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

Jump to bottom

fix: Umap duplicate index #489

Open
wants to merge 7 commits into
base: master
Choose a base branch
from
Open

fix: Umap duplicate index #489

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
db40c9e
fix: Umap duplicate index
tanmoyio Jun 13, 2023
8f3b839
Merge branch 'master' of https://github.com/graphistry/pygraphistry i…
tanmoyio Aug 1, 2023
e2f2b6c
added umap_learn and cuml tests
tanmoyio Aug 2, 2023
6141cd3
removed old index
tanmoyio Aug 2, 2023
aca1c0c
lint
dcolinmorgan Oct 5, 2023
f8bcde0
update branch, lint
dcolinmorgan Oct 5, 2023
394e877
merge updated ci
dcolinmorgan Nov 30, 2023
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
36 changes: 36 additions & 0 deletions graphistry/tests/test_umap_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -74,6 +74,24 @@
node_numeric = node_ints + node_floats
node_target = triangleNodes[["y"]]

node_graph_with_index = pd.DataFrame(
{
"index": range(1, 13),
"a": ["a", "b", "c", "d"] * 3,
"b": ["w", "x", "y", "z"] * 3,
}
)

edge_graph_with_index = pd.DataFrame(
{
"index": range(1, 13),
"a": ["a", "b", "c", "d"] * 3,
"b": ["w", "x", "y", "z"] * 3,
"src": [1, 2, 3, 4] * 3,
"dst": [4, 3, 1, 2] * 3,
}
)

def _eq(df1, df2):
try:
df1 = df1.to_pandas()
Expand Down Expand Up @@ -150,6 +168,15 @@ def setUp(self):
)
self.g2e = g2

# graph with index
self.g_index_nodes = graphistry.nodes(node_graph_with_index)
self.g_index_nodes_umaped = self.g_index_nodes.umap(engine="umap_learn")
assert "_n" == self.g_index_nodes_umaped._node

self.g_index_edges = graphistry.nodes(edge_graph_with_index)
self.g_index_edges_umaped = self.g_index_edges.umap(engine="umap_learn")
assert "_n" == self.g_index_edges_umaped._node


@pytest.mark.skipif(not has_umap, reason="requires umap feature dependencies")
def test_columns_match(self):
Expand Down Expand Up @@ -816,6 +843,15 @@ def test_base(self):
graphistry.nodes(self.df).umap('auto')._node_embedding.shape == (self.samples, 2)
graphistry.nodes(self.df).umap('engine')._node_embedding.shape == (self.samples, 2)

# graph with index
self.g_index_nodes = graphistry.nodes(node_graph_with_index)
self.g_index_nodes_umaped = self.g_index_nodes.umap(engine="cuml")
assert "_n" == self.g_index_nodes_umaped._node

self.g_index_edges = graphistry.nodes(edge_graph_with_index)
self.g_index_edges_umaped = self.g_index_edges.umap(engine="cuml")
assert "_n" == self.g_index_edges_umaped._node


if __name__ == "__main__":
unittest.main()
7 changes: 4 additions & 3 deletions graphistry/umap_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -588,12 +588,13 @@ def umap(

if kind == "nodes":
index = res._nodes.index

if res._node is None:
logger.debug("-Writing new node name")
res._nodes[config.IMPLICIT_NODE_ID] = range(len(res._nodes))

Copy link
Contributor

Choose a reason for hiding this comment

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

Clever

Will this work if cudf?

Will this break somehow downstream if .edges is bound and user was using string name IDs for src dst?

Copy link
Contributor

Choose a reason for hiding this comment

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

Strange, I thought this was already done this way...

res = res.nodes( # type: ignore
res._nodes.reset_index(drop=True)
.reset_index()
.rename(columns={"index": config.IMPLICIT_NODE_ID}),
Copy link
Contributor

Choose a reason for hiding this comment

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

Why not just call it __index_umap__ or something that has low prob of collision? then we can keep it safe for cudf?

res._nodes,
config.IMPLICIT_NODE_ID,
)
res._nodes.index = index
Expand Down
Loading