Open Christonikos opened 1 year ago
rly
commented
1 year ago Hi @Christonikos , thanks for the bug report. Could you share the file that produces the error? If it is too big or sensitive to share, I can try to troubleshoot without it, but since the error appears to be due to the usage of an extension data type, it would be considerably easier to troubleshoot with the file.
Christonikos
commented
1 year ago Hi @rly thanks for the follow up. Please allow me to ask the hosting lab and I will get back to you.
bendichter
commented
1 year ago I really don't think M1 is the issue here. There are certain types of installation snafus with the new M1 macs, but this doesn't hit any of them as far as I can see.
I parcellations was added later in the ndx-ecog extension. My guess is there is a misalignment between the schema used to write the data and the schema that is being cached in the file. The data contains parcellations but the cached schema is older and does not.
bendichter
commented
1 year ago One quick fix might be to update ndx-ecog and import it before reading the file (or if you are already doing that, try not importing it)
Christonikos
commented
1 year ago I just tried with both ndx_ecog #0.1.1 and #0.1.0 and adding the importation prior to reading the file, but that led to the same error. My current pynwb==2.2.0.
Christonikos
commented
1 year ago I am attaching a file for troubleshooting here.
bendichter
commented
1 year ago I get the same error on my computer. It looks like the cached ndx-ecog schema is outdated. When I try importing it, I also get a traceback but it is different:
import pynwb
import ndx_ecog
filepath = "/Users/bendichter/Downloads/sub-TS106_places_ieeg_2.nwb"
with pynwb.NWBHDF5IO(filepath, load_namespaces=True) as io:
nwbfile = io.read()
print(nwbfile)---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1248, in ObjectMapper.construct(self, **kwargs)
1247 try:
-> 1248 obj = self.__new_container__(cls, builder.source, parent, builder.attributes.get(self.__spec.id_key()),
1249 **kwargs)
1250 except Exception as ex:
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1261, in ObjectMapper.__new_container__(self, cls, container_source, parent, object_id, **kwargs)
1259 # obj has been created and is in construction mode, indicating that the object is being constructed by
1260 # the automatic construct process during read, rather than by the user
-> 1261 obj.__init__(**kwargs)
1262 obj._in_construct_mode = False # reset to False to indicate that the construction of the object is complete
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/ndx_ecog/ecog.py:24, in Surface.__init__(self, **kwargs)
23 if np.max(faces) >= len(vertices):
---> 24 raise ValueError('index of faces exceeds number vertices for {}. '
25 'Faces should be 0-indexed, not 1-indexed'.
26 format(name))
27 if np.min(faces) < 0:
ValueError: index of faces exceeds number vertices for std_inflated_LH. Faces should be 0-indexed, not 1-indexed
The above exception was the direct cause of the following exception:
ConstructError Traceback (most recent call last)
Cell In [2], line 7
4 filepath = "/Users/bendichter/Downloads/sub-TS106_places_ieeg_2.nwb"
6 with pynwb.NWBHDF5IO(filepath, load_namespaces=True) as io:
----> 7 nwbfile = io.read()
8 print(nwbfile)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/backends/hdf5/h5tools.py:451, in HDF5IO.read(self, **kwargs)
448 raise UnsupportedOperation("Cannot read from file %s in mode '%s'. Please use mode 'r', 'r+', or 'a'."
449 % (self.source, self.__mode))
450 try:
--> 451 return super().read(**kwargs)
452 except UnsupportedOperation as e:
453 if str(e) == 'Cannot build data. There are no values.': # pragma: no cover
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/backends/io.py:42, in HDMFIO.read(self, **kwargs)
39 if all(len(v) == 0 for v in f_builder.values()):
40 # TODO also check that the keys are appropriate. print a better error message
41 raise UnsupportedOperation('Cannot build data. There are no values.')
