Chris Markiewicz

@yohanchatelain Just checking in today since I'll be out tomorrow.

Thanks @yohanchatelain. That was very helpful, and we declined to try to include this in an LTS release.

@rwblair Just a note that bids-examples is failing dev tests for the PET datasets because we updated the schema without updating the validator.

I don't think we should add these as-is, and would strongly suggest avoiding adding anything but numpy directly. However, we could use protocols to try to identify these and similar...

I mean something like: ```python class ArrayLike(ty.Protocol): def __array__(self, *args, **kwargs): ... @register_serializer def bytes_repr_arraylike(obj: ArrayLike, cache: Cache) -> Iterator[bytes]: yield f"{obj.__class__.__module__}{obj.__class__.__name__}:{obj.size}:".encode() array = np.asanyarray(obj) if array.dtype == "object": yield...

This works for me in a quick IPython session: ```python class ArrayLike(ty.Protocol): def __array__(self, *args, **kwargs): ... @register_serializer def bytes_repr_arraylike(obj: ArrayLike, cache: Cache) -> Iterator[bytes]: yield f"{obj.__class__.__module__}.{obj.__class__.__name__}:".encode() array = np.asanyarray(obj)...

That's really weird. Can you activate `pdb` and see what object is actually causing the problem?

I've never used colab and don't have time to figure it out. I don't understand why you're getting weak references in your environment, hence why I suggest using pdb. It...

This is the BIDS App protocol, which says analysis level is required and can be run, session, participant or group. We only implement per-subject processing. The fmriprep-docker wrapper is a...

> The `--index` option could work, but would involve reading the 4D input file N times. Will that be better than doing the split? I suppose it's a question of...