refactor(sampler): clean up sampler attributes, part 3 #5402
Conversation
This file contains 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
Codecov Report
@@ Coverage Diff @@
## master #5402 +/- ##
==========================================
- Coverage 83.35% 83.33% -0.02%
==========================================
Files 343 342 -1
Lines 18800 18784 -16
==========================================
- Hits 15671 15654 -17
- Misses 3129 3130 +1
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
…ric into remote_backend_4
Padarn
reviewed
Sep 10, 2022
Padarn
approved these changes
Sep 10, 2022
wsad1
approved these changes
Sep 12, 2022
mananshah99
added a commit
that referenced
this pull request
Sep 13, 2022
…n classes, part 4 (#5404) This PR continues the effort to consolidate PyG's sampling interface in preparation for moving `sample(...)` behind the `GraphStore` interface. This effort is somewhat large in scope and will be broken into multiple PRs for ease of review. It builds off of #5402, and makes a significant move to abstract data loading behind a `data: Union[Data, HeteroData, Tuple[FeatureStore, GraphStore]]` and a `sampler: BaseSampler`. It does so by introducing two base implementation classes: `NodeLoader` and `LinkLoader`. `NodeLoader` performs sampling from nodes (using `sample_from_nodes`), and `LinkLoader` does the same from edges (using `sample_from_edges`). They both expose parameters in their initializers that are intended for **loading** (that is, the process of using a sampler to get subgraphs, using a feature fetcher to get features, and joining these together to construct a `HeteroData` object to pass downstream). Samplers are intended to expose parameters that are used for **sampling** (that are particular to the sampling method). The implementations of `NeighborLoader` and `LinkNeighborLoader` are now very simple: they pass the `NeighborSampler` and any necessary initialization parameters directly in `__init__`, with no other change.
14 tasks
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.
This PR continues the effort to consolidate PyG's sampling interface in preparation for moving
sample(...)behind theGraphStoreinterface. This effort is somewhat large in scope and will be broken into multiple PRs for ease of review. It resolves some TODOs from #5365.The major change removes special handling of
input_typeandperm/perm_dictin the neighbor loader. This enables a full separation between the sampler interface and the NeighborLoader (the only other interaction that is not clearly documented or exposed well in the sampler class is*SamplerOutput.metadata, which will take a bit more work to properly define.