---> 42 container = self.__manager.construct(f_builder)
43 return container
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/manager.py:280, in BuildManager.construct(self, **kwargs)
276 result = self.__type_map.construct(builder, self, parent)
277 else:
278 # we are at the top of the hierarchy,
279 # so it must be time to resolve parents
--> 280 result = self.__type_map.construct(builder, self, None)
281 self.__resolve_parents(result)
282 self.prebuilt(result, builder)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/manager.py:789, in TypeMap.construct(self, **kwargs)
787 raise ValueError('No ObjectMapper found for builder of type %s' % dt)
788 else:
--> 789 return obj_mapper.construct(builder, build_manager, parent)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1218, in ObjectMapper.construct(self, **kwargs)
1216 cls = manager.get_cls(builder)
1217 # gather all subspecs
-> 1218 subspecs = self.__get_subspec_values(builder, self.spec, manager)
1219 # get the constructor argument that each specification corresponds to
1220 const_args = dict()
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1147, in ObjectMapper.__get_subspec_values(self, builder, spec, manager)
1145 ret[subspec] = self.__flatten(sub_builder, subspec, manager)
1146 # now process groups and datasets
-> 1147 self.__get_sub_builders(groups, spec.groups, manager, ret)
1148 self.__get_sub_builders(datasets, spec.datasets, manager, ret)
1149 elif isinstance(spec, DatasetSpec):
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1198, in ObjectMapper.__get_sub_builders(self, sub_builders, subspecs, manager, ret)
1195 continue
1196 if dt is None:
1197 # recurse
-> 1198 ret.update(self.__get_subspec_values(sub_builder, subspec, manager))
1199 else:
1200 ret[subspec] = manager.construct(sub_builder)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1147, in ObjectMapper.__get_subspec_values(self, builder, spec, manager)
1145 ret[subspec] = self.__flatten(sub_builder, subspec, manager)
1146 # now process groups and datasets
-> 1147 self.__get_sub_builders(groups, spec.groups, manager, ret)
1148 self.__get_sub_builders(datasets, spec.datasets, manager, ret)
1149 elif isinstance(spec, DatasetSpec):
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1200, in ObjectMapper.__get_sub_builders(self, sub_builders, subspecs, manager, ret)
1198 ret.update(self.__get_subspec_values(sub_builder, subspec, manager))
1199 else:
-> 1200 ret[subspec] = manager.construct(sub_builder)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/manager.py:276, in BuildManager.construct(self, **kwargs)
274 if parent_builder is not None:
275 parent = self._get_proxy_builder(parent_builder)
--> 276 result = self.__type_map.construct(builder, self, parent)
277 else:
278 # we are at the top of the hierarchy,
279 # so it must be time to resolve parents
280 result = self.__type_map.construct(builder, self, None)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/manager.py:789, in TypeMap.construct(self, **kwargs)
787 raise ValueError('No ObjectMapper found for builder of type %s' % dt)
788 else:
--> 789 return obj_mapper.construct(builder, build_manager, parent)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1218, in ObjectMapper.construct(self, **kwargs)
1216 cls = manager.get_cls(builder)
1217 # gather all subspecs
-> 1218 subspecs = self.__get_subspec_values(builder, self.spec, manager)
1219 # get the constructor argument that each specification corresponds to
1220 const_args = dict()
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1147, in ObjectMapper.__get_subspec_values(self, builder, spec, manager)
