xform [WIP]
xform is a library to transform spatial data from one space to another and
provides a common interface to combine different types of transforms.
It was originally written for navis to transform neurons from one brain template space to another and then split off into a separate general-purpose package.
Features
- various supported transforms (see below)
- chaining of transforms
- a template registry that tracks available transforms and plots paths to get from a given source to the desired target template space
Supported transforms
- CMTK warp transforms
- Elastix warp transforms
- H5 deformation fields
- Landmark-based thin-plate spline transforms (powered by morphops)
- Landmark-based least-moving square transforms (powered by molesq)
- Affine transformations
Install
$ pip3 install xform
Additional dependencies:
To use CMTK transforms, you need to have CMTK
installed and its binaries (specifically streamxform) in a path where xform
can find them (e.g. /usr/local/bin).
Usage
Single transforms
At the most basic level you can use individual transform from xform.transforms:
AffineTransformfor affine transforms using a affine matrixCMTKtransformfor CMTK transformsElastixTransformfor Elastix transformsTPStransformorMovingLeastSquaresTransformfor landmark-based transformsH5transformfor deformation-field transforms using Hdf5 files (specs)
A quick example that uses an affine transform to scale coordinates by a factor of 2:
>>> import xform
>>> import numpy as np
>>> # Generate the affine matrix
>>> m = np.diag([2, 2, 2, 2])
>>> # Create the transform
>>> tr = xform.AffineTransform(m)
>>> # Some 3D points to transform
>>> points = np.array([[1,1,1], [2,2,2], [3,3,3]])
>>> # Apply
>>> xf = tr.xform(points)
>>> xf
array([[2., 2., 2.],
[4., 4., 4.],
[6., 6., 6.]])
>>> # Transforms are invertible!
>>> (-tr).xform(xf)
array([[1., 1., 1.],
[2., 2., 2.],
[3., 3., 3.]])
Transform sequences
If you find yourself in a situation where you need to chain some transforms,
you can use xform.transforms.TransformSequence to combine transforms.
For example, let's say we have a CMTK transform that requires spatial data to be in microns but our data is in nanometers:
>>> from xform import CMTKtransform, AffineTransform, TransformSequence
>>> import numpy as np
>>> # Initialize CMTK transform
>>> cmtk = CMTKtransform('~/transform/target_source.list')
>>> # Create an affine transform to go from microns to nanometers
>>> aff = AffineTransform(np.diag([1e3, 1e3, 1e3, 1e3]))
>>> # Create a transform sequence
>>> tr = TransformSequence([-aff, cmtk])
>>> # Apply transform
>>> points = np.array([[1,1,1], [2,2,2], [3,3,3]])
>>> xf = tr.xform(points)
Bridging graphs
When working with many interconnected transforms (e.g. A->B, B->C, B->D, etc.),
you can register the individual transforms and let xform plot the shortest
path to get from a given source to a given target for you:
>>> import xform
>>> from xform import CMTKtransform, AffineTransform, TransformRegistry
>>> import numpy as np
>>> # Create a transform registry
>>> registry = TransformRegistry()
>>> # Generate a couple transforms
>>> # Note that we now provide source and target labels
>>> tr1 = AffineTransform(np.diag([1e3, 1e3, 1e3, 1e3]),
... source_space='A', target_space='B')
>>> cmtk = CMTKtransform('~/transform/C_B.list',
... source_space='B', target_space='C')
>>> # Register the transforms
>>> registry.register_transform([tr1, cmtk])
>>> # Now you ask the registry for the required transforms to move between spaces
>>> path, trans_seq = registry.shortest_bridging_seq(source='A', target='C')
>>> path
array(['A', 'B', 'C'], dtype='<U1')
>>> trans_seq
TransformSequence with 2 transform(s)
Custom transforms
TODO
Metadata
Release files for xform 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| xform-0.1.0.tar.gz | 41.1 kB | Without provenance | Details for xform-0.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| xform-0.1.0-py3-none-any.whl | Python 3 | none | any | Details for xform-0.1.0-py3-none-any.whl |
Total release size: 91.4 kB
Release files / xform-0.1.0.tar.gz
| Download URL | xform-0.1.0.tar.gz |
|---|---|
| Size | 41.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
6d2281b6565e5b2172e7c871269bd84c705e6bf9e0db05ed945e17e882ea717d
|
|
BLAKE2b-256 checksum How to use checksums |
24c810a3014e140ea36294728a0d0c10c2642ee9f0800bc6721234d3d770af67
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.2 CPython/3.9.18
|
Release files / xform-0.1.0-py3-none-any.whl
| Download URL | xform-0.1.0-py3-none-any.whl |
|---|---|
| Size | 50.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
91fa8963fdd0173da18e5537a788112fff7a8330e2799c0e98ead79c64450c11
|
|
BLAKE2b-256 checksum How to use checksums |
65aeff4a6f53e3a46856bee410aec5a5a8047dcf9b665915dc79db79334a91a4
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/4.0.2 CPython/3.9.18
|
Security for xform 0.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.