add command to reshape continuous questions

[lsabor]

closes #3707

adds command: python manage.py reshape_continuous_question
params:

--question_id 12345
[--make_copy]
[--alter_copy]
[--approve_copy_post]
[--nominal_range_min 2020-01-01]
[--nominal_range_max 2300-01-01]
[--convert_to_discrete]
[--step 1.0]
[--new_scheduled_close_time 2300-01-01]

This command will be rarely (hopefully essentially never) used and the code doesn't need to be polished. But it should be checked over for any logic faults.

[lsabor]

For review purposes, don't worry about the math functions regarding splines.
Reviewers can probably start around line 293

[hlbmtc]

Let's replace with questions.services.common.clone_question

[hlbmtc]

Let's move this to the separate clone_post function:

def clone_post(post: Post, **kwargs):
    ...

clone_post(post, question=new_question)

[hlbmtc]

The issue is that you keep the entire new_forecasts list in memory, which can eat a lot of RAM.

I already have a nice util for batched updates — utils.models.ModelBatchUpdater. Maybe you can create a similar helper, but for batched creation?

Something like:

class ModelBatchCreator(ModelBatchUpdater):
    def __init__(
        self,
        model_class: type[DjangoModelType],
        batch_size: int = 100,
    ):
        self.model_class = model_class
        self.batch_size = batch_size

        self._batch: list[DjangoModelType] = []

    def append(self, obj: DjangoModelType) -> None:
        self._batch.append(obj)

        if len(self._batch) >= self.batch_size:
            self.flush()

    def flush(self) -> None:
        if self._batch:
            self.model_class.objects.bulk_create(self._batch)
            self._batch.clear()

    def __enter__(self):
        return self

    def __exit__(self, exc_type, exc_value, traceback):
        self.flush()

And then use it like:

# Updating posts
with ModelBatchCreator(
    model_class=Forecast, batch_size=500
) as creator:
    for idx, forecast in enumerate(original_forecasts.iterator(chunk_size=500)):
        forecast.id = None
        forecast.pk = None
        forecast.question = new_question
        forecast.post = new_post
        creator.append(forecast)

        if idx % batch_size == 0:
            logger.info(f"Created {idx}/{total} forecasts")

[hlbmtc]

assert won't work in python production mode, please replace with explicit exceptions raise

[hlbmtc]

Let's replace it with logging

[hlbmtc]

Same here -- let's use ModelBatchUpdater

[hlbmtc]

if question.type not in QUESTION_CONTINUOUS_TYPES