1145 ret[subspec] = self.__flatten(sub_builder, subspec, manager)
1146 # now process groups and datasets
-> 1147 self.__get_sub_builders(groups, spec.groups, manager, ret)
1148 self.__get_sub_builders(datasets, spec.datasets, manager, ret)
1149 elif isinstance(spec, DatasetSpec):
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1200, in ObjectMapper.__get_sub_builders(self, sub_builders, subspecs, manager, ret)
1198 ret.update(self.__get_subspec_values(sub_builder, subspec, manager))
1199 else:
-> 1200 ret[subspec] = manager.construct(sub_builder)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/manager.py:276, in BuildManager.construct(self, **kwargs)
274 if parent_builder is not None:
275 parent = self._get_proxy_builder(parent_builder)
--> 276 result = self.__type_map.construct(builder, self, parent)
277 else:
278 # we are at the top of the hierarchy,
279 # so it must be time to resolve parents
280 result = self.__type_map.construct(builder, self, None)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/manager.py:789, in TypeMap.construct(self, **kwargs)
787 raise ValueError('No ObjectMapper found for builder of type %s' % dt)
788 else:
--> 789 return obj_mapper.construct(builder, build_manager, parent)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1218, in ObjectMapper.construct(self, **kwargs)
1216 cls = manager.get_cls(builder)
1217 # gather all subspecs
-> 1218 subspecs = self.__get_subspec_values(builder, self.spec, manager)
1219 # get the constructor argument that each specification corresponds to
1220 const_args = dict()
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1147, in ObjectMapper.__get_subspec_values(self, builder, spec, manager)
1145 ret[subspec] = self.__flatten(sub_builder, subspec, manager)
1146 # now process groups and datasets
-> 1147 self.__get_sub_builders(groups, spec.groups, manager, ret)
1148 self.__get_sub_builders(datasets, spec.datasets, manager, ret)
1149 elif isinstance(spec, DatasetSpec):
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1190, in ObjectMapper.__get_sub_builders(self, sub_builders, subspecs, manager, ret)
1188 sub_builder = builder_dt.get(dt)
1189 if sub_builder is not None:
-> 1190 sub_builder = self.__flatten(sub_builder, subspec, manager)
1191 ret[subspec] = sub_builder
1192 else:
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1203, in ObjectMapper.__flatten(self, sub_builder, subspec, manager)
1202 def __flatten(self, sub_builder, subspec, manager):
-> 1203 tmp = [manager.construct(b) for b in sub_builder]
1204 if len(tmp) == 1 and not subspec.is_many():
1205 tmp = tmp[0]
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1203, in <listcomp>(.0)
1202 def __flatten(self, sub_builder, subspec, manager):
-> 1203 tmp = [manager.construct(b) for b in sub_builder]
1204 if len(tmp) == 1 and not subspec.is_many():
1205 tmp = tmp[0]
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/manager.py:276, in BuildManager.construct(self, **kwargs)
274 if parent_builder is not None:
275 parent = self._get_proxy_builder(parent_builder)
--> 276 result = self.__type_map.construct(builder, self, parent)
277 else:
278 # we are at the top of the hierarchy,
279 # so it must be time to resolve parents
280 result = self.__type_map.construct(builder, self, None)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/manager.py:789, in TypeMap.construct(self, **kwargs)
787 raise ValueError('No ObjectMapper found for builder of type %s' % dt)
788 else:
--> 789 return obj_mapper.construct(builder, build_manager, parent)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/utils.py:645, in docval.<locals>.dec.<locals>.func_call(*args, **kwargs)
