Add initial examples

[ElliottKasoar]

Adds example benchmarks for OC157 and Li diffusion barriers (NEB).

To do:

• Fix callback for Li diffusion atom visualisation
• Fix sorting/order of models/storage of data (use dictionary rather than list)
• Add benchmark with multiple tables

[ThomasWarford]

Hi,

the bug I noticed was that points in the parity plots for S24 are given the wrong labels

The following XYZ (found at mlip_testing/app/data/surfaces/S24/mace_mp_0b3/C48H24O26Zn8-H2..csv) has ref adsorption energy of -0.1.

NOTE: I save it as csv because GitHub won't let me upload xyz, change the file extension if you want.
<img width="198" he
C48H24O26Zn8-H2..csv

The point with label C48H24O26Zn8-H2 has very different energy.
ight="73" alt="image" src="https://github.com/user-attachments/assets/103528b8-181f-49c5-806a-801dcbbc9363" />

I fixed this on my branch by changing the get_system_names function in mlip_testing/analysis/surfaces/S24/analyse_S24.py

def system_names() -> list:
    """
    Get list of system names.

    Returns
    -------
    list
        List of all system names.
    """
    # NOTE: The parity plot associates hover data with points in the
    # order the data arrays are provided. In `adsorption_energies` the
    # systems are iterated with `sorted((CALC_PATH / model_name).glob("*.xyz"))`.
    # Originally this function used the unsorted `.glob()` ordering,
    # which can differ from the sorted order, leading to mismatched
    # (mislabelled) hover labels (system names) in the output JSON.
    # We therefore reproduce the exact ordering used in the data
    # collection step by applying the same `sorted(...)` here.
    system_names_list = []
    for model_name in MODELS:
        for system_path in sorted((CALC_PATH / model_name).glob("*.xyz")):
            system_names_list.append(system_path.stem)
        break  # only need the first model to list available systems
    return system_names_list
    ```

[ThomasWarford]

I have not confirmed this fix for S24, but I did for the branch I made, which copied this code, maybe the function should be moved into another file without the global variable dependence.