MNT allow changing the linter in CI

.github/workflows/lint_benchmarks.yml

@@ -2,9 +2,14 @@ name: Lint
 
 on:
   workflow_call:
+    inputs:
+      linter:
+        type: string
+        description: Linter to use for the benchmarks. For now only supports flake8 and ruff.
+        default: flake8
 
 jobs:
-  linter-flake8:
+  linter:
     runs-on: ubuntu-latest
 
     # Cancel in-progress workflows when pushing
@@ -20,7 +25,11 @@ jobs:
       with:
         python-version: 3.8
 
-    - name: Lint with flake8
+    - name: Lint with ${{ inputs.linter }}
       run: |
-        pip install flake8
-        flake8 .
+        pip install ${{ inputs.linter }}
+        if [ "${{ inputs.linter }}" = "ruff" ]; then
+          ruff check .
+        else
+          flake8 .
+        fi

.github/workflows/test_benchmarks.yml

@@ -17,7 +17,7 @@ on:
         type: string
       python_version:
         description: Python version to use for the tests.
-        default: "3.10"
+        default: "3.12"
         required: false
         type: string
       extra_args:
@@ -89,7 +89,7 @@ jobs:
         then
           user=${BENCHOPT_BRANCH%@*}
           branch=${BENCHOPT_BRANCH##*@}
-          pip install -U git+https://github.com/$user/benchopt@$branch
+          pip install -U "benchopt @ git+https://github.com/$user/benchopt@$branch"
         elif [[ "$BENCHOPT_VERSION" == "latest" ]]
         then
           pip install -U benchopt

datasets/simulated.py

@@ -17,7 +17,6 @@ class Dataset(BaseDataset):
             (1000, 500),
             (5000, 200),
         ],
-        'random_state': [27],
     }
 
     # List of packages needed to run the dataset. See the corresponding
@@ -29,8 +28,16 @@ def get_data(self):
         # to `Objective.set_data`. This defines the benchmark's
         # API to pass data. It is customizable for each benchmark.
 
+        # Get a random seed to generate the data. The seed is generated from
+        # the `get_seed` method, which ensures that the same seed is used
+        # across different runs of the benchmark, and different solvers.
+        # The use of `use_repetition=True` ensures that the seed changes across
+        # different repetitions of the benchmark, which can be useful to
+        # generate different data for each repetition.
+        seed = self.get_seed(use_repetition=True)
+
         # Generate pseudorandom data using `numpy`.
-        rng = np.random.RandomState(self.random_state)
+        rng = np.random.RandomState(seed)
         X = rng.randn(self.n_samples, self.n_features)
         y = rng.randn(self.n_samples)