Add TorsionNet500CCSDT benchmark visualisation, MAE metric, and docs

[ellalbolland]

Pre-review checklist for PR author

• I've confirmed the contribution guidelines.

Summary

Adds per-fragment RMSE/MAE scatter plots and per-fragment torsion energy profile curves to the TorsionNet500CCSDT analysis, enabling table-cell → scatter → curve click-through navigation.

Named TorsionNet500CCSDT rather than TorsionNet500 to coexist alongside the existing barrier-height implementation in #642/#659, since this benchmark reports RMSE/MAE of the full relative energy profile against CCSD(T)/CBS reference data - a different metric and reference level of theory, not a duplicate.

Linked issue

Resolves #679

Progress

• Calculations
• Analysis
• Application
• Documentation

Testing

Tested against 16 of the 21 registered models (full 500/500-fragment recompute completed for each). The uma family (uma-s-1p1-omat, uma-m-1p1-omat, uma-s-1p1-omol, uma-s-1p2-omol, uma-m-1p1-omol) could not be run due to lacking HuggingFace authorisation for the gated UMA model weights.
Confirmed working end-to-end by running the app standalone (uv run python ml_peg/app/conformers/TorsionNet500CCSDT/app_TorsionNet500CCSDT.py) and clicking through the table -> scatter -> curve interactivity manually.

New decorators/callbacks

One new local callback, _register_curve_callback in app_TorsionNet500CCSDT.py. plot_from_scatter pre-loads every possible figure upfront (too expensive for 500 curves/model); struct_from_scatter lazy-loads but only for structures, not Plotly figures. This one lazy-loads a curve JSON from file per click instead. Flagged with a TODO to move it into build_callbacks.py if another benchmark needs the same pattern.

[joehart2001]

Thanks @ellalbolland, from a brief scan its looking great! For the uma model access you can request it i think here: https://huggingface.co/facebook/UMA
it would be good to test. i think you then also need to install the hugging face cli etc. but that should be straightforward

• please could you also send us the input files for the calc so we can upload them and test it
• also add the element filtering too

[ellalbolland]

Here's the input dataset used for the calc: TorsionNet500_xyz (1).zip

[joehart2001]

Here's the input dataset used for the calc: TorsionNet500_xyz (1).zip

thanks uploaded

[joehart2001]

this takes about 10 minutes for me locally on cpu, so lets change to less than an hour on cpu and minutes on GPU

[joehart2001]

Hi @ellalbolland, everything is looking really nice so far!

One thing that would be great to add is visualisaiton from the torsion curve plot. you could take a look at some of the NEB curves to see how they use the trajectory mode to visualise data points along the curve. thanks!