feat(ir): experimental egraph integration #4478
Closed
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
BREAKING CHANGE: use other methods for pinging the database
BREAKING CHANGE: functions, methods and classes marked as deprecated are removed now
BREAKING CHANGE: removed Node.root_tables() method, use ibis.expr.analysis.find_immediate_parent_tables() instead
The DestructValue expressions didn't have corresponding operation nodes. BREAKING CHANGE: removed ir.DestructValue, ir.DestructScalar and ir.DestructColumn, use table.unpack() instead
BREAKING CHANGE: ibis.expr.lineage.lineage() is now removed
also make the Signature object validatable
BREAKING CHANGE: use ibis.common.validators.{Patameter, Signature} instead
Unify the result of pandas/dask with the rest of the backends BREAKING CHANGE: Use timezone conversion functions to compute the original machine localized value
The objective of this refactor is to achieve a simpler, stricter and more generic internal representation by removing intermediate Expression objects so there is a single Expression object wrapping an Operation tree only to provide the user-facing API but all the internals work with homogeneous Operation trees. BREAKING CHANGE: Expr(Op(Expr(Op(Expr(Op))))) is now represented as Expr(Op(Op(Op))), so code using ibis internals must be migrated
…imported for types BREAKING CHANGE: Removed Node.output_type in favor of abstractmethod Node.to_expr() which now must be explicitly implemented
…class for ops.Selection BREAKING CHANGE: HasSchema is no longer available, directly subclass ops.TableNode and implement schema property instead
…dling BREAKING CHANGE: Node.blocks() method has been removed.
… of the pandas backend BREAKING CHANGE: removed ops.Node.inputs property, use the multipledispatched get_node_arguments() function in the pandas backend
previously node attributes contained tuple of other nodes which required additional isinstance checks for traversals, now all the traversable elements of the tree are node instances BREAKING CHANGE: removed Node.flat_args(), directly use node.args property instead
…aming semantics BREAKING CHANGE: node.has_resolved_name() is removed, use isinstance(node, ops.Named) instead; node.resolve_name() is removed use node.name instead
Codecov Report
@@ Coverage Diff @@
## 4.x.x #4478 +/- ##
==========================================
- Coverage 92.79% 89.99% -2.81%
==========================================
Files 180 180
Lines 20088 19743 -345
Branches 2853 2907 +54
==========================================
- Hits 18640 17767 -873
- Misses 1090 1540 +450
- Partials 358 436 +78
|
kszucs
force-pushed
the
4.x.x
branch
3 times, most recently
from
e0dd44f to
5054275
Compare
|
@kszucs Should we close the matchpy PR in favor of this one? |
|
Well, I'm not sure which one should be the preferred one. Going to rebase them for now. |
|
Closing this in favor of #5781 |
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.
egraph/egg is an alternative term rewriting system to matchpy thus this integration provides an alternative to #4437
It uses a refactored version of
snake-eggpython bindings from my fork