def util_get_nml():
import wknml
trees = [
wknml.Tree(
id=0,
color=(255, 255, 0, 1),
name="Synapse 1",
nodes=[
wknml.Node(id=0, position=(1, 1, 1), radius=1),
wknml.Node(id=1, position=(1, 1, 2), radius=1),
wknml.Node(id=2, position=(1, 1, 3), radius=1),
wknml.Node(id=3, position=(1, 2, 2), radius=1)
],
edges=[
wknml.Edge(source=0, target=1),
wknml.Edge(source=1, target=2),
wknml.Edge(source=1, target=3)
],
groupId=1,
),
wknml.Tree(
id=1,
color=(255, 0, 255, 1),
name="Synapse 2",
nodes=[
wknml.Node(id=0, position=(2, 1, 1), radius=1),
wknml.Node(id=1, position=(3, 1, 1), radius=1),
wknml.Node(id=2, position=(4, 1, 1), radius=1)
],
edges=[
wknml.Edge(source=0, target=1),
wknml.Edge(source=1, target=2)
],
groupId=1,
),
]
nml = wknml.NML(
parameters=wknml.NMLParameters(
name="Test",
scale=(1, 1, 1),
),
trees=trees,
branchpoints=[],
comments=[],
groups=[],
)
return nml
def test_calc_efpl_simple():
import wknml
nml = wknml.NML(
parameters=wknml.NMLParameters(
name='',
scale=(1, 1, 1),
),
trees=[
wknml.Tree(
id=0,
color=(255, 255, 0, 1),
name='',
nodes=[
wknml.Node(id=4, position=(2, 1, 1), radius=1),
wknml.Node(id=5, position=(3, 1, 1), radius=1)
],
edges=[
wknml.Edge(source=4, target=5),
],
),
],
branchpoints=[],
comments=[],
groups=[],
)
Vlbls = np.ones((5, 5, 5), dtype=np.uint32)
dataset_start = np.zeros((1, 3), dtype=int)
calc_eftpl(nml, Vlbls, dataset_start)
def test__calc_efpl__use_correct_edge_lengths():
""" Github issue: https://github.com/elhuhdron/emdrp/issues/4
Check that the norm between edge node positions is calculated along the
correct axis.
"""
import wknml
trees = [
wknml.Tree(
id=0,
color=(255, 255, 0, 1),
name="neuron1",
nodes=[
wknml.Node(id=0, position=(1, 1, 1), radius=1),
wknml.Node(id=1, position=(1, 1, 3), radius=1),
wknml.Node(id=2, position=(1, 2, 1), radius=1)
],
edges=[
wknml.Edge(source=0, target=1),
wknml.Edge(source=0, target=2)
],
groupId=1,
)
]
nml = wknml.NML(
parameters=wknml.NMLParameters(
name="Test",
scale=(1, 1, 1),
),
trees=trees,
branchpoints=[],
comments=[],
groups=[],
)
Vlbls = np.ones((5, 5, 5), dtype=np.uint32)
dataset_start = np.zeros((1, 3), dtype=int)
Vlbls[:, 1:, :] = 2
efpl = calc_eftpl(nml, Vlbls, dataset_start)
assert (efpl == 2 / 3)
def _skeleton_to_nml(self):
""" Converts skeleton to wknml data structures."""
trees = []
for tree_idx, tree_id in enumerate(self.tree_ids):
nml_nodes = Skeleton._nodes_to_nml_nodes(self.nodes[tree_idx])
nml_edges = Skeleton._edges_to_nml_edges(self.edges[tree_idx])
tree = wknml.Tree(id=tree_id,
color=self.colors[tree_idx],
name=self.names[tree_idx],
groupId=self.group_ids[tree_idx],
nodes=nml_nodes,
edges=nml_edges)
trees.append(tree)
nml = wknml.NML(
parameters=wknml.NMLParameters(**self.parameters._asdict()),
trees=trees,
branchpoints=self.branchpoints,
comments=self._skeleton_to_nml_comments(),
groups=self.groups)
return nml
old_new_mapping = defaultdict(list)
new_trees = []
for i in range(n_components):
node_ids, = np.where(labels == i)
node_ids = node_ids.tolist()
if len(node_ids) == 1 and node_ids[0] not in all_node_ids:
continue
old_tree = find(lambda t: any(n.id in node_ids for n in t.nodes),
nml.trees)
new_tree = wknml.Tree(
id=i,
color=old_tree.color,
name=old_tree.name,
groupId=old_tree.groupId,
nodes=[n for n in all_nodes if n.id in node_ids],
edges=[
e for e in all_edges
if e.source in node_ids or e.target in node_ids
],
)
old_new_mapping[old_tree.id].append(i)
new_trees.append(new_tree)
new_trees_with_groups = []
new_groups = []
for i, (old_id, new_ids) in enumerate(old_new_mapping.items()):
group_id = i + 1
old_tree = find(lambda t: t.id == old_id, nml.trees)
new_groups.append(wknml.Group(id=group_id, name=old_tree.name,
children=[]))
