Basic functionality (#14)

* Added default parameter sets to be used for transforms (from brainregister [IBL] or ARA)

* Basic magicgui that can import an atlas, translate the atlas/sample, and perform selected transforms with either elastix defaults or provided ARA defaults

* Added a button to reset the image back to origin

* Added the ability to rotate the image and reset back to original state

* Added start alignment button

* Added start alignment button

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Transitioning basic function to qpty directly. Widget in registration_widget.py now uses qtpy directly to replicate the majority of _widget.py functionality.

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Napari image layer is now rotated around the center not at origin

* Tests for setup_parameter_object done

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Wrote basic tests for registration_widget and register

* Re-factored the core registration_widget.py to split off each individual UI component into its own file

* Added a run settings widget to allow selection of transform types and default param settings

* Added tabs on the left side of the main widget, parameter_list_view represents parameters in a table widget

* Can dynamically load default values from the 3 provided sources and load them into the parameter list view

* Param dictionaries are updated when fields are altered in the table tab view

* Parameter table now expands when new entries are added

* Plugin runs successfully after renaming

* Param dictionaries passed to the backend directly

* First version of transform selection widget allowing arbitrary number of transforms with unique defaults

* Each column now dynamically imports the correct default file based on the selection in the second column of the widget

* Parameters are now read as lists from files, arbitrary numbers of parameter maps can be run in one go

* Moving image adjusts immediately after values are entered, removed adjust image button

* When reading parameter files anything after a '\' will now be discarded to allow comments in parameter file

* Updated tests for braingblobe_registration/register.py, added tests for brainglobe_registration/widgets/adjust_moving_image.py

* Added tests for brainglobe_registration/widgets/select_images_view.py

* Started tests for brainglobe_registration/widgets/parameter_list_view.py

* Finished tests for brainglobe/registration/widgets/parameter_list_view.py

* Started tests for transform_select_view.py

* Updating README.md

* Update README.md

* Updated README.md

* Added tests, added dynamic fetching for sample selection

* Update README.md

* Refactored test_registration_widget.py to account for funcitons moving to utils

* Fixed line lengths

* WIP transforming annotations after registration

* WIP transform annotations after registration

* Transformed annotation and boundaries images from atlas to sample space

* Registered boundaries are now added as an image layer, boundaries overlaid on sample image after run

* Changed the source link in the widget header

Co-authored-by: Adam Tyson <code@adamltyson.com>

* Removed tutorial section of the header

---------

Co-authored-by: IgorTatarnikov <i.tatarnikov@ucl.ac.uk>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Adam Tyson <code@adamltyson.com>

.gitignore

@@ -82,3 +82,6 @@ venv/
 
 # written by setuptools_scm
 **/_version.py
+
+# Other test files
+Logs/

README.md

@@ -11,29 +11,51 @@ Registration to a BrainGlobe atlas using Elastix
 
 ----------------------------------
 
-This [napari] plugin was generated with [Cookiecutter] using [@napari]'s [cookiecutter-napari-plugin] template.
-
-<!--
-Don't miss the full getting started guide to set up your new package:
-https://github.com/napari/cookiecutter-napari-plugin#getting-started
-
-and review the napari docs for plugin developers:
-https://napari.org/stable/plugins/index.html
--->
-
+A [napari] plugin for registering images to a BrainGlobe atlas.
+
+![brainglobe-registration](./imgs/brainglobe_registration_main.png)
+
+## Usage
+
+1. Open `napari`.
+2. Install the plugin with `pip install git+https://github.com/brainglobe/brainglobe-registration.git`.
+3. Open the widget by selecting `Plugins > BrainGlobe Registration` in the napari menu bar near the
+top left of the window.
+![brainglobe-registration-plugin](./imgs/brainglobe_registration_plugin_window.png)
+The `BrainGlobe Registration` plugin will appear on the right hand side of the napari window.
+4. Open the image you want to register in napari (a sample 2D image can be found by selecting `File > Open Sample > Sample Brain Slice`).
+5. Select the atlas you want to register to from the dropdown menu.
+![brainglobe-registration-atlas-selection](./imgs/brainglobe_registration_atlas_selection.png)
+The atlas will appear in the napari viewer. Select the approximate `Z` slice of the atlas that you want to register to,
+using the slider at the bottom of the napari viewer.
+![brainglobe-registration-atlas-selection](./imgs/brainglobe_registration_atlas_selection_2.png)
+6. Adjust the sample image to roughly match the atlas image.
+You can do this by adjusting X and Y translation as well as rotating around the centre of the image.
+You can overlay the two images by toggling `Grid` mode in the napari viewer (Ctrl+G).
+You can then adjust the color map and opacity of the atlas image to make manual alignment easier.
+![brainglobe-registration-overlay](./imgs/brainglobe_registration_overlay.png)
+The sample image can be reset to its original position and orientation by clicking `Reset Image` in the `BrainGlobe Registration` plugin window.
+7. Select the transformations you want to use from the dropdown menu. Set the transformation type to empty to remove a step.
+Select from one of the three provided default parameter sets (elastix, ARA, or IBL). Customise the parameters further in the
+`Parameters` tab.
+8. Click `Run` to register the image. The registered image will appear in the napari viewer.
+![brainglobe-registration-registered](./imgs/brainglobe_registration_registered.png)
+![brainglobe-registration-registered](./imgs/brainglobe_registration_registered_stacked.png)
 ## Installation
 
