ov.pp.qc()

[zhaofy123]

Describe the bug
KeyError: 'False: boolean label can not be used without a boolean index'

here is report

---

KeyError Traceback (most recent call last)
File :1

File ~/anaconda3/envs/omicvs/lib/python3.10/site-packages/omicverse/pp/_qc.py:195, in qc(adata, **kwargs)
193 if settings.mode == 'gpu':
194 print('GPU mode activated')
--> 195 return qc_gpu(adata,**kwargs)
196 else:
197 print('CPU mode activated')

File ~/anaconda3/envs/omicvs/lib/python3.10/site-packages/omicverse/pp/_qc.py:464, in qc_gpu(adata, mode, min_cells, min_genes, nmads, max_cells_ratio, max_genes_ratio, batch_key, doublets, doublets_method, path_viz, tresh, mt_startswith, mt_genes)
462 # QC plot
463 QC_test = (adata.obs['passing_mt']) & (adata.obs['passing_nUMIs']) & (adata.obs['passing_ngenes'])
--> 464 removed = QC_test.loc[lambda x : x is False]
465 removed_cells.extend(list(removed.index.values))
466 print(f'Total cell filtered out with this last --mode {mode} QC (and its chosen options):
467 {n1-np.sum(QC_test)}')

File ~/anaconda3/envs/omicvs/lib/python3.10/site-packages/pandas/core/indexing.py:1191, in _LocationIndexer.getitem(self, key)
1189 maybe_callable = com.apply_if_callable(key, self.obj)
1190 maybe_callable = self._check_deprecated_callable_usage(key, maybe_callable)
-> 1191 return self._getitem_axis(maybe_callable, axis=axis)

File ~/anaconda3/envs/omicvs/lib/python3.10/site-packages/pandas/core/indexing.py:1430, in _LocIndexer._getitem_axis(self, key, axis)
...
1244 isinstance(key.start, bool) or isinstance(key.stop, bool)
1245 ):
1246 raise TypeError(f"{key}: boolean values can not be used in a slice")

[zhaofy123]

I find this is a really interesting bug.It is only happened when i use 'gpu mode' because everything is ok if i choose 'cpu mode'.My gpu is RTX4070tisuper, and the cuda version is 12.4.

[aaa7260]

每次使用GPU就会卡在这个界面，很久都没反应，但是使用rsc就不会

[zhaofy123]

所以ov的预处理环节，你们使用gpu是没有出现类似的报错吗，想知道你使用的的版本。

[aaa7260]

所以ov的预处理环节，你们使用gpu是没有出现类似的报错吗，想知道你使用的的版本。

我使用的是omicverse 1.6.10，我按照你的源代码，在python里分开执行，rsc.pp.highly_variable_genes(adata, n_top_genes=3000, flavor="pearson_residuale",layer='counts')这一步过不去，不指定layers并且使用flavor=‘seurat_V3’就可以继续下去，使用layer=‘counts’，会提示TypeError: The input is not a CuPy ndarray or CuPy sparse matrix. Rapids-singlecell only supports GPU matrices in this function, so your input must be either a CuPy ndarray or a CuPy sparse matrix. Please checkout rapids_singlecell.get.anndata_to_GPU to convert your data to GPU. If you're working with CPU-based matrices, please consider using Scanpy instead.
使用flavor="pearson_residuale"会flavor="pearson_residuale"

感谢大佬的回答

[Starlitnightly]

Maybe this issue caused by the updated of rapid_single_cell