Fix MessagePassing output shape for isolated target cells - #375
Merged
Conversation
The scatter aggregation inferred the output size from the largest message index, silently dropping target cells without incoming messages and contradicting the documented output shape. Pass the number of target cells from the neighborhood matrix as dim_size so that every target cell receives an output row, zero-filled when its neighborhood is empty. Fixes #136 Also resolves the reported Conv scenario of #228, where an isolated last cell previously led to a wrong output shape.
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #375 +/- ##
=======================================
Coverage 96.43% 96.44%
=======================================
Files 58 58
Lines 2078 2079 +1
=======================================
+ Hits 2004 2005 +1
Misses 74 74 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
There was a problem hiding this comment.
Pull request overview
This PR updates MessagePassing aggregation to preserve the documented output shape (..., n_target_cells, out_channels) by ensuring scatter aggregation uses the neighborhood matrix’s target dimension as dim_size, so isolated target cells still produce an output row (zero-filled).
Changes:
- Pass
dim_sizeinto the scatter aggregation so output includes all target cells, including those with empty neighborhoods. - Record
n_target_cellsfromneighborhood.shape[0]duringforward()for use during aggregation. - Add a regression test covering an isolated target cell output shape.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.
| File | Description |
|---|---|
| topomodelx/base/message_passing.py | Forces scatter output to include all target cells by supplying dim_size based on neighborhood shape. |
| test/base/test_message_passing.py | Adds a regression test for correct output shape when a target cell is isolated. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Comment on lines
191
to
+197
| aggr = scatter(self.aggr_func) | ||
| return aggr(x_message, self.target_index_i, 0) | ||
| return aggr( | ||
| x_message, | ||
| self.target_index_i, | ||
| 0, | ||
| dim_size=getattr(self, "n_target_cells", None), | ||
| ) |
Comment on lines
+125
to
+126
| result = self.mp.forward(x_source, neighborhood) | ||
| assert result.shape == (2, 10) |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The scatter aggregation inferred the output size from the largest message index, silently dropping target cells without incoming messages and contradicting the documented output shape.
Pass the number of target cells from the neighborhood matrix as dim_size so that every target cell receives an output row, zero-filled when its neighborhood is empty.
Fixes #136
Also resolves the reported Conv scenario of #228, where an isolated last cell previously led to a wrong output shape.