Adrian Henle

It would be nice to get a workaround on this... Downgrading to an obsolete version of PyTorch isn't an option for me, and this is breaking my CI. And this...

> Let's use Julia 1.8.0 !!! It seems our issue is solved. > ... > LIBM: libopenlibm > LLVM: libLLVM-13.0.1 (ORCJIT, broadwell) > Threads: 1 on 4 virtual cores >...

I have the same issue.

Having done this, I wouldn't recommend it to users. I wouldn't *discourage* anyone from doing it, but I wouldn't *advertise* it either. After doing this, the user has to restart...

Here is the complete list of MACCS rules that are returning false-positive for the molecule shown above. Each rule is a tuple that gives the SMARTS query and the count...

Notebook showing benchmarks: [🎈 Fun lecture.jl — Pluto.jl.pdf](https://github.com/mojaie/MolecularGraph.jl/files/9601607/Fun.lecture.jl.Pluto.jl.pdf) I think the best solution is to store the correct bond aromaticity when the `GraphMol` is created. It will make object instantiation...

Another version of the notebook with the previous example SMILES encoded with explicit aromaticity and a large fused ring example: [🎈 Surprising blueprint.jl — Pluto.jl.pdf](https://github.com/mojaie/MolecularGraph.jl/files/9604532/Surprising.blueprint.jl.Pluto.jl.pdf) The explicit aromaticity encoding makes...

 Explicitly encoding the aromaticity in the SMILES input has no effect on the information stored in the `GraphMol` struct. More important, though, is the fact that by defaulting to...

 That still does not change the values stored in `Bond.isaromatic`, and it is still incorrect either way to say that a bond is "specified" as *not aromatic* in SMILES...

 Confirmed that using explicit aromatic bonds in SMILES string gives the correct result; I must have made a mistake before, sorry. > Please consider to use parse(SMILES, mol) instead...