-You can install `brainglobe-registration` via [pip]:
+We strongly recommend to use a virtual environment manager (like `conda` or `venv`). The installation instructions below
+will not specify the Qt backend for napari, and you will therefore need to install that separately. Please see the
+[`napari` installation instructions](https://napari.org/stable/tutorials/fundamentals/installation.html) for further advice on this.
 
-    pip install brainglobe-registration
+[WIP] You can install `brainglobe-registration` via [pip]:
 
+    pip install brainglobe-registration
 
 
 To install latest development version :
 
     pip install git+https://github.com/brainglobe/brainglobe-registration.git
 
-
 ## Contributing
 
 Contributions are very welcome. Tests can be run with [tox], please ensure
@@ -48,6 +70,19 @@ Distributed under the terms of the [BSD-3] license,
 
 If you encounter any problems, please [file an issue] along with a detailed description.
 
+
+## Acknowledgements
+
+This [napari] plugin was generated with [Cookiecutter] using [@napari]'s [cookiecutter-napari-plugin] template.
+
+<!--
+Don't miss the full getting started guide to set up your new package:
+https://github.com/napari/cookiecutter-napari-plugin#getting-started
+
+and review the napari docs for plugin developers:
+https://napari.org/stable/plugins/index.html
+-->
+
 [napari]: https://github.com/napari/napari
 [Cookiecutter]: https://github.com/audreyr/cookiecutter
 [@napari]: https://github.com/napari

pyproject.toml

@@ -4,11 +4,17 @@ build-backend = "setuptools.build_meta"
 
 
 [tool.setuptools_scm]
-write_to = "src/bg_elastix/_version.py"
+write_to = "src/brainglobe_registration/_version.py"
 
 [tool.black]
 line-length = 79
 
 [tool.isort]
 profile = "black"
 line_length = 79
+
+[tool.pytest.ini_options]
+testpaths = "src/tests"
+markers = [
+    "slow: mark test as slow"
+]

setup.cfg

@@ -36,6 +36,7 @@ install_requires =
     qtpy
     itk-elastix
     bg-atlasapi
+    pytransform3d
 
 python_requires = >=3.8
 include_package_data = True

src/brainglobe_registration/elastix/register.py

@@ -1,66 +1,120 @@
 import itk
 import numpy as np
+from bg_atlasapi import BrainGlobeAtlas
+from typing import List
+
+
+def get_atlas_by_name(atlas_name: str) -> BrainGlobeAtlas:
+    """
+    Get a BrainGlobeAtlas object by its name.
+
+    Parameters
+    ----------
+    atlas_name : str
+        The name of the atlas.
+
+    Returns
+    -------
+    BrainGlobeAtlas
+        The BrainGlobeAtlas object.
+    """
+    atlas = BrainGlobeAtlas(atlas_name)
+
+    return atlas
 
 
 def run_registration(
-    fixed_image,
+    atlas_image,
     moving_image,
-    rigid=True,
-    affine=True,
-    bspline=True,
-    affine_iterations="2048",
-    log=False,
-):
+    annotation_image,
+    parameter_lists: List[tuple[str, dict]] = None,
+) -> tuple[np.ndarray, itk.ParameterObject, np.ndarray]:
+    """
+    Run the registration process on the given images.
+
+    Parameters
+    ----------
+    atlas_image : np.ndarray
+        The atlas image.
+    moving_image : np.ndarray
+        The moving image.
+    annotation_image : np.ndarray
+        The annotation image.
+    parameter_lists : List[tuple[str, dict]], optional
+        The list of parameter lists, by default None
+
+    Returns
+    -------
+    np.ndarray
+        The result image.
+    itk.ParameterObject
+        The result transform parameters.
+    """
     # convert to ITK, view only
-    fixed_image = itk.GetImageViewFromArray(fixed_image).astype(itk.F)
+    atlas_image = itk.GetImageViewFromArray(atlas_image).astype(itk.F)
     moving_image = itk.GetImageViewFromArray(moving_image).astype(itk.F)
 
