Skip to content

fix(recognition): allow numpy 2.x to unblock Python 3.13 install#1336

Open
mikahanninen wants to merge 1 commit into
masterfrom
fix/recognition-numpy-py313
Open

fix(recognition): allow numpy 2.x to unblock Python 3.13 install#1336
mikahanninen wants to merge 1 commit into
masterfrom
fix/recognition-numpy-py313

Conversation

@mikahanninen

Copy link
Copy Markdown
Collaborator

Summary

rpaframework-recognition pinned numpy >=1.19.3, <2.0.0, which forces numpy 1.26.4 — the last 1.x release. numpy 1.26.4 has no cp313 wheel, so on Python 3.13 pip falls back to a source build that fails without a C toolchain (e.g. VS Build Tools on Windows). This makes the package uninstallable on 3.13.

opencv-python-headless (4.10.x, already in our range) supports numpy 2, so dropping the upper bound resolves cleanly.

Changes

  • packages/recognition/pyproject.toml: numpy >=1.19.3, <2.0.0numpy >=1.19.3; add Programming Language :: Python :: 3.13 classifier.
  • packages/recognition/uv.lock: relocked — numpy now forks to 2.2.6 (py<3.11) / 2.4.6 (py>=3.11), with cp313 wheels present.

Verification

uv pip compile on Python 3.13 now resolves:

numpy==2.4.6
opencv-python-headless==4.10.0.84
pillow==12.2.0

No conflicts.

Fixes #1331

🤖 Generated with Claude Code

rpaframework-recognition pinned numpy>=1.19.3,<2.0.0, forcing numpy
1.26.4 which has no cp313 wheel. On Python 3.13 pip then attempts a
source build that fails without a C toolchain (e.g. VS Build Tools on
Windows). opencv-python-headless 4.10.x already supports numpy 2, so
drop the upper bound and add the 3.13 classifier.

Fixes #1331

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

rpaframework-recognition: numpy<2.0.0 constraint breaks installation on Python 3.13

1 participant