def do_bulk_inference(
self,
model_name: str,
objects: List[dict],
top_n: int = TOP_N,
retry: bool = False,
) -> List[dict]:
"""
Performs bulk inference for larger collections.
For *objects* collections larger than *LIMIT_OBJECTS_PER_CALL*, splits
the data into several smaller Inference requests.
Returns the aggregated values of the *predictions* of the original API response
as returned by :meth:`create_inference_request`.
:param model_name: name of the model used for inference
:param objects: Objects to be classified
:param top_n: How many predictions to return per object
:return: the aggregated ObjectPrediction dictionaries
"""
result = [] # type: List[dict]
for work_package in split_list(objects, LIMIT_OBJECTS_PER_CALL):
response = self.create_inference_request(model_name,
work_package,
top_n=top_n,
retry=retry)
result.extend(response["predictions"])
return result
def test_slice_size_invalid(self):
for invalid_slice_size in [-1000, -1, 0]:
with pytest.raises(ValueError):
list(split_list(["a", "b"], invalid_slice_size))
def test_slice_size_bigger_than_list(self):
res = list(split_list(["a", "b", "c", "d"], 6))
assert res == [["a", "b", "c", "d"]]
def test_list_uneven(self):
res = list(split_list(["a", "b", "c", "d"], 3))
assert res == [["a", "b", "c"], ["d"]]
def test_regular_case(self):
res = list(split_list(["a", "b", "c", "d"], 2))
assert res == [["a", "b"], ["c", "d"]]
def test_empty_list(self):
res = list(split_list([], slice_size=1))
assert res == [[]]