ebi-gene-expression-group/scxa-workflows

cannot open h5ad

SilasK opened this issue · 1 comments

Hello, I was very happy to see that you make for all experiments h5ad objects available.

However, I tried to open the one and got the following error:

(self, X, obs, var, uns, obsm, varm, varp, obsp, raw, layers, dtype, shape, filename, filemode)
    446 else:
    447     class_names = ", ".join(c.__name__ for c in StorageType.classes())
--> 448     raise ValueError(
    449         f"`X` needs to be of one of {class_names}, not {type(X)}."
    450     )
    451 if shape is not None:
    452     raise ValueError("`shape` needs to be `None` if `X` is not `None`.")

ValueError: `X` needs to be of one of ndarray, MaskedArray, spmatrix, Array, ZappyArray, Array, not <class 'str'>.

I use the latest anndata version 0.8 and py 3.8

pcm32 commented

Hi, apologies, just saw this. So our galaxy setup is still most likely producing AnnData 0.7.x, could that be the culprit? We deposit all the AnnData files per release per dataset here http://ftp.ebi.ac.uk/pub/databases/microarray/data/atlas/sc_experiments/ (just go inside one of the directories, you should find an .h5ad file there), in case it could be a case of a specific anndata file being corrupted.