def generate_multiple_landmark_annotation_sets(retina_grader=False):
graders = (UserFactory(), UserFactory(), UserFactory())
if retina_grader:
add_to_graders_group(graders)
landmarksets = (
LandmarkAnnotationSetFactory(grader=graders[0]),
LandmarkAnnotationSetFactory(grader=graders[1]),
LandmarkAnnotationSetFactory(grader=graders[0]),
LandmarkAnnotationSetFactory(grader=graders[2]),
)
# Create child models for landmark annotation set
singlelandmarkbatches = (
SingleLandmarkAnnotationFactory.create_batch(
2, annotation_set=landmarksets[0]),
SingleLandmarkAnnotationFactory.create_batch(
5, annotation_set=landmarksets[1]),
[],
[],
)
images = [
Image.objects.filter(
singlelandmarkannotation__annotation_set=landmarksets[0].id),
Image.objects.filter(
singlelandmarkannotation__annotation_set=landmarksets[1].id),
[],
[],
]
# Create singlelandmarkannotations with the images of landmarkset1
for image in images[0]:
singlelandmarkbatches[2].append(
SingleLandmarkAnnotationFactory.create(
annotation_set=landmarksets[2], image=image))
images[2].append(image)
singlelandmarkbatches[3].append(
SingleLandmarkAnnotationFactory.create(
annotation_set=landmarksets[3], image=image))
images[3].append(image)
singlelandmarkbatches[2].append(
SingleLandmarkAnnotationFactory.create(annotation_set=landmarksets[2]))
images[2].append(singlelandmarkbatches[2][-1].image)
return MultipleLandmarkAnnotationSets(
grader1=graders[0],
grader2=graders[1],
grader3=graders[2],
landmarkset1=landmarksets[0],
landmarkset1images=images[0],
landmarkset2=landmarksets[1],
landmarkset2images=images[1],
landmarkset3=landmarksets[2],
landmarkset3images=images[2],
landmarkset4=landmarksets[3],
landmarkset4images=images[3],
)
def generate_multiple_landmark_annotation_sets(retina_grader=False):
graders = (UserFactory(), UserFactory())
if retina_grader:
add_to_graders_group(graders)
landmarksets = (
LandmarkAnnotationSetFactory(grader=graders[0]),
LandmarkAnnotationSetFactory(grader=graders[1]),
LandmarkAnnotationSetFactory(grader=graders[0]),
)
# Create child models for landmark annotation set
singlelandmarkbatches = (
SingleLandmarkAnnotationFactory.create_batch(
2, annotation_set=landmarksets[0]
),
SingleLandmarkAnnotationFactory.create_batch(
5, annotation_set=landmarksets[1]
),
[],
)
images = [
Image.objects.filter(
singlelandmarkannotation__annotation_set=landmarksets[0].id
),
Image.objects.filter(
singlelandmarkannotation__annotation_set=landmarksets[1].id
),
[],
]
# Create singlelandmarkannotations with the images of landmarkset1
for image in images[0]:
singlelandmarkbatches[2].append(
SingleLandmarkAnnotationFactory.create(
annotation_set=landmarksets[2], image=image
)
)
images[2].append(image)
singlelandmarkbatches[2].append(
SingleLandmarkAnnotationFactory.create(annotation_set=landmarksets[2])
)
images[2].append(singlelandmarkbatches[2][-1].image)
return MultipleLandmarkAnnotationSets(
grader1=graders[0],
grader2=graders[1],
landmarkset1=landmarksets[0],
landmarkset1images=images[0],
landmarkset2=landmarksets[1],
landmarkset2images=images[1],
landmarkset3=landmarksets[2],
landmarkset3images=images[2],
)
def test_landmarks_single(self):
u = UserFactory()
i1 = ImageFactory()
i2 = ImageFactory()
las = LandmarkAnnotationSetFactory(grader=u)
SingleLandmarkAnnotationFactory(annotation_set=las, image=i1)
SingleLandmarkAnnotationFactory(annotation_set=las, image=i2)
landmarks = RetinaImageSerializer(i1).data["landmark_annotations"]
assert len(landmarks) == 1
assert landmarks == [i2.pk]
def test_landmarks_multiple(self):
u = UserFactory()
i1 = ImageFactory()
img_pks = set()
for _ in range(4):
img = ImageFactory()
las = LandmarkAnnotationSetFactory(grader=u)
SingleLandmarkAnnotationFactory(annotation_set=las, image=img)
SingleLandmarkAnnotationFactory(annotation_set=las, image=i1)
img_pks.add(img.pk)
landmarks = RetinaImageSerializer(i1).data["landmark_annotations"]
assert len(landmarks) == 4
assert set(landmarks) == img_pks
def generate_annotation_set(retina_grader=False):
grader = UserFactory()
