Exhaustive fragmenter saturation fix and optimization

[ToLeWeiss]

Fix for Saturation Issues in ExhaustiveFragmenter (#1119)

Aims to resolve the saturation-related issues in the ExhaustiveFragmenter, as described in issue #1119.

Fixes & Improvements

Ensures Correct Fragment Saturation:

• The generated fragments are now explicitly copied, preventing unintended modifications.
• Added the ability to set and control the saturation state of returned fragments.

Optimized Bond Splitting Logic:

• Implemented an optimization based on the observation that a specific combination of bonds only needs to be split once.
• This reduces redundant computations, improving overall efficiency.

Improvements

• Fixes inconsistencies in fragment saturation.
• Improves performance by avoiding unnecessary recomputations.
• Enhances flexibility by allowing users to specify saturation preferences.

Testing & Verification

• Verified that fragment saturation is correctly applied across different configurations.
• Tested for performance improvements with large molecular structures.
• Ensured no regressions in existing fragmentation behavior.

[JonasSchaub]

Review part I, continuation later.

[JonasSchaub]

Second and final part of first review.

[JonasSchaub]

@ToLeWeiss here is our code from Scaffold Generator that transfers stereo information: https://github.com/cdk/cdk-scaffold/blob/0847806199324735355d508c296b1b3c9c4d0ff0/src/main/java/org/openscience/cdk/tools/scaffold/ScaffoldGenerator.java#L1176

[JonasSchaub]

@ToLeWeiss another thing to think about: is there any chance to have a substructure/fragment that is there multiple times in the input structure also appear multiple times in the fragments, despite the deduplication using SMILES?
E.g. if I split C1CCCCC1C2CCCCC2, I would like to have two cyclohexanes in the output, not just one.

[JonasSchaub]

@ToLeWeiss I finished my review. Please have a look at the comments and talk to me if anything is unclear.

[JonasSchaub]

Review of ExhaustiveFragmenter class.

[JonasSchaub]

@ToLeWeiss the code is becoming really nice!
The main remaining points are the preservation of stereochemistry, introducing another constructor, and clearing up some confusion about the inclusive max tree depth (see my comments). Everything else was mostly comments.