TypeError: object of type 'NoneType' has no len()

Sentry Issue: SIR-LANCEBOT-8R

TypeError: object of type 'NoneType' has no len()
  File "discord/ext/commands/core.py", line 190, in wrapped
    ret = await coro(*args, **kwargs)
  File "bot/exts/holidays/halloween/scarymovie.py", line 23, in random_movie
    selection = await self.select_movie()
  File "bot/exts/holidays/halloween/scarymovie.py", line 50, in select_movie
    selection_id = random.choice(data.get("results")).get("id")
  File "random.py", line 378, in choice
    return seq[self._randbelow(len(seq))]

Unhandled command error: object of type 'NoneType' has no len()

[TizzySaurus]

I don’t think there’s any reason it should be None so probably send a message to the user saying no results were found and maybe log a warning so we can look into it.

Yep, that sounds good.

If we used the results from the first request we’d be much more limited in the results we got.

Right, I see. For some reason I was thinking that the initial response would include all entries, which isn’t actually the case as you stated (it seems to return 20).

Side note: we may also want a fix like https://github.com/python-discord/sir-lancebot/issues/1055, it’s the same API and similar use case, just in a different place. It’s possible it isn’t an issue in practice for this genre if the total count isn’t high enough though.

It seems like a smart thing to do for future-proofing. Seems it needs to be limited to a max of 500 (not the 1000 stated on the API’s docs)?.

[wookie184]

random.choice will error on an empty sequence, so we should handle it being None/an empty array before trying to do that. Right, that makes sense.

What do we actually want to do in the event that .get("results") or selection_id ends up being None?

I don’t think there’s any reason it should be None so probably send a message to the user saying no results were found and maybe log a warning so we can look into it.

In fact, reading through the code and API documentation, I’m not entirely sure why we’re not using data from the first GET request and indexing that instead of the second GET request(which is the one raising the error above). data should contain a list of results, which we can index and then get the id from for what’s currently the third GET request).

The first request is to get the number of pages so one can be selected for the request. If we used the results from the first request we’d be much more limited in the results we got.

Side note: we may also want a fix like https://github.com/python-discord/sir-lancebot/issues/1055, it’s the same API and similar use case, just in a different place. It’s possible it isn’t an issue in practice for this genre if the total count isn’t high enough though.