Merge pull request #75 from klarman-cell-observatory/yiming

Avoid item assignment on NoneType object error

pegasusio/citeseq_data.py

@@ -23,20 +23,20 @@ def __init__(
         feature_metadata: Union[dict, pd.DataFrame],
         matrices: Dict[str, csr_matrix],
         metadata: dict,
-        barcode_multiarrays: Dict[str, np.ndarray] = None,
-        feature_multiarrays: Dict[str, np.ndarray] = None,
-        barcode_multigraphs: Dict[str, csr_matrix] = None,
-        feature_multigraphs: Dict[str, csr_matrix] = None,
+        barcode_multiarrays: Dict[str, np.ndarray] = dict(),
+        feature_multiarrays: Dict[str, np.ndarray] = dict(),
+        barcode_multigraphs: Dict[str, csr_matrix] = dict(),
+        feature_multigraphs: Dict[str, csr_matrix] = dict(),
         cur_matrix: str = "raw.count",
     ) -> None:
         assert metadata["modality"] == "citeseq"
         super().__init__(barcode_metadata, feature_metadata, matrices, metadata, barcode_multiarrays, feature_multiarrays, barcode_multigraphs, feature_multigraphs, cur_matrix)
-        
+
 
     def from_anndata(self, data: anndata.AnnData, genome: str = None, modality: str = None) -> None:
         raise ValueError("Cannot convert an AnnData object to a CITESeqData object!")
 
-    
+
     def to_anndata(self) -> anndata.AnnData:
         raise ValueError("Cannot convert a CITESeqData object ot an AnnData object!")
 
@@ -52,7 +52,7 @@ def set_aside(self, params: List[str]) -> None:
         assert len(self.matrices) == 1 and "raw.count" in self.matrices
         assert "_other_names" not in self.metadata
 
-        locs = self.feature_metadata.index.get_indexer(params) 
+        locs = self.feature_metadata.index.get_indexer(params)
         if (locs < 0).sum() > 0:
             raise ValueError(f"Detected unknown antibodies {params[locs < 0]}!")
         self.metadata["_other_names"] = self.feature_metadata.index.values[locs] # with loc: List[int], this should be a copy not a reference
@@ -61,17 +61,17 @@ def set_aside(self, params: List[str]) -> None:
         obs_keys = self.metadata.get("_obs_keys", [])
         obs_keys.append("_other_counts")
         self.metadata["_obs_keys"] = obs_keys
-        
+
         idx = np.ones(self._shape[1], dtype = bool)
         idx[locs] = False
         self._inplace_subset_var(idx)
 
 
     def arcsinh_transform(self, cofactor: float = 5.0, jitter = False, random_state = 0, select: bool = True) -> None:
         """Conduct arcsinh transform on the raw.count matrix.
-        
+
         Add arcsinh transformed matrix 'arcsinh.transformed'. If jitter == True, instead add a 'arcsinh.jitter' matrix in dense format, jittering by adding a randomized value in U([-0.5, 0.5)). Mimic Cytobank.
-        
+
         Parameters
         ----------
         cofactor: ``float``, optional, default: ``5.0``

pegasusio/cyto_data.py

@@ -22,10 +22,10 @@ def __init__(
         feature_metadata: Union[dict, pd.DataFrame],
         matrices: Dict[str, np.ndarray],
         metadata: dict,
-        barcode_multiarrays: Dict[str, np.ndarray] = None,
-        feature_multiarrays: Dict[str, np.ndarray] = None,
-        barcode_multigraphs: Dict[str, csr_matrix] = None,
-        feature_multigraphs: Dict[str, csr_matrix] = None,
+        barcode_multiarrays: Dict[str, np.ndarray] = dict(),
+        feature_multiarrays: Dict[str, np.ndarray] = dict(),
+        barcode_multigraphs: Dict[str, csr_matrix] = dict(),
+        feature_multigraphs: Dict[str, csr_matrix] = dict(),
         cur_matrix: str = "raw.data",
     ) -> None:
         assert metadata["modality"] == "cyto"
@@ -35,7 +35,7 @@ def __init__(
     def from_anndata(self, data: anndata.AnnData, genome: str = None, modality: str = None) -> None:
         raise ValueError("Cannot convert an AnnData object to a CytoData object!")
 
-    
+
     def to_anndata(self) -> anndata.AnnData:
         raise ValueError("Cannot convert a CytoData object ot an AnnData object!")
 
@@ -51,7 +51,7 @@ def set_aside(self, params: List[str] = ["Time"]) -> None:
         assert len(self.matrices) == 1 and "raw.data" in self.matrices
         assert "_parameter_names" not in self.metadata
 
-        locs = self.feature_metadata.index.get_indexer(params) 
+        locs = self.feature_metadata.index.get_indexer(params)
         if (locs < 0).sum() > 0:
             raise ValueError(f"Detected unknown parameters {params[locs < 0]}!")
         self.barcode_multiarrays["_parameters"] = self.matrices["raw.data"][:, locs]
@@ -63,9 +63,9 @@ def set_aside(self, params: List[str] = ["Time"]) -> None:
 
     def arcsinh_transform(self, cofactor: float = 5.0, jitter = False, random_state = 0, select: bool = True) -> None:
         """Conduct arcsinh transform on the raw.count matrix.
-        
+
         Add arcsinh transformed matrix 'arcsinh.transformed'. If jitter == True, instead add a 'arcsinh.jitter' matrix in dense format, jittering by adding a randomized value in U([-0.5, 0.5)). Mimic Cytobank.
-        
+
         Parameters
         ----------
         cofactor: ``float``, optional, default: ``5.0``

