visualize_chemspace is missing

[suneelbvs]

Hi,

I have been trying to plot scaffold split using visualize_chemspace (as mentioned in the tutorial). visualize_chemspace code is not there in utili.py. Is it changed?

[cwognum]

Hi @suneelbvs ! It is not part of the official Splito code-base, but rather a utility function that is defined specifically for the tutorials.

See:

splito/docs/tutorials/utils.py

Lines 14 to 23 in 0d9c2df

	def visualize_chemspace(data: pd.DataFrame, split_names: List[str], mol_col: str = "smiles", size_col=None):
	figs = plt.figure(num=3)
	features = [dm.to_fp(mol) for mol in data[mol_col]]
	embedding = umap.UMAP().fit_transform(features)
	data["UMAP_0"], data["UMAP_1"] = embedding[:, 0], embedding[:, 1]
	for split_name in split_names:
	plt.figure()
	fig = sns.scatterplot(data=data, x="UMAP_0", y="UMAP_1", style=size_col, hue=split_name, alpha=0.7)
	fig.set_title(f"UMAP Embedding of compounds for {split_name}")
	return figs

I don't think this function should be part of the official Splito code base since it is not directly relevant to splitting, but if you feel otherwise I would be happy to discuss this!