Uff python bindings

[scal444]

No description provided.

[greptile-apps]

Greptile Summary

This PR adds GPU-accelerated UFF force field optimization mirroring the existing MMFF implementation, and takes the opportunity to consolidate previously duplicated utility code (vectorToList, extractMolecules, ThreadLocalBuffers, etc.) into shared headers (boost_python_utils.h, mmff_python_utils.h, bfgs_common.h/cpp). The refactoring is clean and the new UFF path is structurally sound.

Confidence Score: 5/5

• Safe to merge; the one finding is a defensive-coding suggestion with no current impact.
• No P0 or P1 issues found. The only comment is a P2 on the newly shared extractMMFFPropertiesList being more lenient than the old strict check, but the Python wrapper always supplies exactly numMols entries so the looser C++ validation has no real-world impact. Prior concerns (Any-typed stubs, unchecked EmbedMultipleConfs) are resolved.
• No files require special attention.

Important Files Changed

Filename	Overview
src/minimizer/bfgs_uff.cpp	New UFF BFGS minimizer; mirrors bfgs_mmff.cpp closely with correct per-molecule vdwThreshold/ignoreInterfrag indexing via confInfo.molIdx. Uses shared bfgs_common infrastructure. No issues found.
src/minimizer/bfgs_common.cpp	New shared BFGS infrastructure extracted from bfgs_mmff.cpp. ensureCapacity now uses energiesSize1.3 for the energy buffer (previously used positionsSize1.3), which is a correct improvement. Thread-safety for writeBackResults is preserved.
nvmolkit/uffOptimization.py	Python wrapper with solid input validation (None + UFFHasAllMoleculeParams checks) and scalar/sequence normalization. Mirrors mmffOptimization.py structure. No issues found.
nvmolkit/boost_python_utils.h	New shared header consolidating previously duplicated vectorToList, vectorOfVectorsToList, extractMolecules, extractDoubleList, extractBoolList helpers from multiple .cpp files. Clean refactoring.
nvmolkit/mmff_python_utils.h	New header for MMFF Python utilities. extractMMFFPropertiesList silently fills missing entries with defaults and ignores excess ones — lenient behavior vs. the strict equality check that was previously in mmffOptimization.cpp. Python wrapper always supplies exactly numMols items, so in practice this is benign.

_{Reviews (3): Last reviewed commit: "Merge branch 'main' into uff_python" | Re-trigger Greptile}

[evasnow1992]

Changes look good to me. Two minor commends regarding code duplication. Also we are missing updating docs/index.rst and docs/api/nvmolkit.rst with UFF. Would prefer to have them together with the current PR but not required.

[scal444]

Updated docs