Dependency Issue in example - Vertical FL

[TheVidz]

Describe the bug

There is a dependency conflict in the example for 'vertical-fl'. The following error is received on pip install -e .

Steps/Code to Reproduce

After following the steps given in README of vertical-fl:
Start by cloning the example project:

git clone --depth=1 https://github.com/adap/flower.git _tmp \
        && mv _tmp/examples/vertical-fl . \
        && rm -rf _tmp \
        && cd vertical-fl

This will create a new directory called vertical-fl with the following structure: following files:

vertical-fl
├── vertical_fl
│   ├── __init__.py
│   ├── client_app.py   # Defines your ClientApp
│   ├── server_app.py   # Defines your ServerApp
│   ├── strategy.py     # Defines your Strategy
│   └── task.py         # Defines your model, training and data loading
├── pyproject.toml      # Project metadata like dependencies and configs
├── data/train.csv
└── README.md

Install dependencies and project

Install the dependencies defined in pyproject.toml as well as the mlxexample package.

pip install -e .

Expected Results

The result expected is that The required packages are installed.
But here encountered dependency conflict.

Actual Results

The following error is encountered:

INFO: pip is looking at multiple versions of numpy to determine which version is compatible with other requirements. This could take a while.
INFO: pip is looking at multiple versions of vertical-fl to determine which version is compatible with other requirements. This could take a while.
ERROR: Cannot install flwr-datasets, flwr[simulation]==1.13.1, vertical-fl and vertical-fl==1.0.0 because these package versions have conflicting dependencies.

The conflict is caused by:
    vertical-fl 1.0.0 depends on numpy==1.24.4
    pandas 2.0.3 depends on numpy>=1.21.0; python_version >= "3.10"
    scikit-learn 1.3.2 depends on numpy<2.0 and >=1.17.3
    flwr-datasets 0.3.0 depends on numpy<2.0.0 and >=1.21.0
    pyarrow 16.1.0 depends on numpy>=1.16.6
    flwr[simulation] 1.13.1 depends on numpy<3.0.0 and >=1.26.0

To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict

ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

[TheVidz]

Oh I was able to successfully install
If I change the pyproject.toml file:

    "flwr[simulation]>=1.13.1",
    "flwr-datasets>=0.3.0",
    "numpy==1.24.4",
    "pandas==2.0.3",
    "scikit-learn==1.3.2",
    "torch==2.1.0",
]

If I change the numpy dependency to >=1.24.4 instead of ==1.24.4,
This could solve the dependency conflict with flwr[simulation] depending on numpy>=1.26.0

Kindly tell if this is a valid fix?

[jafermarq]

Hi @TheVidz, thanks for opening the issue. This issue arises because until recently flower had to use numpy<2.0 but we removed this limitation a couple of releases ago. I'm inclined to say that if we can remove the numpy specification in the pyproject.toml that's even better.

If you have time, could you open a PR with that fix? I think the same will be needed for the quickstart-pandas example

[TheVidz]

Sure, I will open a PR removing the numpy specification right now, thank you for explaining!

[jafermarq]

Closing, many thanks once more for the fix!