diff --git a/model_server/extensions/chaeo/params.py b/model_server/extensions/chaeo/params.py
index 71b9497dd76ef8f2a116e1573d37d55a22012520..9e06b286498a21bf1115b15f26f99b10547d9808 100644
--- a/model_server/extensions/chaeo/params.py
+++ b/model_server/extensions/chaeo/params.py
@@ -34,8 +34,8 @@ class RoiSetMetaParams(BaseModel):
class RoiSetExportParams(BaseModel):
pixel_probabilities: bool = False
patches_3d: Union[PatchParams, None] = None
- patches_2d_for_annotation: Union[PatchParams, None] = None
- patches_2d_for_training: Union[PatchParams, None] = None
+ annotated_patches_2d: Union[PatchParams, None] = None
+ patches_2d: Union[PatchParams, None] = None
patch_masks: Union[PatchParams, None] = None
annotated_zstacks: Union[AnnotatedZStackParams, None] = None
object_classes: bool = False
diff --git a/model_server/extensions/chaeo/products.py b/model_server/extensions/chaeo/products.py
index a99e8c63b2bf9702146d8e3c375a1e998b3129be..bd643df7775d91c556885b51f43f49647ec73b10 100644
--- a/model_server/extensions/chaeo/products.py
+++ b/model_server/extensions/chaeo/products.py
@@ -227,21 +227,13 @@ def export_patches_from_zstack(
make_3d = kwargs.get('make_3d', False)
patches_df = get_patches_from_zmask_meta(roiset, **kwargs)
- pc = roiset.acc_raw.chroma # TODO: this should follow from generated patches, not roiset
- # patches = list(patches_df['patch'])
- # if not make_3d and pc == 1:
- # patches_acc = MonoPatchStack(patches)
- # else:
- # patches_acc = Multichannel3dPatchStack(patches)
- patches_acc = roiset.get_raw_patches(make_3d=make_3d)
-
def _export_patch(roi):
- patch = patches_acc.iat_yxcz(i)
- ext = 'tif' if make_3d or patches_acc.chroma > 3 else 'png'
+ patch = InMemoryDataAccessor(roi.patch)
+ ext = 'tif' if make_3d or patch.chroma > 3 else 'png'
fname = f'{prefix}-la{roi.label:04d}-zi{roi.zi:04d}.{ext}'
if patch.dtype is np.dtype('uint16'):
- write_patch_to_file(where, fname, resample_to_8bit(patch))
+ write_patch_to_file(where, fname, resample_to_8bit(patch.data))
else:
write_patch_to_file(where, fname, patch)
@@ -252,7 +244,7 @@ def export_patches_from_zstack(
})
exported = []
- for i, roi in enumerate(patches_df.itertuples()): # just used for label info
+ for roi in patches_df.itertuples(): # just used for label info
_export_patch(roi)
return exported
diff --git a/model_server/extensions/chaeo/tests/test_zstack.py b/model_server/extensions/chaeo/tests/test_zstack.py
index 61565ec4006fadf3d5e8007f5b70908922015529..c07deb807d44afc0c13a0b3fa811eaf0d36725bd 100644
--- a/model_server/extensions/chaeo/tests/test_zstack.py
+++ b/model_server/extensions/chaeo/tests/test_zstack.py
@@ -47,13 +47,14 @@ class TestRoiSetMonoProducts(BaseTestRoiSetMonoProducts, unittest.TestCase):
id_map,
self.stack_ch_pa,
params=RoiSetMetaParams(
- mask_type=mask_type, filters=kwargs.get('filters')
+ mask_type=mask_type,
+ filters=kwargs.get('filters', {'area': {'min': 1e3, 'max': 1e4}})
)
)
return roiset
def test_zmask_makes_correct_boxes(self, **kwargs):
- roiset = self._make_roi_set(mask_type='boxes', **kwargs)
+ roiset = self._make_roi_set(**kwargs)
