-
Notifications
You must be signed in to change notification settings - Fork 302
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.
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
Add node2vec
wrapper to cugraph
#2093
Add node2vec
wrapper to cugraph
#2093
Conversation
…bcugraph Pylibcugraph wrappers for cugraph dev
…bcugraph Branch 22.04 fea node2vec pylibcugraph
…bcugraph Branch 22.04 fea node2vec pylibcugraph
Codecov Report
@@ Coverage Diff @@
## branch-22.04 #2093 +/- ##
================================================
+ Coverage 73.70% 73.99% +0.29%
================================================
Files 155 157 +2
Lines 10373 10496 +123
================================================
+ Hits 7645 7767 +122
- Misses 2728 2729 +1
Continue to review full report at Codecov.
|
Fetching from 22.04 after random walk bug was merged
Fetching from 22.04 after random walk bug was merged
rerun tests |
Automatically clone raft when the raft pinned tag changes (rapidsai#2087)
rerun tests |
# are contained in Simple_1 | ||
_test_data = {"karate.csv": { | ||
"seeds": cp.asarray([0, 0], dtype=np.int32), | ||
"paths": cp.asarray([0, 8, 33, 29, 26, 0, 1, 3, 13, 33], |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think this is good for now but I am a little concern with this approach of testing because of the randomness of the results.
- did you make sure there are only 2 paths with depth 5 starting from seed 0?
- Even if it is the case, the testing seems hardcoded and only a few seeds are tested.
An alternative approach could be to pick random seeds and verify the accuracy of the paths by checking the predecessors. To illustrate this, let's take the first path
[0, 8, 33, 29, 26]. you verify that 0 is a predecessor of 8, then 8 being a predecessor of 33 and so on... but I am not sure how to easily do this without involving cugraph.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This approach of testing was also in the cugraph level, but yes when this was done I did not check that the walks taken were in fact valid, just putting placeholder values and focusing on making sure the start of each path was each of the seeds. I've since made the testing more robust on the cugraph side, but for pylibcugraph it's more complicated, I'm still figuring that out right now but I can include a FIXME or add an issue to revisit later
…bcugraph Branch 22.04 fea node2vec pylibcugraph
Update for building
…bcugraph Branch 22.04 fea node2vec pylibcugraph
rerun tests |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great job on this. Testing especially.
rerun tests |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is looking good, thanks. I have a few comments below based on what I found so far. Aside from that, I'd also like to look into the renumbering comments Joseph posted so I may have a follow-up review.
Include commits made mar 8-9
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I noticed one other minor change to make, plus verifying the renumbering questions, then it looks ready to approve.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me
@gpucibot merge |
Merged PR #2093 requires `pylibcugraph`. This PR adds `pylibcugraph` as a run dependency to the `cugraph` conda package. Authors: - Rick Ratzel (https://github.com/rlratzel) Approvers: - Jordan Jacobelli (https://github.com/Ethyling) URL: #2121
Helps close #2058, when combined with #2085. Requires #2085 to be merged before merging, as the cugraph implementation requires the pylibcugraph node2vec wrapper