refactor[next]: Global ordering relation of dimensions #1847
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
SF-N
commented
Feb 7, 2025
| else: | ||
| assert isinstance(lhs, ts.FieldType) and isinstance(rhs, ts.FieldType) | ||
| assert lhs.dtype == rhs.dtype | ||
| return ts.FieldType(dims=common.promote_dims(*[lhs.dims, rhs.dims]), dtype=lhs.dtype) | ||
|
|
There was a problem hiding this comment.
This was necessary for the new test (im.plus(im.ref("inp1", float_i_field), im.ref("inp2", float_j_field)), float_ij_field) in test_type_inference.py. I am not sure if supporting this is actually desired.
SF-N
commented
Feb 7, 2025
| @@ -185,6 +189,8 @@ def expression_test_cases(): | |||
| ), | |||
| ts.DeferredType(constraint=None), | |||
| ), | |||
| # (im.as_fieldop(im.lambda_("x", "y")(im.plus(im.deref("x"), im.deref("y"))))( | |||
| # im.ref("inp1", float_i_field), im.ref("inp2", float_j_field)), float_ij_field), | |||
There was a problem hiding this comment.
Should something like this, i.e. not returning DeferredType but FieldType, work in the end as well ?
SF-N
changed the title
SF-N
changed the title
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Outdated suggestions cannot be applied.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 introduces a global ordering relation of dimensions, replacing the previous mechanism in promote_dims. The ordering relation is: sorting first by
Dimension.kind(HORIZONTAL<VERTICAL<LOCAL) and then lexicographically byDimension.value.Reason:
An as_fieldop call as emitted by the frontend has no domain, and inferring the type of the domain was not possible, since no global ordering relation of dimensions existed, e.g. for an as_fieldop operating on a
Vertexand aKfield it was unclear if the dimensions of the output were(Vertex, K)or (K, Vertex), which lead to many negative consequences in other places, that will be tackled in PR 1853 and following ones.