Installing package that requires numpy>=2 is incompatible with Evidently Integration.

[JustGitting]

Contact Details [Optional]

No response

System Information

(.venv) PS C:\...\> zenml info -a -s
Daemon functionality is currently not supported on Windows.
ZENML_LOCAL_VERSION: 0.91.1
ZENML_SERVER_VERSION: 0.91.1
ZENML_SERVER_DATABASE: mysql
ZENML_SERVER_DEPLOYMENT_TYPE: other
ZENML_CONFIG_DIR: C:\Users\user1\AppData\Roaming\zenml
ZENML_LOCAL_STORE_DIR: /mnt/c/Users/user1/AppData/Roaming/zenml/zenml
ZENML_SERVER_URL: http://192.168.1.150:8080
ZENML_ACTIVE_REPOSITORY_ROOT: C:\Users\user1\Documents\Projects\test
PYTHON_VERSION: 3.11.2
ENVIRONMENT: native
SYSTEM_INFO: {'os': 'windows', 'windows_version_release': '10', 'windows_version': '10.0.22631', 'windows_version_service_pack': 'SP0', 'windows_version_os_type': 'Multiprocessor Free'}

What happened?

I want to install the feature selection and hyper-parameter optimisation package Shap-Hypetune package.

https://github.com/cerlymarco/shap-hypetune

However, shap-hypetune requires numpy>=2 and the ZenML Evidently integration only supports numpy<2.0.0. Hence, I'm unable to run both.

> zenml integration list
Daemon functionality is currently not supported on Windows.
╭───────────┬─────────────────────┬──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ INSTALLED │ INTEGRATION         │ REQUIRED_PACKAGES                                                                                                                                                                                                    │
├───────────┼─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┤
...
├───────────┼─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┤
│    ✅     │ evidently           │ evidently>=0.4.16,<=0.4.22, tenacity!=8.4.0, numpy<2.0.0, pandas>=2.0.0                                                                                                                                              │
├───────────┼─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┤
...

Reproduction steps

Install evidently integration.

zenml integration install evidently -y

Install shap-hypetune which uninstalls zeneml evidently integration (or deactivates it?)

pip install --upgrade shap-hypetune

zenml integration list
Daemon functionality is currently not supported on Windows.
╭───────────┬─────────────────────┬──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ INSTALLED │ INTEGRATION         │ REQUIRED_PACKAGES                                                                                                                                                                                                    │
├───────────┼─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┤
...
├───────────┼─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┤
│    ❌     │ evidently           │ evidently>=0.4.16,<=0.4.22, tenacity!=8.4.0, numpy<2.0.0, pandas>=2.0.0                                                                                                                                              │
├───────────┼─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┤

...

Relevant log output

The first sign of the problem was when I ran my pipeline and got the following error because evidently <0.4.22 uses numpy < 2.0.
And the shap package insatlled numpy > 2.0.

│ C:\Users\user1\Documents\Projects\test1\.venv\Lib\site-packages\evidently\core.py:96 in        │
│ <module>                                                                                         │
│                                                                                                  │
│    93 #     return pd.Index(value)                                                               │
│    94                                                                                            │
│    95                                                                                            │
│ ❱  96 @pydantic_type_validator(np.float_)                                                        │
│    97 def np_inf_valudator(value):                                                               │
│    98 │   return np.float(value)                                                                 │
│    99                                                                                            │
│                                                                                                  │
│ C:\Users\uesr1\Documents\Projects\test1\.venv\Lib\site-packages\numpy\__init__.py:794 in       │
│ __getattr__                                                                                      │
│                                                                                                  │
│   791 │   │   │   raise AttributeError(__former_attrs__[attr], name=None)                        │
│   792 │   │                                                                                      │
│   793 │   │   if attr in __expired_attributes__:                                                 │
│ ❱ 794 │   │   │   raise AttributeError(                                                          │
│   795 │   │   │   │   f"`np.{attr}` was removed in the NumPy 2.0 release. "                      │
│   796 │   │   │   │   f"{__expired_attributes__[attr]}",                                         │
│   797 │   │   │   │   name=None                                                                  │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
AttributeError: `np.float_` was removed in the NumPy 2.0 release. Use `np.float64` instead.

Code of Conduct

• I agree to follow this project's Code of Conduct

[htahir1]

Good point we should upgrade evidently.. This should unblock you for now @JustGitting #4243