def set_attr(self, name: str, values: np.ndarray, axis: int = 0, dtype: str = None) -> None:
"""
**DEPRECATED** - Use `ds.ra.key = values` or `ds.ca.key = values` instead
"""
deprecated("'set_attr' is deprecated. Use 'ds.ra.key = values' or 'ds.ca.key = values' instead")
if axis == 0:
self.ra[name] = values
else:
self.ca[name] = values
def delete_attr(self, name: str, axis: int = 0) -> None:
"""
**DEPRECATED** - Use `del ds.ra.key` or `del ds.ca.key` instead, where `key` is replaced with the attribute name
"""
deprecated("'delete_attr' is deprecated. Use 'del ds.ra.key' or 'del ds.ca.key' instead")
if axis == 0:
del self.ra[name]
else:
del self.ca[name]
def batch_scan_layers(self, cells: np.ndarray = None, genes: np.ndarray = None, axis: int = 0, batch_size: int = 1000, layers: Iterable = None) -> Iterable[Tuple[int, np.ndarray, Dict]]:
"""
**DEPRECATED** - Use `scan` instead
"""
deprecated("'batch_scan_layers' is deprecated. Use 'scan' instead")
if cells is None:
cells = np.fromiter(range(self.shape[1]), dtype='int')
if genes is None:
genes = np.fromiter(range(self.shape[0]), dtype='int')
if layers is None:
layers = self.layers.keys()
if axis == 1:
cols_per_chunk = batch_size
ix = 0
while ix < self.shape[1]:
cols_per_chunk = min(self.shape[1] - ix, cols_per_chunk)
selection = cells - ix
# Pick out the cells that are in this batch
selection = selection[np.where(np.logical_and(selection >= 0, selection < cols_per_chunk))[0]]
if selection.shape[0] == 0:
ix += cols_per_chunk
continue
# Load the whole chunk from the file, then extract genes and cells using fancy indexing
vals = dict()
for key in layers:
vals[key] = self.layers[key][:, ix:ix + cols_per_chunk]
vals[key] = vals[key][genes, :]
vals[key] = vals[key][:, selection]
yield (ix, ix + selection, vals)
ix += cols_per_chunk
if axis == 0:
rows_per_chunk = batch_size
ix = 0
while ix < self.shape[0]:
rows_per_chunk = min(self.shape[0] - ix, rows_per_chunk)
selection = genes - ix
# Pick out the genes that are in this batch
selection = selection[np.where(np.logical_and(selection >= 0, selection < rows_per_chunk))[0]]
if selection.shape[0] == 0:
ix += rows_per_chunk
continue
# Load the whole chunk from the file, then extract genes and cells using fancy indexing
vals = dict()
for key in layers:
vals[key] = self.layers[key][ix:ix + rows_per_chunk, :]
vals[key] = vals[key][selection, :]
vals[key] = vals[key][:, cells]
yield (ix, ix + selection, vals)
ix += rows_per_chunk
def list_edges(self, *, axis: int) -> List[str]:
"""
**DEPRECATED** - Use `ds.row_graphs.keys()` or `ds.col_graphs.keys()` instead
"""
deprecated("'list_edges' is deprecated. Use 'ds.row_graphs.keys()' or 'ds.col_graphs.keys()' instead")
if axis == 0:
return self.row_graphs.keys()
elif axis == 1:
return self.col_graphs.keys()
else:
return []
def delete_attr(self, name: str, axis: int = 0) -> None:
"""
DEPRECATED
"""
deprecated(
"'delete_attr' is deprecated. Use 'del ds.ra.key' or 'del ds.ca.key' instead"
)
if axis == 0:
del self.ra[name]
else:
del self.ca[name]
def get_edges(self, name: str, *, axis: int) -> Tuple[np.ndarray, np.ndarray, np.ndarray]:
"""
**DEPRECATED** - Use `ds.row_graphs[name]` or `ds.col_graphs[name]` instead
"""
deprecated("'get_edges' is deprecated. Use 'ds.row_graphs[name]' or 'ds.col_graphs[name]' instead")
if axis == 0:
g = self.row_graphs[name]
return (g.row, g.col, g.data)
if axis == 1:
g = self.col_graphs[name]
return (g.row, g.col, g.data)
raise ValueError("Axis must be 0 or 1")
def set_layer(self,
name: str,
matrix: np.ndarray,
chunks: Tuple[int, int] = (64, 64),
chunk_cache: int = 512,
dtype: str = "float32",
compression_opts: int = 2) -> None:
"""
DEPRECATED.
"""
deprecated(
"'set_layer' is deprecated. Use 'ds.layer.Name = matrix' or 'ds.layer['Name'] = matrix' instead"
)
self.layers[name] = matrix
def set_edges(self, name: str, a: np.ndarray, b: np.ndarray, w: np.ndarray, *, axis: int) -> None:
"""
**DEPRECATED** - Use `ds.row_graphs[name] = g` or `ds.col_graphs[name] = g` instead
"""
deprecated("'set_edges' is deprecated. Use 'ds.row_graphs[name] = g' or 'ds.col_graphs[name] = g' instead")
try:
g = scipy.sparse.coo_matrix((w, (a, b)), (self.shape[axis], self.shape[axis]))
except Exception:
raise ValueError("Input arrays could not be converted to a sparse matrix")
if axis == 0:
self.row_graphs[name] = g
elif axis == 1:
self.col_graphs[name] = g
else:
raise ValueError("axis must be 0 (rows) or 1 (columns)")
