invalid class “dgRMatrix” object: slot j is not increasing inside a column

[scottgigante]

I'm getting an error in scipy2ri.py2r on a CSR matrix.

scIB.preprocessing.normalize(adata)

----------------------
  File "/home/travis/virtualenv/python3.7-dev/lib/python3.7/site-packages/scIB/preprocessing.py", line 168, in normalize
    ro.globalenv['data_mat'] = adata.X.T
  File "/home/travis/virtualenv/python3.7-dev/lib/python3.7/site-packages/rpy2/robjects/environments.py", line 32, in __setitem__
    robj = conversion.converter.py2rpy(value)
  File "/opt/python/3.7-dev/lib/python3.7/functools.py", line 840, in wrapper
    return dispatch(args[0].__class__)(*args, **kw)
  File "/home/travis/virtualenv/python3.7-dev/lib/python3.7/site-packages/anndata2ri/scipy2ri/py2r.py", line 43, in wrapper
    return f(obj)
  File "/home/travis/virtualenv/python3.7-dev/lib/python3.7/site-packages/anndata2ri/scipy2ri/py2r.py", line 70, in csr_to_rmat
    Dim=as_integer(list(csr.shape)),
  File "/home/travis/virtualenv/python3.7-dev/lib/python3.7/site-packages/rpy2/robjects/functions.py", line 198, in __call__
    .__call__(*args, **kwargs))
  File "/home/travis/virtualenv/python3.7-dev/lib/python3.7/site-packages/rpy2/robjects/functions.py", line 125, in __call__
    res = super(Function, self).__call__(*new_args, **new_kwargs)
  File "/home/travis/virtualenv/python3.7-dev/lib/python3.7/site-packages/rpy2/rinterface_lib/conversion.py", line 44, in _
    cdata = function(*args, **kwargs)
  File "/home/travis/virtualenv/python3.7-dev/lib/python3.7/site-packages/rpy2/rinterface.py", line 624, in __call__
    raise embedded.RRuntimeError(_rinterface._geterrmessage())
rpy2.rinterface_lib.embedded.RRuntimeError: Error in validObject(.Object) : 
  invalid class “dgRMatrix” object: slot j is not increasing inside a column

[flying-sheep]

Why is adata.X.T.indices not sorted? Shouldn’t that be guaranteed?

[scottgigante]

indices is not, but indptr is.

import tempfile
import anndata
import os
import scprep
import numpy as np

URL = (
    "https://ndownloader.figshare.com/files/23992451?private_link=e3921450ec1bd0587870"
)
with tempfile.TemporaryDirectory() as tempdir:
    filepath = os.path.join(tempdir, "zebrafish.h5ad")
    scprep.io.download.download_url(URL, filepath)
    adata = anndata.read_h5ad(filepath)

np.all(adata.X.indices == np.sort(adata.X.indices))
# False
np.all(adata.X.indptr == np.sort(adata.X.indptr))
# True
np.all(adata.X.T.indices == np.sort(adata.X.T.indices))
# False
np.all(adata.X.T.indptr == np.sort(adata.X.T.indptr))
# True

[ivirshup]

They're not always sorted, but there are a couple methods for checking and sorting these: csr_matrix.has_sorted_indices, csr_matrix.sort_indices, etc.

[ivirshup]

Just a note for future reference, the results of those comparisons is pretty reasonable. The check for sorted-ness of indices should probably look more like:

all(np.array_equal(x, np.sort(x)) for x in np.split(s.indices, s.indptr[1:-1]))

And indptr is the only thing that has to be sorted.