Add support for MRtrix MIF file format#1489
Draft
tsalo wants to merge 4 commits into
Draft
Conversation
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #1489 +/- ##
==========================================
- Coverage 95.44% 95.38% -0.07%
==========================================
Files 209 211 +2
Lines 29981 30616 +635
Branches 4483 4567 +84
==========================================
+ Hits 28616 29202 +586
- Misses 930 960 +30
- Partials 435 454 +19 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Member
Author
|
@mattcieslak any chance you have some very small MIF files (< 50K) that we can use for testing here? |
effigies
reviewed
Apr 13, 2026
effigies
left a comment
effigies
left a comment
Member
There was a problem hiding this comment.
I didn't get very far in reviewing this, but I had written a comment. I can try to look at this again on Wednesday, but hitting post in case this is helpful to you now.
Comment on lines
+44
to
+109
| _MIF_DTYPE_MAP: dict[str, str] = { | ||
| 'Int8': 'i1', | ||
| 'UInt8': 'u1', | ||
| 'Int16': 'i2', | ||
| 'UInt16': 'u2', | ||
| 'Int32': 'i4', | ||
| 'UInt32': 'u4', | ||
| 'Int64': 'i8', | ||
| 'UInt64': 'u8', | ||
| 'Float32': 'f4', | ||
| 'Float64': 'f8', | ||
| 'CFloat32': 'c8', | ||
| 'CFloat64': 'c16', | ||
| } | ||
|
|
||
| _NUMPY_TO_MIF_BASE: dict[tuple[str, int], str] = { | ||
| ('i', 1): 'Int8', | ||
| ('u', 1): 'UInt8', | ||
| ('i', 2): 'Int16', | ||
| ('u', 2): 'UInt16', | ||
| ('i', 4): 'Int32', | ||
| ('u', 4): 'UInt32', | ||
| ('i', 8): 'Int64', | ||
| ('u', 8): 'UInt64', | ||
| ('f', 4): 'Float32', | ||
| ('f', 8): 'Float64', | ||
| ('c', 8): 'CFloat32', | ||
| ('c', 16): 'CFloat64', | ||
| } | ||
|
|
||
|
|
||
| def _mif_parse_dtype(dtype_str: str) -> np.dtype: | ||
| """Convert a MIF datatype string (e.g. ``'Float32LE'``) to a numpy dtype.""" | ||
| dtype_str = dtype_str.strip() | ||
| if dtype_str.endswith('LE'): | ||
| endian, base = '<', dtype_str[:-2] | ||
| elif dtype_str.endswith('BE'): | ||
| endian, base = '>', dtype_str[:-2] | ||
| else: | ||
| endian = '<' if sys.byteorder == 'little' else '>' | ||
| base = dtype_str | ||
|
|
||
| if base not in _MIF_DTYPE_MAP: | ||
| raise ValueError(f'Unknown MIF datatype: {dtype_str!r}') | ||
|
|
||
| type_char = _MIF_DTYPE_MAP[base] | ||
| if type_char in ('i1', 'u1'): # single-byte types have no endianness | ||
| return np.dtype(type_char) | ||
| return np.dtype(endian + type_char) | ||
|
|
||
|
|
||
| def _mif_dtype_to_str(dtype: np.dtype) -> str: | ||
| """Convert a numpy dtype to a MIF datatype string.""" | ||
| dtype = np.dtype(dtype) | ||
| base_name = _NUMPY_TO_MIF_BASE.get((dtype.kind, dtype.itemsize)) | ||
| if base_name is None: | ||
| raise ValueError(f'Cannot represent numpy dtype {dtype!r} in MIF format') | ||
| if dtype.itemsize == 1: | ||
| return base_name | ||
|
|
||
| byte_order = dtype.byteorder | ||
| if byte_order == '=': | ||
| byte_order = '<' if sys.byteorder == 'little' else '>' | ||
| elif byte_order == '|': | ||
| return base_name | ||
| return base_name + ('LE' if byte_order == '<' else 'BE') |
Member
There was a problem hiding this comment.
A lot of this work is already done in nibabel.volumeutils. Building off your first structure:
_dtdefs = [
(f'{mift}{end}', (label := f'{end and recode[end]}{dt}'), np.dtype(label).type)
for mift, dt in _MIF_DTYPE_MAP.items()
for end in ('', 'BE', 'LE')
]
data_type_codes = nb.volumeutils.make_dt_codes(_dtdefs)This allows you to replace your functions:
_mif_dtype_to_str = lambda dtype: data_type_codes[dtype]
_mif_parse_dtype = lambda dtype_str: data_type_codes.dtype[dtype_str]See, for example:
>>> data_type_codes.dtype['Float32']
dtype('float32')
>>> data_type_codes.dtype['Float32BE']
dtype('>f4')
>>> data_type_codes.dtype['Float32LE']
dtype('float32')
>>> data_type_codes[np.dtype('f8')]
'Float64LE'This doesn't do the same validation. Bad names will show up as KeyErrors, instead of custom-raised ValueErrors, and you could theoretically have a file with a bad name that matches a numpy dtype string, and that would be accepted.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Related to #927 (does not close because this doesn't include MIH support). I used Claude to write this for
ModelArrayIO, but I don't have much experience with the file format so I'm happy to make any changes.I don't have any small MIF files that would be appropriate for nibabel's tests on hand, but I can hunt some down.