CHANG-Shaole

Thank you for your in-time reply! I will give it a try.

Sorry for bothering you again. It seems that the links of exp data-[Baron_human.h5ad](http://download.gao-lab.org/Cell_BLAST/tutorial/Baron_human.h5ad) and [Lawlor.h5ad](http://download.gao-lab.org/Cell_BLAST/tutorial/Lawlor.h5ad) are invalid. I would much appreciate it if you could give me another link for...

Thank you for the reply! Recently I tried to follow your [demo](https://cblast.readthedocs.io/en/latest/BLAST.html#Preparing-database). But whenever I run such a block: ``` blast = cb.blast.BLAST(models, baron_human) ``` it came with an error:...

Hi! Thank you for your reply. I haven't completed that. But I have some ideas for that. We need to re-write some code to adapt to the issue. We need...

Hi, @youcefBouraoui Not very clear about your problem. But I remember after extracting the pattern feat, we need to concatenate it in the integrated data files for the EGAT train/predict....

> Sorry for the late reply. To run inference, you need first prepare a molecule graph and predict bond disconnection using EGAT. Once you have the disconnection, edit the product...

Hi, @chaoyan1037 Sorry for bothering you again. There is another problem with [canonicalize_products.py](https://github.com/uta-smile/RetroXpert/blob/main/canonicalize_products.py). I tried the rxn_smiles without mapping numbers: `OB(O)C1=C(OC)C=C(C(F)(F)F)C=C1.O=C(N1)NN=C(Br)C1=O>>O=C(N1)NN=C(C2=CC=C(C(F)(F)F)C=C2OC)C1=O` after [canonicalize_products.py](https://github.com/uta-smile/RetroXpert/blob/main/canonicalize_products.py), I got `COc1cc(C(F)(F)F)ccc1B(O)O.O=c1[nH]nc(Br)c(=O)[nH]1>>[CH3:1][O:2][c:3]1[cH:4][c:5]([C:6]([F:7])([F:8])[F:9])[cH:10][cH:11][c:12]1-[c:13]1[n:14][nH:15][c:16](=[O:17])[nH:18][c:19]1=[O:20]` I am not sure...