Detect if adata.X is log normalised or not.

Hi

I wanted to know how the check ‘Did not modify X as it looks preprocessed already’ is working. In the code it is comparing counts of spliced and X. log_advised = np.allclose(adata.X[:10].sum(), adata.layers['spliced'][:10].sum())

Can you comment on why X would be log normalised if these counts are equal?

[VolkerBergen]

Would be quite heuristic and covering only few cases. Since log is common preprocessing procedure, I’d assume that the user has done so if he merges his pre-analysed data.

The most important cases are already covered in filter_and_normalize, see https://github.com/theislab/scvelo/blob/d64a55d455cfaab0e61c8c41b7528216263720cd/scvelo/preprocessing/utils.py#L410

[saksham219]

A simple method can check if data was log scaled. For checking if data is log scaled we can simply take the range of counts ; max(X) - min(X). If this range is greater than 100(another assumption) it can be said that the counts have not been log transformed I agree with your point that if the data was scaled to a unit variance beforehand, this method would fail.