pegasusio/nanostring_data.py

@@ -20,10 +20,10 @@ def __init__(
         feature_metadata: Union[dict, pd.DataFrame],
         matrices: Dict[str, np.ndarray],
         metadata: dict,
-        barcode_multiarrays: Dict[str, np.ndarray] = None,
-        feature_multiarrays: Dict[str, np.ndarray] = None,
-        barcode_multigraphs: Dict[str, csr_matrix] = None,
-        feature_multigraphs: Dict[str, csr_matrix] = None,
+        barcode_multiarrays: Dict[str, np.ndarray] = dict(),
+        feature_multiarrays: Dict[str, np.ndarray] = dict(),
+        barcode_multigraphs: Dict[str, csr_matrix] = dict(),
+        feature_multigraphs: Dict[str, csr_matrix] = dict(),
         cur_matrix: str = "Q3Norm",
     ) -> None:
         assert metadata["modality"] == "nanostring"
@@ -33,7 +33,7 @@ def __init__(
     def from_anndata(self, data: anndata.AnnData, genome: str = None, modality: str = None) -> None:
         raise ValueError("Cannot convert an AnnData object to a NanostringData object!")
 
-    
+
     def to_anndata(self) -> anndata.AnnData:
         raise ValueError("Cannot convert a NanostringData object ot an AnnData object!")
 
@@ -76,9 +76,9 @@ def norm_hk(self, select: bool = True) -> None:
 
     def log_transform(self, select: bool = True) -> None:
         """Conduct log transformation on the selected matrix: log(x + 1). Selected matrix can be either Q3Norm or HKNorm
-        
+
         Add log-transformed matrix 'LogMatrix'.
-        
+
         Parameters
         ----------
         select: ``bool``, optional, default: ``True``

pegasusio/unimodal_data.py

@@ -37,14 +37,14 @@ def _set_modality(metadata: dict, modality: str):
 class UnimodalData:
     def __init__(
         self,
-        barcode_metadata: Union[dict, pd.DataFrame, anndata.AnnData] = None,
-        feature_metadata: Union[dict, pd.DataFrame] = None,
-        matrices: Dict[str, csr_matrix] = None,
-        metadata: dict = None,
-        barcode_multiarrays: Dict[str, np.ndarray] = None,
-        feature_multiarrays: Dict[str, np.ndarray] = None,
-        barcode_multigraphs: Dict[str, csr_matrix] = None,
-        feature_multigraphs: Dict[str, csr_matrix] = None,
+        barcode_metadata: Union[dict, pd.DataFrame, anndata.AnnData] = pd.DataFrame(),
+        feature_metadata: Union[dict, pd.DataFrame] = pd.DataFrame(),
+        matrices: Dict[str, csr_matrix] = dict(),
+        metadata: dict = dict(),
+        barcode_multiarrays: Dict[str, np.ndarray] = dict(),
+        feature_multiarrays: Dict[str, np.ndarray] = dict(),
+        barcode_multigraphs: Dict[str, csr_matrix] = dict(),
+        feature_multigraphs: Dict[str, csr_matrix] = dict(),
         cur_matrix: str = "X",
         genome: str = None,
         modality: str = None,
@@ -55,15 +55,11 @@ def __init__(
             self.from_anndata(barcode_metadata, genome = genome, modality = modality)
             return None
 
-        def replace_none_df(value):
-            return value if value is not None else pd.DataFrame()
-
-
         _set_genome(metadata, genome)
         _set_modality(metadata, modality)
 
-        self.barcode_metadata = replace_none_df(barcode_metadata) # barcode metadata
-        self.feature_metadata = replace_none_df(feature_metadata) # feature metadata
+        self.barcode_metadata = barcode_metadata # barcode metadata
+        self.feature_metadata = feature_metadata # feature metadata
 
         if len(self.barcode_metadata) > 0:
             if isinstance(self.barcode_metadata, MutableMapping):
@@ -594,7 +590,7 @@ def _copy_view(self, viewobj: UnimodalDataView) -> "UnimodalData":
                             viewobj.obsm[...],
                             viewobj.varm[...],
                             viewobj.obsp[...],
-                            viewobj.varp[...],                         
+                            viewobj.varp[...],
                             viewobj._cur_matrix)
 
 

pegasusio/vdj_data.py

@@ -66,10 +66,10 @@ def __init__(
         feature_metadata: Union[dict, pd.DataFrame],
         matrices: Dict[str, csr_matrix],
         metadata: dict,
-        barcode_multiarrays: Dict[str, np.ndarray] = None,
-        feature_multiarrays: Dict[str, np.ndarray] = None,
-        barcode_multigraphs: Dict[str, csr_matrix] = None,
-        feature_multigraphs: Dict[str, csr_matrix] = None,
+        barcode_multiarrays: Dict[str, np.ndarray] = dict(),
+        feature_multiarrays: Dict[str, np.ndarray] = dict(),
+        barcode_multigraphs: Dict[str, csr_matrix] = dict(),
+        feature_multigraphs: Dict[str, csr_matrix] = dict(),
         cur_matrix: str = "umis",
     ) -> None:
         assert metadata["modality"] in {"tcr", "bcr"}
@@ -131,7 +131,7 @@ def construct_clonotype(self, min_umis: int = 2) -> None:
     def from_anndata(self, data: anndata.AnnData, genome: str = None, modality: str = None) -> None:
         raise ValueError("Cannot convert an AnnData object to a VDJData object!")
 
-    
+
     def to_anndata(self) -> anndata.AnnData:
         raise ValueError("Cannot convert a VDJData object ot an AnnData object!")