643 def func_call(*args, **kwargs):
644 pargs = _check_args(args, kwargs)
--> 645 return func(args[0], **pargs)
File ~/opt/miniconda3/lib/python3.9/site-packages/hdmf/build/objectmapper.py:1252, in ObjectMapper.construct(self, **kwargs)
1250 except Exception as ex:
1251 msg = 'Could not construct %s object due to: %s' % (cls.__name__, ex)
-> 1252 raise ConstructError(builder, msg) from ex
1253 return obj
ConstructError: (root/general/subject/cortical_surfaces/std_inflated_LH GroupBuilder {'attributes': {'namespace': 'ndx-ecog', 'neurodata_type': 'Surface', 'object_id': 'aac7d8fb-ad9e-4632-9fa3-f897392d846f'}, 'groups': {'parcellations': root/general/subject/cortical_surfaces/std_inflated_LH/parcellations GroupBuilder {'attributes': {'namespace': 'ndx-ecog', 'neurodata_type': 'Parcellations', 'object_id': '5475cfbc-cf80-4eaa-a616-e5b9560e7cd7'}, 'groups': {}, 'datasets': {'HCP': root/general/subject/cortical_surfaces/std_inflated_LH/parcellations/HCP DatasetBuilder {'attributes': {'labels': array(['???', 'V1', 'MST', 'V6', 'V2', 'V3', 'V4', 'V8', '4', '3b', 'FEF',
'PEF', '55b', 'V3A', 'RSC', 'POS2', 'V7', 'IPS1', 'FFC', 'V3B',
'LO1', 'LO2', 'PIT', 'MT', 'A1', 'PSL', 'SFL', 'PCV', 'STV', '7Pm',
'7m', 'POS1', '23d', 'v23ab', 'd23ab', '31pv', '5m', '5mv', '23c',
'5L', '24dd', '24dv', '7AL', 'SCEF', '6ma', '7Am', '7PL', '7PC',
'LIPv', 'VIP', 'MIP', '1', '2', '3a', '6d', '6mp', '6v', 'p24pr',
'33pr', 'a24pr', 'p32pr', 'a24', 'd32', '8BM', 'p32', '10r', '47m',
'8Av', '8Ad', '9m', '8BL', '9p', '10d', '8C', '44', '45', '47l',
'a47r', '6r', 'IFJa', 'IFJp', 'IFSp', 'IFSa', 'p9-46v', '46',
'a9-46v', '9-46d', '9a', '10v', 'a10p', '10pp', '11l', '13l',
'OFC', '47s', 'LIPd', '6a', 'i6-8', 's6-8', '43', 'OP4', 'OP1',
'OP2-3', '52', 'RI', 'PFcm', 'PoI2', 'TA2', 'FOP4', 'MI', 'Pir',
'AVI', 'AAIC', 'FOP1', 'FOP3', 'FOP2', 'PFt', 'AIP', 'EC', 'PreS',
'H', 'ProS', 'PeEc', 'STGa', 'PBelt', 'A5', 'PHA1', 'PHA3',
'STSda', 'STSdp', 'STSvp', 'TGd', 'TE1a', 'TE1p', 'TE2a', 'TF',
'TE2p', 'PHT', 'PH', 'TPOJ1', 'TPOJ2', 'TPOJ3', 'DVT', 'PGp',
'IP2', 'IP1', 'IP0', 'PFop', 'PF', 'PFm', 'PGi', 'PGs', 'V6A',
'VMV1', 'VMV3', 'PHA2', 'V4t', 'FST', 'V3CD', 'LO3', 'VMV2',
'31pd', '31a', 'VVC', '25', 's32', 'pOFC', 'PoI1', 'Ig', 'FOP5',
'p10p', 'p47r', 'TGv', 'MBelt', 'LBelt', 'A4', 'STSva', 'TE1m',
'PI', 'a32pr', 'p24'], dtype=object), 'namespace': 'ndx-ecog', 'neurodata_type': 'Parcellation', 'object_id': '10436e73-6d00-41e4-aa83-0490fcb78a3e'}, 'data': <Closed HDF5 dataset>}}, 'links': {}}}, 'datasets': {'faces': root/general/subject/cortical_surfaces/std_inflated_LH/faces DatasetBuilder {'attributes': {}, 'data': <Closed HDF5 dataset>}, 'vertices': root/general/subject/cortical_surfaces/std_inflated_LH/vertices DatasetBuilder {'attributes': {}, 'data': <Closed HDF5 dataset>}}, 'links': {}}, 'Could not construct Surface object due to: index of faces exceeds number vertices for std_inflated_LH. Faces should be 0-indexed, not 1-indexed')The file uses parcellations, which are defined in the extension.
I see a strange dataset in there though. What is "curv" doing here?
bendichter
commented
1 year ago As the traceback suggests, it looks like the "faces" array is 1-indexed instead of 0-indexed, since it goes from 1 to n_vertices+1. The solution is to
Christonikos
commented
1 year ago Thanks, @bendichter—could you please explain a bit more the first step? I tried using the latest ndx-ecog version, but I still could not load the file. A code snippet with a requirements.txt would be very helpful.
bendichter
commented
1 year ago @Christonikos apologies for the sporadic support during the holiday break.
