feat(projects): Improve project relationship iteration with VersionComparator

[Ankush1oo8]

Please provide a summary of your changes here.

• Added VersionComparator in common/utils for accurate version sorting
• Updated iterateProjectRelationShips methods to use VersionComparator
• Improved sorting logic for project links based on name and version
• Ensured compatibility with existing project relationship handling

Issue: #236

Suggest Reviewer

How To Test?

1. Run the SW360 backend and verify project relationships are sorted correctly.
2. Ensure no functionality is broken by running tests.
3. Validate the sorting logic using different project versions.

Checklist

Must:

• All related issues are referenced in commit messages and in PR

[GMishx]

A small change is still required.

[Ankush1oo8]

@GMishx can you check now

[Ankush1oo8]

Why This Works

If part = "43-2+deb8u2", numericPrefix = "43", which can be parsed safely.

If part has no numeric prefix (e.g., "deb8u2"), it gets a default -1 for lower priority.

It ensures that versions like "3.16.43-2+deb8u2" are still sorted meaningfully.

[Ankush1oo8]

@GMishx can you check now

[Ankush1oo8]

@GMishx should I also optimise for version with aplha or beta

[GMishx]

Why This Works

If part = "43-2+deb8u2", numericPrefix = "43", which can be parsed safely.

If part has no numeric prefix (e.g., "deb8u2"), it gets a default -1 for lower priority.

It ensures that versions like "3.16.43-2+deb8u2" are still sorted meaningfully.

I agree but current solution works for cases like 1.2.2 < 1.2.3

Will work for the Debian version strings.

But would say v1.3.2 > v2.1.1

The solution should be generic enough.

[Ankush1oo8]

@GMishx I have also added for
Handles epoch versions (1:2.3.4 vs 2.3.4) correctly
Ignores build metadata (1.2.3+build.4 == 1.2.3)
Normalizes leading zeros (01.02.003 == 1.2.3)
Supports configurable pre-release precedence (alpha < beta < rc)
Optimized with caching for performance

[GMishx]

I'd call it "over-engineered" solution to a simple problem of returning 0 if the conversion to int failed.

However, it looks good to me. But I'd like to have second opinion from @shi9qiu or @hoangnt2

[GMishx]

@shi9qiu can you please help test / review this branch.