Allow numpy 2.3+ for Python 3.14 compatibility #409
Merged
Conversation
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
NumPy 2.1.x has a bug on Python 3.14 where "temporary elision" optimization incorrectly identifies arrays as temporary due to the new LOAD_FAST_BORROW bytecode instruction, causing destructive in-place modifications. This caused dividend_income_tax to return £0 instead of £22bn in UK microsimulations. Fix: numpy/numpy#28748 (released in NumPy 2.3.0) Issue: numpy/numpy#28681 Closes PolicyEngine#407 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #409 +/- ##
==========================================
+ Coverage 81.14% 81.75% +0.61%
==========================================
Files 197 200 +3
Lines 10228 10438 +210
Branches 1057 1049 -8
==========================================
+ Hits 8299 8534 +235
- Misses 1640 1643 +3
+ Partials 289 261 -28 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Collaborator
|
@MaxGhenis I just want to run a few things with it |
Closed
baogorek
approved these changes
Dec 2, 2025
Collaborator
baogorek
left a comment
baogorek
left a comment
There was a problem hiding this comment.
The things I wanted to test are running fine.
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Outdated suggestions cannot be applied.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.
Summary
~=2.1.0to>=2.1.0,<3to allow NumPy 2.3+Problem
NumPy 2.1.x has a bug on Python 3.14 where the "temporary elision" optimization incorrectly identifies arrays as temporary due to the new
LOAD_FAST_BORROWbytecode instruction, causing destructive in-place modifications during arithmetic operations.This caused
dividend_income_taxto return £0 instead of £22.55bn in UK microsimulations.Root Cause
When formulas execute operations like:
NumPy's temporary elision optimization checks
refcount==1to decide if it can do in-place ops. Python 3.14'sLOAD_FAST_BORROWbytecode creates temporary references that don't increment refcount, breaking this heuristic.Fix
Test plan
dividend_income_taxcorrectly returns £22.55bnCloses #407
🤖 Generated with Claude Code