napari image arithmetic widget¶
napari is a fast, interactive, multi-dimensional image viewer for python. It uses Qt for the GUI, so it’s easy to extend napari with small, composable widgets created with
magicgui. Here we’re going to build this simple image arithmetic widget with a few additional lines of code.
For napari-specific magicgui documentation, see the napari docs
This example demonstrates how to:
⬇️ Create a
magicgui widget that can be used in another program (napari)
⬇️ Use an
Enum to create a dropdown menu
⬇️ Connect some event listeners to create interactivity.
Code follows, with explanation below… You can also get this example at github.
1from enum import Enum 2 3import numpy 4import napari 5from napari.types import ImageData 6 7from magicgui import magicgui 8 9class Operation(Enum): 10 """A set of valid arithmetic operations for image_arithmetic. 11 12 To create nice dropdown menus with magicgui, it's best 13 (but not required) to use Enums. Here we make an Enum 14 class for all of the image math operations we want to 15 allow. 16 """ 17 add = numpy.add 18 subtract = numpy.subtract 19 multiply = numpy.multiply 20 divide = numpy.divide 21 22 23# here's the magicgui! We also use the additional 24# `call_button` option 25@magicgui(call_button="execute") 26def image_arithmetic( 27 layerA: ImageData, operation: Operation, layerB: ImageData 28) -> ImageData: 29 """Add, subtracts, multiplies, or divides to image layers.""" 30 return operation.value(layerA, layerB) 31 32# create a viewer and add a couple image layers 33viewer = napari.Viewer() 34viewer.add_image(numpy.random.rand(20, 20), name="Layer 1") 35viewer.add_image(numpy.random.rand(20, 20), name="Layer 2") 36 37# add our new magicgui widget to the viewer 38viewer.window.add_dock_widget(image_arithmetic) 39 40# keep the dropdown menus in the gui in sync with the layer model 41viewer.layers.events.inserted.connect(image_arithmetic.reset_choices) 42viewer.layers.events.removed.connect(image_arithmetic.reset_choices) 43 44napari.run()
We’re going to go a little out of order so that the other code makes more sense. Let’s start with the actual function we’d like to write to do some image arithmetic.
Our function takes two
numpy arrays (in this case, from Image layers), and some mathematical operation (we’ll restrict the options using an
Enum). When called, our function calls the selected operation on the data.
def image_arithmetic(array1, operation, array2): return operation.value(array1, array2)
magicgui works particularly well with type annotations, and allows third-party libraries to register widgets and behavior for handling their custom types (using
napari provides support for
magicgui by registering a dropdown menu whenever a function parameter is annotated as one of the basic napari
Layer types, or, in this case,
ImageData indicates we just the
data attribute of the layer. Furthermore, it recognizes when a function has a
LayerData return type annotation, and will add the result to the viewer. So we gain a lot by annotating the above function with the appropriate
from napari.types import ImageData def image_arithmetic( layerA: ImageData, operation: Operation, layerB: ImageData ) -> ImageData: return operation.value(layerA, layerB)
the magic part¶
Finally, we decorate the function with
@magicgui and tell it we’d like to have a
call_button that we can click to execute the function.
@magicgui(call_button="execute") def image_arithmetic(layerA: ImageData, operation: Operation, layerB: ImageData): return operation.value(layerA, layerB)
That’s it! The
image_arithmetic function is now a
FunctionGui that can be shown, or incorporated into other GUIs (such as the napari GUI shown in this example)
create dropdowns with Enums¶
We’d like the user to be able to select the operation (
divide) using a dropdown menu.
Enums offer a convenient way to restrict values to a strict set of options, while providing
name: value pairs for each of the options. Here, the value for each choice is the actual function we would like to have called when that option is selected.
class Operation(enum.Enum): add = numpy.add subtract = numpy.subtract multiply = numpy.multiply divide = numpy.divide
add it to napari¶
When we decorated the
image_arithmetic function above, it became a
FunctionGui. Napari recognizes this type, so we can simply add it to the napari viewer as follows:
This api has changed slightly with version 0.2.0 of magicgui. See the migration guide if you are migrating from a previous version.
connect event listeners for interactivity¶
What fun is a GUI without some interactivity? Let’s make stuff happen.
We connect the
image_arithmetic.reset_choices function to the
viewer.layers.events.inserted/removed event from
napari, to make sure that the dropdown menus stay in sync if a layer gets added or removed from the napari window:
An additional offering from
magicgui here is that the decorated function also acquires a new attribute “
called” that can be connected to callback functions of your choice. Then, whenever the gui widget or the original function are called, the result will be passed to your callback function:
@image_arithmetic.called.connect def print_mean(value): """Callback function that accepts an event""" # the value attribute has the result of calling the function print(np.mean(value))
>>> image_arithmetic() 1.0060037881040373