pySIML gives you easy access to SIML, an extremely fast method for computing LINGO chemical similarities [Vidal05]. There are existing implementations of LINGO, such as that included in OpenEye’s OEChem toolkit - so why bother with SIML?
Speed.
For several classes of important problems, SIML is significantly faster than existing implementations of LINGOs. In particular, for M x N similarity problems, in which one needs to compare every molecule in a set of size M against every other molecule in a set of size N (and M and N are reasonably large, on the order of hundreds), SIML on a single-core CPU is several times as fast as existing implementations of LINGO. SIML also supports computing LINGOs on a CUDA-capable GPU, which allows over eightyfold speedup relative to even fast CPUs.
SIML and pySIML live at https://simtk.org/home/siml.