def test_max_str_len(dtype, chunks, backend, data):
shape = data.draw(st.lists(st.integers(0, 8), min_size=0, max_size=3))
ndim = len(shape)
x = data.draw(arrays(dtype=dtype if dtype != "O" else "U", shape=shape))
x = backend(x)
if dtype == "O":
x = x.astype(object)
if chunks is not None and backend is xr.DataArray:
x = x.chunk(chunks=(chunks, ) * ndim)
if chunks is not None and backend is da.array:
x = x.rechunk((chunks, ) * ndim)
if x.size == 0:
with pytest.raises(
ValueError,
match="Max string length cannot be calculated for empty array"
):
max_str_len(x)
else:
expected = max(map(len, np.asarray(x).ravel()))
actual = int(max_str_len(x))
assert expected == actual
def update_dataset(self,
ds: xr.Dataset,
add_str_max_length_attrs: bool = False) -> None:
# cyvcf2 represents missing Integer values as the minimum int32 value
# so change these to be the fill value
if self.array.dtype == np.int32:
self.array[self.array == np.iinfo(np.int32).min] = self.fill_value
ds[self.variable_name] = (self.dims, self.array)
if len(self.description) > 0:
ds[self.variable_name].attrs["comment"] = self.description
if add_str_max_length_attrs and self.array.dtype.kind == "O":
max_length = max_str_len(self.array)
ds.attrs[f"max_length_{self.variable_name}"] = max_length
def test_max_str_len__invalid_dtype():
with pytest.raises(ValueError, match="Array must have string dtype"):
max_str_len(np.array([1]))