zmask = roiset.get_zmask()
meta = roiset.zmask_meta
interm = roiset.interm
@@ -96,7 +97,7 @@ class TestRoiSetMonoProducts(BaseTestRoiSetMonoProducts, unittest.TestCase):
return self._make_roi_set(mask_type='contours')
def test_zmask_makes_correct_boxes_with_filters(self):
- return self._make_roi_set(filters={'area': {'min': 1e3, 'max': 1e4}})
+ return self._make_roi_set()
def test_zmask_makes_correct_expanded_boxes(self):
return self._make_roi_set(expand_box_by=(64, 2))
@@ -110,18 +111,18 @@ class TestRoiSetMonoProducts(BaseTestRoiSetMonoProducts, unittest.TestCase):
def test_zmask_rel_slices_are_valid(self):
roiset = self._make_roi_set()
- for i, s in enumerate(roiset.get_slices()):
- ebb = roiset.acc_raw.data[s]
+ # for i, s in enumerate(roiset.get_slices()):
+ for roi in roiset.get_df().itertuples():
+ ebb = roiset.acc_raw.data[roi.slice]
self.assertEqual(len(ebb.shape), 4)
self.assertTrue(np.all([si >= 1 for si in ebb.shape]))
- rel_slices = roiset.get_df()['relative_slice']
- rbb = ebb[rel_slices[i]]
+ rbb = ebb[roi.relative_slice]
self.assertEqual(len(rbb.shape), 4)
self.assertTrue(np.all([si >= 1 for si in rbb.shape]))
def test_make_2d_patches_from_zmask(self):
roiset = self._make_roi_set(
- filters={'area': {'min': 1e3, 'max': 1e4}},
+ # filters={'area': {'min': 1e3, 'max': 1e4}},
expand_box_by=(64, 2)
)
files = export_patches_from_zstack(
@@ -133,7 +134,7 @@ class TestRoiSetMonoProducts(BaseTestRoiSetMonoProducts, unittest.TestCase):
def test_make_3d_patches_from_zmask(self):
roiset = self._make_roi_set(
- filters={'area': {'min': 1e3, 'max': 1e4}},
+ # filters={'area': {'min': 1e3, 'max': 1e4}},
expand_box_by=(64, 2),
)
files = export_patches_from_zstack(
@@ -142,6 +143,17 @@ class TestRoiSetMonoProducts(BaseTestRoiSetMonoProducts, unittest.TestCase):
make_3d=True)
self.assertGreaterEqual(len(files), 1)
+ def test_export_annotated_zstack(self):
+ roiset = self._make_roi_set(
+ # filters={'area': {'min': 1e3, 'max': 1e4}},
+ expand_box_by=(64, 2),
+ )
+ file = roiset.export_annotated_zstack(
+ output_path / 'annotated_stack',
+ )
+ result = generate_file_accessor(Path(file['location']) / file['filename'])
+ self.assertEqual(result.shape, roiset.acc_raw.shape)
+
def test_flatten_image(self):
id_map = get_label_ids(self.seg_mask)
@@ -167,7 +179,7 @@ class TestRoiSetMonoProducts(BaseTestRoiSetMonoProducts, unittest.TestCase):
def test_make_binary_masks_from_zmask(self):
roiset = self._make_roi_set(
- filters={'area': {'min': 1e3, 'max': 1e4}},
+ # filters={'area': {'min': 1e3, 'max': 1e4}},
expand_box_by=(128, 2)
)
files = roiset.export_patch_masks(output_path / '2d_mask_patches', )
@@ -175,11 +187,7 @@ class TestRoiSetMonoProducts(BaseTestRoiSetMonoProducts, unittest.TestCase):
def test_classify_by(self):
roiset = self._make_roi_set()
- roiset.classify_by(
- 'dummy_class',
- pipeline_params['patches_channel'],
- DummyInstanceSegmentationModel()
- )
+ roiset.classify_by('dummy_class', 0, DummyInstanceSegmentationModel())
self.assertTrue(all(roiset.get_df()['classify_by_dummy_class'].unique() == [1.]))
