Fix composite license expression detection for AND/OR cases

[Kaushik-Kumar-CEG]

Fixes #4690

The Issue

ScanCode produced incorrect results when normalizing certain short composite license references, involving AND/OR

The "AND" Failure
* Input: Apache-2.0 AND MIT
* Old Behavior: Dropped "MIT" as noise. Output: Apache-2.0
* Expected: Apache-2.0 AND MIT

The "OR" Failure
* Input: Apache-2.0 OR MIT
* Old Behavior: Reported both the combined license AND the individual part. Output: Apache-2.0 OR MIT + Apache-2.0
* Expected: Only Apache-2.0 OR MIT

Summary of Changes

This PR addresses the issue by adding explicit license reference rules for the complete expressions
Added rules to match the full phrases:

licensed under Apache-2.0 AND MIT
licensed under Apache-2.0 OR MIT

Treating these phrases as single license references ensures that the full composite expressions are detected correctly,

Verification

Added datadriven license detection tests covering both scenarios.

The tests confirm that the expected composite license expressions are detected as intended.

I've regenerated test expectations that were directly affected by the new rules. The changes show the fix working correctly (redundant apache-2.0 detection removed). All tests pass locally. CI failures appear to be environment-related

Tasks

• Reviewed contribution guidelines
• PR is descriptively titled 📑 and links the original issue above 🔗
• Tests pass -- look for a green checkbox ✔️ a few minutes after opening your PR
• Commits are in uniquely-named feature branch and has no merge conflicts 📁
• Updated documentation pages (if applicable)
• Updated CHANGELOG.rst (if applicable)

[AyanSinhaMahapatra]

@Kaushik-Kumar-CEG IMHO the main fix for this particular issue would be to add rules.

1. you have lots of test failures, have you even checked these? you need to regenerate the test expectations and see if your changes make any sense/they break other tests
2. for a case like apache-2.0 AND (apache-2.0 OR mit) you need to ensure these are perfect matches, on the same line/right next to each other to merge these. And the expressions matter
3. You have not crafted any test so show that the issue is actually fixed.

[AyanSinhaMahapatra]

We need to add license rules for the whole texts instead:

• licensed under Apache-2.0 AND MIT
• licensed under Apache-2.0 OR MIT

[Kaushik-Kumar-CEG]

@Kaushik-Kumar-CEG IMHO the main fix for this particular issue would be to add rules.

1. you have lots of test failures, have you even checked these? you need to regenerate the test expectations and see if your changes make any sense/they break other tests
2. for a case like apache-2.0 AND (apache-2.0 OR mit) you need to ensure these are perfect matches, on the same line/right next to each other to merge these. And the expressions matter
3. You have not crafted any test so show that the issue is actually fixed.

@AyanSinhaMahapatra
Thanks for your review!

I updated the PR and my approach based on your review:

• Switched to a rules-only fix
• Added explicit AND/OR rules
• No detection logic changes
• Added data driven tests
• Verified locally (screenshots attached)
• Regenerated the test expectations (resulted in modification of 3 .yaml files automatically)

The PR is ready for review. Please let me know if any further changes are needed!

AND Case

OR Case