dims.Data creates XTensorSharedVariables#8096
Merged
jessegrabowski merged 3 commits intoFeb 21, 2026
Merged
Conversation
4 tasks
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #8096 +/- ##
=======================================
Coverage 91.49% 91.50%
=======================================
Files 123 123
Lines 19832 19821 -11
=======================================
- Hits 18146 18137 -9
+ Misses 1686 1684 -2
🚀 New features to boost your workflow:
|
ricardoV94
changed the title
ricardoV94
changed the title
ricardoV94
force-pushed
the
handle_xtensor_variables_more_naturally
branch
from
ba31b01 to
abfdd9e
Compare
ricardoV94
changed the title
ricardoV94
force-pushed
the
handle_xtensor_variables_more_naturally
branch
from
abfdd9e to
04e2f24
Compare
ricardoV94
changed the title
ricardoV94
changed the title
dims.Data creates XTensorSharedVariables directly
ricardoV94
changed the title
dims.Data creates XTensorSharedVariables directlydims.Data creates XTensorSharedVariables
Documentation build overview
Show files changed (9 files in total): 📝 9 modified | ➕ 0 added | ➖ 0 deleted
|
jessegrabowski
approved these changes
Feb 21, 2026
Member
jessegrabowski
left a comment
jessegrabowski
left a comment
There was a problem hiding this comment.
approved with some stupid questions and a nitpick about a comment, typical PR review smh
| """ | ||
| model = modelcontext(model) | ||
| var = var.copy(model.name_for(name)) | ||
| var = view_op(var, name=model.name_for(name)) |
Member
There was a problem hiding this comment.
What's the difference? Maybe a better question - why were we copying in the first place?
Member
Author
There was a problem hiding this comment.
You need an op sometimes because you can't have x be both a free rv and a deterministic, but x and view(x) is fine. Copy was another way of achieving this but it's Elemwise or XElemwise Op depending on the type. View instead works with any input type (and achieves the same).
Even variables without a .copy method would work in the future.
The immediate motivation for changing though was to keep pretending xtensorvariables don't exist as much as possible in the rest of the codebase
| coords = { | ||
| "group": range(3), | ||
| "knots": range(N_knots), | ||
| "knot": range(N_knots), |
Member
There was a problem hiding this comment.
singular dims gang rise up!
jessegrabowski
force-pushed
the
handle_xtensor_variables_more_naturally
branch
from
9b06508 to
0a6f365
Compare
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
2 participants
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.
Depends on pymc-devs/pytensor#1882
We were registering regular TensorSharedVariables in
pymc.dims.Data, and then wrapping those inas_xtensor. The problem with that is that when you then try to act onm["x"]you were not acting on the same variable thatpymc.dims.Datareturned to you, instead you were acting on the input ofas_xtensor...This made it hard to adopt it in pymc-marketing