I was able to get past your error by pip-installing ndx-ecog and importing it. I'm not quite sure yet why your original attempt isn't working- it should. It may be revealing a weakness in our automatic class generation, but I am able to get past it for now with this import (import ndx_ecog) before the NWBHDF5IO. This produces another very long traceback but you'll notice at the end that the error is different. This traceback reveals that there is a problem with the data that is preventing this file from being opened.
We put a checker in to make sure that the faces array has the correct values. This array defines triangles between vertices by listing triplets of vertices for each triangle. The complex 3-D shapes are constructed from these triangles. These indices should be 0-indexed, meaning they should go from 0 to n_vertices - 1. This is a bit against the grain for MATLAB users but it is more common in the broader programming community and is the norm e.g. in Python. However the faces array here is clearly 1-indexed, since it has a range 1 to n_verticies.
I put a checker in the loader function to check for this condition and throw an error if it occurs. (Perhaps this is a bit heavy-handed. Maybe a warning would have been better than an error.) But the data is incorrect. All that would need to be done to correct these arrays is to subtract all of the faces values by 1. I could create a patch script but I think you are better off going to the data producers and requesting that they fix it on their side.
bendichter
commented
1 year ago I also don't know what's going on with the "curv" datasets
Christonikos
commented
1 year ago Hi @bendichter thanks for the follow-up and happy new year. I can confirm that I get the same traceback when importing ndx_ecog prior to NWBHDF5IO. What puzzles me is that I asked the hosting lab to load the data on their side, and they faced no such issue, nor did I when using an Ubuntu PC. I have also informed the hosting lab, so they can jump in the discussion and follow-up on their side.
bendichter
commented
1 year ago It's probably a package version issue. If they are using a version of ndx-ecog before that check was in place then they won't get that error. The data is still incorrect though even if they don't get the error, so I think the right way forward is to correct it.
Christonikos
commented
1 year ago @bendichter I am adding @OWoolnough in the conversation, as a representative of the hosting lab.
Christonikos
commented
1 year ago Hi everyone, I'm still unable to use these files on my M1, whereas in the past I was able to load and work with it (despite the faces index problem) and @owoolnough is able to load the dataset on his side. Any advice on how to proceed?
bendichter
commented
1 year ago @Christonikos and @owoolnough the data in the NWB file supplied is incorrect.
import h5py
file = h5py.File("/Users/bendichter/Downloads/sub-TS106_places_ieeg_2.nwb")
print(np.min(file["/general/subject/cortical_surfaces/std_pial_LH/faces"]))
print(np.max(file["/general/subject/cortical_surfaces/std_pial_LH/faces"]))
print(file["/general/subject/cortical_surfaces/std_pial_LH/vertices"].shape)1
156252
(156252, 3)faces should take values from 0 to shape[0] - 1. That needs to be fixed before this is readable. It is possible that you were able to read this file using an outdated ndx-ecog package on another computer, but the data is still wrong and I don't think it has to do with M1.
What happened?
When trying to load an nwb file that contains ECOG data, I receive a
ConstructErrororiginating from aTypeError: CustomClassGenerator.set_init.<locals>.__init__: missing argument 'parcellations.Notably, the file could be previously loaded on a dell, ubuntu-based laptop, and the problem occurs when trying to load the file on an apple m1 macbook air.
The code originally used to load the file was:
After interacting a bit with @bendichter I tried his suggestion (that works on his m1):
Which led to the same exact error on my side.
I am bit puzzled by this, so I opened it as an issue in the hopes that this can also be useful for the members of the community.
I am using pynwb #2.2.0 (but also tried with many other versions) and h5py #2.10.0
Steps to Reproduce
Traceback
Operating System
macOS
Python Executable
Conda
Python Version
3.9
Package Versions
pynwb #2.2.0 h5py #2.10.0
Code of Conduct