Amboss Anki Add-On groups related cards in study sessions

When you create a study session based on previously matured cards, it seems to group all the related cards together. This can make the answers of subsequent questions obvious. (For example you answer a question on rheumatoid arthritis and the next questions are also about RA, giving away the answer in many cases). It would be nice if it randomized the order of the cards.

In performance.py:

def sample_items(items: list, max_items: int, reproducible: bool = True) -> list:
    if max_items >= len(items):
        return items
    if reproducible:
        seed(md5(dumps(items).encode("utf-8")).hexdigest())
    sampled = [items[i] for i in sorted(sample(range(len(items)), max_items))]
    if reproducible:
        seed()
    return sampled

It seems sample_items will keep the ids sorted. It seems this is done for the sake of reproducibility. However if qbank IDs from certain topics on Amboss are often grouped together, numerically, this will result in 

In controller.py, the qbank is created from a list of question_ids.

            self.start_question_session(
                question_ids=question_ids,
                statuses=self._question_statuses,
                request_origin=RequestOrigin.anki_home_qbank_widget,
                sidepanel=True,
                max_size=max_size,
            )

There's quite a lot of steps between these two pieces, but my thought is that these questions_ids are never being shuffled.