def test_object_map_workflow(self):
@@ -214,13 +222,13 @@ class TestRoiSetMonoProducts(BaseTestRoiSetMonoProducts, unittest.TestCase):
export_params = RoiSetExportParams(**{
'pixel_probabilities': True,
'patches_3d': {},
- 'patches_2d_for_annotation': {
+ 'annotated_patches_2d': {
'draw_bounding_box': True,
'rgb_overlay_channels': [3, None, None],
'rgb_overlay_weights': [0.2, 1.0, 1.0],
'pad_to': 512,
},
- 'patches_2d_for_training': {
+ 'patches_2d': {
'draw_bounding_box': False,
'draw_mask': False,
},
diff --git a/model_server/extensions/chaeo/zmask.py b/model_server/extensions/chaeo/zmask.py
index 02e5922adc1adadc9dfaef567b90d05cc1aada7c..31d5166b1ecc4ab0a6936ba1975f0b350306e7f8 100644
--- a/model_server/extensions/chaeo/zmask.py
+++ b/model_server/extensions/chaeo/zmask.py
@@ -167,12 +167,11 @@ class RoiSet(object):
self._df[name] = se
def get_multichannel_projection(self): # TODO: document and test
- dff = self.df[self.df['keeper']]
if self.count:
projected = project_stack_from_focal_points(
- dff['centroid-0'].to_numpy(),
- dff['centroid-1'].to_numpy(),
- dff['zi'].to_numpy(),
+ self._df['centroid-0'].to_numpy(),
+ self._df['centroid-1'].to_numpy(),
+ self._df['zi'].to_numpy(),
self.acc_raw,
degree=4,
)
@@ -200,6 +199,13 @@ class RoiSet(object):
else:
return Multichannel3dPatchStack(patches)
+ def export_annotated_zstack(self, where, prefix='zstack', **kwargs):
+ annotated = InMemoryDataAccessor(
+ draw_boxes_on_3d_image(self.acc_raw.data, self.zmask_meta, **kwargs) # TODO remove zmask_meta ref
+ )
+ success = write_accessor_data_to_file(where / (prefix + '.tif'), annotated)
+ return {'location': where.__str__(), 'filename': prefix + '.tif'}
+
def get_zmask(self, mask_type='boxes'):
"""
Return a mask of same dimensionality as raw data
@@ -237,6 +243,7 @@ class RoiSet(object):
return zi_st
+
# TODO: channel restriction as an argument
def classify_by(self, name: str, channel: int, object_classification_model: InstanceSegmentationModel, ):
@@ -248,12 +255,10 @@ class RoiSet(object):
lamap = self.object_id_labels
om = np.zeros(lamap.shape, dtype=lamap.dtype)
- # self.df['instance_class'] = np.nan
df = self.get_df()
idx = df.index
se = pd.Series(data=np.nan, index=idx)
- # df['classify_by_' + f]
# assign labels to object map:
for i in range(0, len(idx)):
@@ -281,31 +286,28 @@ class RoiSet(object):
files = export_patches_from_zstack(
subdir, self, white_channel=channel, prefix=pr, make_3d=True, **kp
)
- if k == 'patches_2d_for_annotation':
+ if k == 'annotated_patches_2d':
files = export_multichannel_patches_from_zstack(
- subdir, self.acc_raw, self.zmask_meta, prefix=pr, make_3d=False, rgb_white_channel=channel,
+ subdir, self, prefix=pr, make_3d=False, white_channel=channel,
bounding_box_channel=1, bounding_box_linewidth=2, **kp,
)
- if k == 'patches_2d_for_training':
+ if k == 'patches_2d':
files = export_multichannel_patches_from_zstack(
- subdir, self.acc_raw, self.zmask_meta, rgb_white_channel=channel, prefix=pr, make_3d=False, **kp
+ subdir, self, white_channel=channel, prefix=pr, make_3d=False, **kp
)
df_patches = pd.DataFrame(files)
- self.df = pd.merge(self.df, df_patches, left_index=True, right_on='df_index').drop(columns='df_index')
- self.df['patch_id'] = self.df.apply(lambda _: uuid4(), axis=1)
+ self._df = pd.merge(self._df, df_patches, left_index=True, right_on='df_index').drop(columns='df_index')
+ self._df['patch_id'] = self._df.apply(lambda _: uuid4(), axis=1)
if k == 'patch_masks':
- self.export_patch_masks(subdir, prefix=pr, **params.patch_masks)
+ self.export_patch_masks(subdir, prefix=pr, **kp)
if k == 'annotated_zstacks':
- annotated = InMemoryDataAccessor(
- draw_boxes_on_3d_image(raw_ch.data, self.zmask_meta, **kp) # TODO remove zmask_meta ref
- )
- write_accessor_data_to_file(subdir / (pr + '.tif'), annotated)
+ self.export_annotated_zstack(prefix=pr, **kp)
if k == 'object_classes':
write_accessor_data_to_file(subdir / (pr + '.tif'), self.object_class_map)
if k == 'dataframe':
dfpa = subdir / (pr + '.csv')
dfpa.parent.mkdir(parents=True, exist_ok=True)
- self.df.to_csv(dfpa, index=False)
+ self._df.to_csv(dfpa, index=False)
def project_stack_from_focal_points(