Evoked IO cannot handle channel names with colons

[jasmainak]

Description of the problem

I have data where the channel names contain :. While MNE is able to successfully write this data, it throws a weird error when reading. While I can fix the channel names, I think the error (if at all) should be raised when writing the data, not reading.

Steps to reproduce

import mne

root = mne.datasets.sample.data_path() / 'MEG' / 'sample'
evk_file = root / 'sample_audvis-ave.fif'
evoked = mne.read_evokeds(evk_file, baseline=(None, 0), proj=True,
                          verbose=False)[0]
# comment out the following line to see now error
evoked.rename_channels(lambda x: x.replace(' ', ':'))

evoked.info['bads'] = [evoked.info['ch_names'][0]]
mne.write_evokeds('test_ave.fif', evoked, overwrite=True)

evoked2 = mne.read_evokeds('test_ave.fif')[0]

Link to data

No response

Expected results

No error

Actual results

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
File /autofs/space/meghnn_001/users/mjas/github_repos/mne-opm/mne_bug.py:14
     11 evoked.info['bads'] = [evoked.info['ch_names'][0]]
     12 mne.write_evokeds('test_ave.fif', evoked, overwrite=True)
---> 14 evoked2 = mne.read_evokeds('test_ave.fif')[0]

File <decorator-gen-255>:12, in read_evokeds(fname, condition, baseline, kind, proj, allow_maxshield, verbose)

File /autofs/space/meghnn_001/users/mjas/github_repos/mne-python/mne/evoked.py:1186, in read_evokeds(fname, condition, baseline, kind, proj, allow_maxshield, verbose)
   1184 return_list = True
   1185 if condition is None:
-> 1186     evoked_node = _get_evoked_node(fname)
   1187     condition = range(len(evoked_node))
   1188 elif not isinstance(condition, list):

File /autofs/space/meghnn_001/users/mjas/github_repos/mne-python/mne/evoked.py:1005, in _get_evoked_node(fname)
   1003 f, tree, _ = fiff_open(fname)
   1004 with f as fid:
-> 1005     _, meas = read_meas_info(fid, tree, verbose=False)
   1006     evoked_node = dir_tree_find(meas, FIFF.FIFFB_EVOKED)
   1007 return evoked_node

File <decorator-gen-35>:10, in read_meas_info(fid, tree, clean_bads, verbose)

File /autofs/space/meghnn_001/users/mjas/github_repos/mne-python/mne/io/meas_info.py:1552, in read_meas_info(fid, tree, clean_bads, verbose)
   1549             acq_stim = tag.data
   1551 #   Load the SSP data
-> 1552 projs = _read_proj(
   1553     fid, meas_info, ch_names_mapping=ch_names_mapping)
   1555 #   Load the CTF compensation data
   1556 comps = _read_ctf_comp(
   1557     fid, meas_info, chs, ch_names_mapping=ch_names_mapping)

File <decorator-gen-15>:12, in _read_proj(fid, node, ch_names_mapping, verbose)

File /autofs/space/meghnn_001/users/mjas/github_repos/mne-python/mne/io/proj.py:528, in _read_proj(fid, node, ch_names_mapping, verbose)
    525     data = data.T
    527 if data.shape[1] != len(names):
--> 528     raise ValueError('Number of channel names does not match the '
    529                      'size of data matrix')
    531 # just always use this, we used to have bugs with writing the
    532 # number correctly...
    533 nchan = len(names)

ValueError: Number of channel names does not match the size of data matrix

Additional information

Platform:         Linux-3.10.0-1160.76.1.el7.x86_64-x86_64-with-glibc2.10
Python:           3.8.6 | packaged by conda-forge | (default, Oct  7 2020, 19:08:05)  [GCC 7.5.0]
Executable:       /autofs/space/meghnn_001/users/mjas/anaconda3/envs/mne/bin/python3.8
CPU:              x86_64: 64 cores
Memory:           125.4 GB

Exception ignored on calling ctypes callback function: <function _ThreadpoolInfo._find_modules_with_dl_iterate_phdr.<locals>.match_module_callback at 0x7f921e1c2790>
Traceback (most recent call last):
  File "/autofs/space/meghnn_001/users/mjas/anaconda3/envs/mne/lib/python3.8/site-packages/threadpoolctl.py", line 400, in match_module_callback
    self._make_module_from_path(filepath)
  File "/autofs/space/meghnn_001/users/mjas/anaconda3/envs/mne/lib/python3.8/site-packages/threadpoolctl.py", line 515, in _make_module_from_path
    module = module_class(filepath, prefix, user_api, internal_api)
  File "/autofs/space/meghnn_001/users/mjas/anaconda3/envs/mne/lib/python3.8/site-packages/threadpoolctl.py", line 606, in __init__
    self.version = self.get_version()
  File "/autofs/space/meghnn_001/users/mjas/anaconda3/envs/mne/lib/python3.8/site-packages/threadpoolctl.py", line 646, in get_version
    config = get_config().split()
AttributeError: 'NoneType' object has no attribute 'split'
mne:              1.4.dev0
numpy:            1.23.0 {unknown linalg bindings}
scipy:            1.5.3
matplotlib:       3.3.3 {backend=Qt5Agg}

sklearn:          0.23.2
numba:            Not found
nibabel:          3.2.1
nilearn:          0.7.0
dipy:             1.3.0
openmeeg:         Not found
cupy:             Not found
pandas:           1.1.5
pyvista:          0.36.1 {OpenGL 4.5.0 NVIDIA 455.45.01 via Quadro P5000/PCIe/SSE2}
pyvistaqt:        0.9.0
ipyvtklink:       Not found
vtk:              9.0.1
qtpy:             2.0.1 {PyQt5=5.12.9}
ipympl:           Not found
/autofs/space/meghnn_001/users/mjas/anaconda3/envs/mne/lib/python3.8/site-packages/pyqtgraph/colors/palette.py:1: RuntimeWarning: PyQtGraph supports Qt version >= 5.15, but 5.12.9 detected.
  from ..Qt import QtGui
pyqtgraph:        0.13.1
pooch:            v1.5.2

mne_bids:         Not found
mne_nirs:         Not found
mne_features:     Not found
mne_qt_browser:   0.4.0
mne_connectivity: Not found
mne_icalabel:     Not found