error while processing raw.h5ad

[dmator]

Hi, this is my first time reporting an issue on Github so please bear with me.
Please let me know if you need more information from me to address this issue:

• OS: windows 10
• Anaconda is added to PATH
• I use MobaXterm instead of Anaconda Prompt
• executed the following command some time ago so that MobaXterm will default to python3 that is installed by Anaconda:

echo "alias python=python.exe" >> ~/.bashrc
source ~/.bashrc
echo "if [[ -f ~/.bashrc ]]; then source ~/.bashrc; fi" > ~/.bash_profile

• In order to install cell x gene, I did:
  pip install cellxgene[prepare] and got an error message ("get an Environment error with a suggestion to add --user at the end"). So I did pip install cellxgene[prepare] --user, found out that cellxgene is not added to PATH. So I manually added it to PATH by following the instructions on https://www.architectryan.com/2018/03/17/add-to-the-path-on-windows-10/

Okay, that's the background.

On python, I converted mtx file to h5ad file by doing:

import scanpy as sc

adata = sc.read_10x_mtx(
'./filtered_feature_bc_matrix/',  # the directory with the `.mtx` file
var_names='gene_symbols',              
cache=True)                               
adata.write('write/10x_mHep_raw.h5ad', compression='gzip')

Then I used the following command on MobaXTerm to process the raw file:

cellxgene prepare 10x_mHep_raw.h5ad --output=10x_mHep_processed.h5ad

I got a very long error message (please see the attached txt file)
The error message ended with this:

numba.errors.LoweringError: Failed in nopython mode pipeline (step: nopython mode backend)
Operands must be the same type, got (i64, i32)

File "..\..\..\..\..\..\..\..\programdata\anaconda3\lib\site-packages\scanpy\preprocessing\_qc.py", line 392:
def _top_segment_proportions_sparse_csr(data, indptr, ns):
    <source elided>
    # Just to keep it simple, as a dense matrix
    partitioned = np.zeros((indptr.size - 1, maxidx), dtype=data.dtype)
    ^

[1] During: lowering "$0.43 = call $0.33($0.39, func=$0.33, args=[Var($0.39, c:\programdata\anaconda3\lib\site-pa                       ckages\scanpy\preprocessing\_qc.py (392))], kws=[('dtype', Var($0.41, c:\programdata\anaconda3\lib\site-packages\                       scanpy\preprocessing\_qc.py (392)))], vararg=None)" at c:\programdata\anaconda3\lib\site-packages\scanpy\preproce                       ssing\_qc.py (392)
-------------------------------------------------------------------------------
This should not have happened, a problem has occurred in Numba's internals.
You are currently using Numba version 0.45.1.

Please report the error message and traceback, along with a minimal reproducer
at: https://github.com/numba/numba/issues/new

If more help is needed please feel free to speak to the Numba core developers
directly at: https://gitter.im/numba/numba

Thanks in advance for your help in improving Numba!
[numba error.txt](https://github.com/numba/numba/files/3982574/numba.error.txt)

[dmator]

Thank you in advance for your help!

[stuartarchibald]

Thanks for the report, this will be fixed by #4729 (the issue is np.zeros() getting an int64 type and an int32 type as arguments which is currently not working).

[stuartarchibald]

Closing this issue as it seems to be resolved by #4729. If this is not the case please re-open with a comment about any item that appears to be unresolved. Many thanks.