if retina_grader:
add_to_graders_group([grader])
measurement = MeasurementAnnotationFactory(grader=grader)
boolean = BooleanClassificationAnnotationFactory(grader=grader)
integer = IntegerClassificationAnnotationFactory(grader=grader)
polygon = PolygonAnnotationSetFactory(grader=grader)
coordinatelist = CoordinateListAnnotationFactory(grader=grader)
landmark = LandmarkAnnotationSetFactory(grader=grader)
etdrs = ETDRSGridAnnotationFactory(grader=grader)
# Create child models for polygon annotation set
SinglePolygonAnnotationFactory.create_batch(10, annotation_set=polygon)
# Create child models for landmark annotation set (3 per image)
for i in range(5):
image = ImageFactory()
SingleLandmarkAnnotationFactory(annotation_set=landmark, image=image)
return AnnotationSet(
grader=grader,
measurement=measurement,
boolean=boolean,
polygon=polygon,
coordinatelist=coordinatelist,
landmark=landmark,
etdrs=etdrs,
integer=integer,
)
def create_load_data(data_type, ds, grader):
if data_type == "Registration":
model = LandmarkAnnotationSetFactory(grader=grader)
SingleLandmarkAnnotationFactory(
annotation_set=model, image=ds["image_cf"]
),
if ds["archive"].name == "Australia":
# Australia does not allow obs images so create a new cf image for Australia test
img = ImageFactory(study=ds["study"])
SingleLandmarkAnnotationFactory(annotation_set=model, image=img)
else:
SingleLandmarkAnnotationFactory(
annotation_set=model, image=ds["image_obs"]
),
elif data_type == "ETDRS":
model = ETDRSGridAnnotationFactory(grader=grader, image=ds["image_cf"])
elif data_type == "GA" or data_type == "kappa":
model_macualar = PolygonAnnotationSetFactory(
grader=grader, image=ds["image_cf"], name="macular"
)
SinglePolygonAnnotationFactory(annotation_set=model_macualar)
SinglePolygonAnnotationFactory(annotation_set=model_macualar)
SinglePolygonAnnotationFactory(annotation_set=model_macualar)
model_peripapillary = PolygonAnnotationSetFactory(
grader=grader, image=ds["image_cf"], name="peripapillary"
)
SinglePolygonAnnotationFactory(annotation_set=model_peripapillary)
SinglePolygonAnnotationFactory(annotation_set=model_peripapillary)
SinglePolygonAnnotationFactory(annotation_set=model_peripapillary)
model = [model_macualar, model_peripapillary]
elif data_type == "Measure":
model = [
MeasurementAnnotationFactory(grader=grader, image=ds["image_cf"]),
MeasurementAnnotationFactory(grader=grader, image=ds["image_cf"]),
MeasurementAnnotationFactory(grader=grader, image=ds["image_cf"]),
]
elif data_type == "Fovea":
model = BooleanClassificationAnnotationFactory(
grader=grader, image=ds["image_cf"], name="fovea_affected"
)
return model
def generate_annotation_set(retina_grader=False, image=False):
grader = UserFactory()
if retina_grader:
add_to_graders_group([grader])
create_options = {"grader": grader}
if image:
create_options_with_image = {"image": image, **create_options}
else:
create_options_with_image = create_options
measurement = MeasurementAnnotationFactory(**create_options_with_image)
boolean = BooleanClassificationAnnotationFactory(
**create_options_with_image
)
integer = IntegerClassificationAnnotationFactory(
**create_options_with_image
)
polygon = PolygonAnnotationSetFactory(**create_options_with_image)
coordinatelist = CoordinateListAnnotationFactory(
**create_options_with_image
)
etdrs = ETDRSGridAnnotationFactory(**create_options_with_image)
landmark = LandmarkAnnotationSetFactory(**create_options)
# Create child models for polygon annotation set
SinglePolygonAnnotationFactory.create_batch(10, annotation_set=polygon)
# Create child models for landmark annotation set (3 per image)
single_landmarks = []
for i in range(5):
if i > 0 or not image:
image = ImageFactory()
single_landmarks.append(
SingleLandmarkAnnotationFactory(
annotation_set=landmark, image=image
)
)
return AnnotationSet(
grader=grader,
measurement=measurement,
boolean=boolean,
polygon=polygon,
coordinatelist=coordinatelist,
landmark=landmark,
singlelandmarks=single_landmarks,
etdrs=etdrs,
integer=integer,
)