trk-io  

trk-io implements a TrackVis (.trk) reader and writer.

Highlights

• Can read and write TrackVis files. Handles affine transformation as nibabel.streamlines and MI-Brain would.
• Reading and writing is tested as much as in nibabel.streamlines.
• Reader can read all streamlines at once or can be used as a generator. Handles endianness.
• Write all at once or streamline per streamline.
• Follows nibabel.streamlines architecture (all 3D points are in a single Vec![Point3D]). Currently, this is only useful for performance, but it may lead to easier changes when and if we support BLAS.
• Handles endianness.

Examples

rust // Read complete streamlines to memory let streamlines = Reader::new("bundle.trk").read_all(); for streamline in &streamlines { println!("Nb points: {}", streamline.len()); for point in streamline { println!("{}", point); } } rust // Simple read/write. Using a generator (read one streamline at a time) let reader = Reader::new("full_brain.trk"); let mut writer = Writer::new("copy.trk", Some(reader.header.clone())); for streamline in reader.into_iter() { writer.write(streamline); } // The new file will be completed only at the end of the scope. The // 'n_count' field is written in the destructor because we don't // know how many streamlines the user will write.

Roadmap

There's still a lot of work to do but it should work perfectly for simple use cases. In particular, future versions should be able to:

• Support TCK reading/writing
• Better handling of scalars and properties. There's currently a way to access them but I wouldn't call it conveniant. They are in the header so you need to zip over them yourself.
• Create some binary tools using this lib, e.g. showaffine, counttracks, pruning, strip_info, color, etc.
• Support for ops.Range, e.g. streamlines[0..10]

Your help is much appreciated. Consider filing an issue in case something is missing for your use case to work. Pull requests are also welcome.