pip dependency conflict (numpy) in the test pipeline #5164
Description
First Check
- I added a very descriptive title here.
- This is not a Q&A. I am sure something is wrong with NiceGUI or its documentation.
- I used the GitHub search to find a similar issue and came up empty.
Example Code
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
contourpy 1.1.1 requires numpy<2.0,>=1.16; python_version <= "3.11", but you have numpy 2.0.2 which is incompatible.
matplotlib 3.7.5 requires numpy<2,>=1.20, but you have numpy 2.0.2 which is incompatible.Description
Check https://github.com/evnchn/nicegui/actions/runs/17738993290/job/50407952857.
As it turns out, opencv-python requires numpy<2.3.0,>=2, and everything falls apart afterwards.
NiceGUI Version
zauberzeug/main
Python Version
3.9 - 3.12
Browser
Other
Operating System
Linux
Additional Context
@falkoschindler Do NOT clear your pipeline's cache. If it is cleared then this issue would affect main's pipeline as well.
Immediate course of action I recommend to run tests no longer than 7 days apart to ensure our cache is not evicted (https://docs.github.com/en/actions/reference/workflows-and-actions/dependency-caching#usage-limits-and-eviction-policy).
Long term action we need to think about how to better manage the dependencies used for tests. At least, bare poetry run pip does not well-account for the possibility that updated dependencies, such as opencv-python, conflict with poetry.lock stuff.