed altogether...

[CasperWA]

I think I got most of the codecov messages now. Unfortunately, it seems codecov has not uploaded the checks for the latest commit. But as far as I know, there is only one left, which concerns a < resolve based on pre-release information. I think that's fine, to be honest.

[ajeklund]

Excellent! Just one question added.

[ajeklund]

Thank you!