Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Node grouping API #4427

Open
wants to merge 10 commits into
base: main
Choose a base branch
from
+135 −0
Open

Node grouping API #4427

Show file tree
Hide file tree
Changes from 7 commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
2ff33b2
Add deployment related attributes
ankatiyar Jan 17, 2025
0d6461a
Merge branch 'main' into grouping
ankatiyar Jan 27, 2025
ef4d664
Update with feedback
ankatiyar Jan 29, 2025
a299ef3
Minor formatting
ankatiyar Jan 31, 2025
b79f1f5
Add test
ankatiyar Feb 3, 2025
a8b1f45
Update test
ankatiyar Feb 3, 2025
9d736b9
Merge branch 'main' into grouping
ankatiyar Feb 3, 2025
5e0bc8c
Merge branch 'main' into grouping
ankatiyar Feb 10, 2025
ada46e6
Update with feedback
ankatiyar Feb 10, 2025
cee4a2c
break up test
ankatiyar Feb 10, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 25 additions & 0 deletions kedro/pipeline/pipeline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -369,6 +369,31 @@ def grouped_nodes(self) -> list[list[Node]]:

return [list(group) for group in self._toposorted_groups]

@property
def grouped_nodes_by_namespace(self) -> dict[str, dict[str, Any]]:
"""Return a dictionary of the pipeline nodes grouped by namespace with
information about the nodes, their type, and dependencies."""
grouped_nodes: dict[str, dict[str, Any]] = defaultdict(dict)
for node in self.nodes:
key = node.namespace or node.name
if key not in grouped_nodes:
grouped_nodes[key] = {}
grouped_nodes[key]["name"] = key
grouped_nodes[key]["type"] = "namespace" if node.namespace else "node"
grouped_nodes[key]["nodes"] = [*grouped_nodes[key].get("nodes", []), node]
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we have O(Nˆ2) complexity here, as every time we add a node, we recreate a list based on what was already there. So maybe it makes sense to initialize grouped_nodes[key]["nodes"] and then just append to it.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You mean when essentially every node would be in a separate group? Maybe we can get around that by checking straight away if there are any namespaces in the pipeline and if not just return all nodes separately.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I mean just don't unpack the list but rather append to it, like

for node in self.nodes:
    key = node.namespace or node.name
    if key not in grouped_nodes:
        grouped_nodes[key] = {}
        grouped_nodes[key]["name"] = key
        grouped_nodes[key]["type"] = "namespace" if node.namespace else "node"
        grouped_nodes[key]["nodes"] = []
    grouped_nodes[key]["nodes"].append(node)

Nothing should change structure-wise from the current implementation.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But maybe we should also check if there are any namespaces at all. Don't know without looking into the code if that's easy. If there are no namespaces in the pipeline at all, it doesn't make sense to call this method and the user should probably be made aware.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added the suggestion by @ElenaKhaustova
@merelcht
Even to check for namespaces, I believe we'd have to check all the nodes for node.namespace - if the node doesn't have a namespace, it's treated as its own group. Maybe this check belongs on the plugin side?

dependencies = set()
for parent in self.node_dependencies[node]:
if parent.namespace and parent.namespace != key:
dependencies.add(parent.namespace)
elif parent.namespace and parent.namespace == key:
continue
else:
dependencies.add(parent.name)
grouped_nodes[key]["dependencies"] = (
ankatiyar marked this conversation as resolved.
Show resolved Hide resolved
grouped_nodes[key].get("dependencies", set()) | dependencies
)
return grouped_nodes

def only_nodes(self, *node_names: str) -> Pipeline:
"""Create a new ``Pipeline`` which will contain only the specified
nodes by name.
Expand Down
31 changes: 31 additions & 0 deletions tests/pipeline/test_pipeline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -376,6 +376,37 @@ def test_node_dependencies(self, complex_pipeline):
}
assert actual == expected

def test_node_grouping_by_namespace(self):
pipeline = modular_pipeline(
[
node(identity, "A", "B", name="node1", namespace="name_1"),
node(identity, "B", "C", name="node2", namespace="name_1"),
node(identity, "C", "D", name="node3", namespace="name_2"),
node(identity, "D", "E", name="node4", namespace="name_2"),
node(identity, "E", "G", name="node5"),
node(identity, "G", "H", name="node6"),
]
)
grouped = pipeline.grouped_nodes_by_namespace
# Validate keys for namespace groups
for key in ["name_1", "name_2"]:
assert key in grouped
assert grouped[key]["name"] == key
assert grouped[key]["type"] == "namespace"
assert len(grouped[key]["nodes"]) == 2

# Validate dependencies for namespace groups
assert grouped["name_1"]["dependencies"] == set()
assert grouped["name_2"]["dependencies"] == {"name_1"}

# Validate nodes for namespace groups
assert grouped["node5"]["type"] == "node"
assert grouped["node5"]["name"] == "node5"
assert len(grouped["node5"]["nodes"]) == 1
assert grouped["node5"]["dependencies"] == {"name_2"}
# Validate when node depends on node
assert grouped["node6"]["dependencies"] == {"node5"}


@pytest.fixture
def pipeline_with_circle():
Expand Down
Loading