     # This syntax needed for 3D images
     elastix_object = itk.ElastixRegistrationMethod.New(
-        fixed_image, moving_image
+        moving_image, atlas_image
     )
 
-    parameter_object = setup_parameter_object(
-        rigid=rigid,
-        affine=affine,
-        bspline=bspline,
-        affine_iterations=affine_iterations,
-    )
+    parameter_object = setup_parameter_object(parameter_lists=parameter_lists)
+
     elastix_object.SetParameterObject(parameter_object)
-    elastix_object.SetLogToConsole(log)
 
     # update filter object
     elastix_object.UpdateLargestPossibleRegion()
 
     # get results
     result_image = elastix_object.GetOutput()
     result_transform_parameters = elastix_object.GetTransformParameterObject()
-    return np.asarray(result_image), result_transform_parameters
+    temp_interp_order = result_transform_parameters.GetParameter(
+        0, "FinalBSplineInterpolationOrder"
+    )
+    result_transform_parameters.SetParameter(
+        "FinalBSplineInterpolationOrder", "0"
+    )
 
+    annotation_image_transformix = itk.transformix_filter(
+        annotation_image.astype(np.float32, copy=False),
+        result_transform_parameters,
+    )
+
+    result_transform_parameters.SetParameter(
+        "FinalBSplineInterpolationOrder", temp_interp_order
+    )
+
+    return (
+        np.asarray(result_image),
+        result_transform_parameters,
+        np.asarray(annotation_image_transformix),
+    )
 
-def setup_parameter_object(
-    rigid=True,
-    affine=True,
-    bspline=True,
-    affine_iterations="2048",
-):
+
+def setup_parameter_object(parameter_lists: List[tuple[str, dict]] = None):
+    """
+    Set up the parameter object for the registration process.
+
+    Parameters
+    ----------
+    parameter_lists : List[tuple[str, dict]], optional
+        The list of parameter lists, by default None
+
+    Returns
+    -------
+    itk.ParameterObject
+        The parameter object.#
+    """
     parameter_object = itk.ParameterObject.New()
 
-    if rigid:
-        parameter_map_rigid = parameter_object.GetDefaultParameterMap("rigid")
-        parameter_object.AddParameterMap(parameter_map_rigid)
-
-    if affine:
-        parameter_map_affine = parameter_object.GetDefaultParameterMap(
-            "affine"
-        )
-        parameter_map_affine["MaximumNumberOfIterations"] = [affine_iterations]
-        parameter_object.AddParameterMap(parameter_map_affine)
-
-    if bspline:
-        parameter_map_bspline = parameter_object.GetDefaultParameterMap(
-            "bspline"
-        )
-        parameter_object.AddParameterMap(parameter_map_bspline)
+    for transform_type, parameter_dict in parameter_lists:
+        parameter_map = parameter_object.GetDefaultParameterMap(transform_type)
+        parameter_map.clear()
+
+        for k, v in parameter_dict.items():
+            parameter_map[k] = v
+
+        parameter_object.AddParameterMap(parameter_map)
 
     return parameter_object

src/brainglobe_registration/napari.yaml

@@ -1,10 +1,14 @@
 name: brainglobe-registration
-display_name: BrainGlobe Elastix Registration
+display_name: BrainGlobe Registration
 contributions:
   commands:
-    - id: brainglobe-registration.register
-      python_name: brainglobe_registration._widget:register
-      title: BrainGlobe Elastix Registration
+    - id: brainglobe-registration.make_registration_widget
+      python_name: brainglobe_registration.registration_widget:RegistrationWidget
+      title: BrainGlobe Registration
+  sample_data:
+    - key: example
+      display_name: Sample Brain Slice
+      uri: src/brainglobe_registration/resources/sample_hipp.tif
   widgets:
-    - command: brainglobe-registration.register
-      display_name: BrainGlobe Elastix Registration
+    - command: brainglobe-registration.make_registration_widget
+      display_name: BrainGlobe Registration