Edge-based temporal sampling #280
Conversation
for more information, see https://pre-commit.ci
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## master #280 +/- ##
==========================================
- Coverage 85.74% 85.65% -0.10%
==========================================
Files 32 32
Lines 1087 1115 +28
==========================================
+ Hits 932 955 +23
- Misses 155 160 +5 ☔ View full report in Codecov by Sentry. |
| @@ -12,6 +12,8 @@ | |||
| #include "pyg_lib/csrc/utils/cpu/convert.h" | |||
| #include "pyg_lib/csrc/utils/types.h" | |||
|
|
|||
| #include <iostream> | |||
| pyg::sampler::Mapper<node_t, scalar_t>& dst_mapper, | ||
| pyg::random::RandintEngine<scalar_t>& generator, | ||
| std::vector<node_t>& out_global_dst_nodes) { | ||
| void edge_temporal_sample(const node_t global_src_node, |
There was a problem hiding this comment.
Please keep the order of argument same as in the temporal_sample function:
const int64_t count, const temporal_t seed_time, const temporal_t* edge_time,
| if ((row_end - row_start == 0) || (count == 0)) | ||
| return; | ||
| // Find new `row_end` such that all neighbors fulfill temporal constraints: | ||
| std::vector<int> v(row_end - row_start); |
There was a problem hiding this comment.
| std::vector<int> v(row_end - row_start); | |
| std::vector<scalar_t> v(row_end - row_start); |
|
@kgajdamo Thanks for the review. I think all of your comments are already resolved in latest commit :) |
I came back to this review after a while and didn't notice it was merged and left the comments 🥲 |
This PR is to enable the edge-based temporal sampling for NeighborSampler. This PR covers both homogeneous and heterogeneous cases. The associated PYG-LIB PR is pyg-team/pyg-lib#280 . Thanks, Poovaiah --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Akihiro Nitta <[email protected]> Co-authored-by: Matthias Fey <[email protected]>
This PR is to enable edge-based temporal sampling for both homogeneous and heterogeneous graphs.
Thanks,
Poovaiah