style: pin pre commit and ruff to recent versions. (#1358)

* update and pin ruff and pre-commit.

* linting and formatting with new ruff

* apply ruff to tests and tutorials, ignore long lines.

Author: janfb

.pre-commit-config.yaml

@@ -1,12 +1,12 @@
 repos:
 - repo: https://github.com/astral-sh/ruff-pre-commit
-  rev: v0.3.3
+  rev: v0.9.0
   hooks:
     - id: ruff
     - id: ruff-format
       args: [--diff]
 - repo: https://github.com/pre-commit/pre-commit-hooks
-  rev: v4.5.0
+  rev: v5.0.0
   hooks:
     - id: check-added-large-files
     - id: check-merge-conflict

pyproject.toml

@@ -64,10 +64,10 @@ doc = [
 dev = [
     "ffmpeg",
     # Lint
-    "pre-commit == 3.5.0",
+    "pre-commit == 4.0.1",
     "pyyaml",
     "pyright",
-    "ruff>=0.3.3",
+    "ruff==0.9.0",
     # Test
     "pytest",
     "pytest-cov",
@@ -106,6 +106,7 @@ ignore = [
 [tool.ruff.lint.extend-per-file-ignores]
 "__init__.py" = ["E402", "F401", "F403"]  # allow unused imports and undefined names
 "test_*.py" = ["F403", "F405"]
+"tutorials/*.ipynb" = ["E501"]  # allow long lines in notebooks
 
 [tool.ruff.lint.isort]
 case-sensitive = true

sbi/analysis/plot.py

@@ -775,9 +775,9 @@ def pairplot(
 
     # checks.
     if fig_kwargs_filled["legend"]:
-        assert len(fig_kwargs_filled["samples_labels"]) >= len(
-            samples
-        ), "Provide at least as many labels as samples."
+        assert len(fig_kwargs_filled["samples_labels"]) >= len(samples), (
+            "Provide at least as many labels as samples."
+        )
     if offdiag is not None:
         warn("offdiag is deprecated, use upper or lower instead.", stacklevel=2)
         upper = offdiag
@@ -1594,9 +1594,9 @@ def _sbc_rank_plot(
             ranks_list[idx]: np.ndarray = rank.numpy()  # type: ignore
 
     plot_types = ["hist", "cdf"]
-    assert (
-        plot_type in plot_types
-    ), "plot type {plot_type} not implemented, use one in {plot_types}."
+    assert plot_type in plot_types, (
+        "plot type {plot_type} not implemented, use one in {plot_types}."
+    )
 
     if legend_kwargs is None:
         legend_kwargs = dict(loc="best", handlelength=0.8)
@@ -1609,9 +1609,9 @@ def _sbc_rank_plot(
         params_in_subplots = True
 
     for ranki in ranks_list:
-        assert (
-            ranki.shape == ranks_list[0].shape
-        ), "all ranks in list must have the same shape."
+        assert ranki.shape == ranks_list[0].shape, (
+            "all ranks in list must have the same shape."
+        )
 
     num_rows = int(np.ceil(num_parameters / num_cols))
     if figsize is None:
@@ -1636,9 +1636,9 @@ def _sbc_rank_plot(
             )
             ax = np.atleast_1d(ax)  # type: ignore
         else:
-            assert (
-                ax.size >= num_parameters
-            ), "There must be at least as many subplots as parameters."
+            assert ax.size >= num_parameters, (
+                "There must be at least as many subplots as parameters."
+            )
             num_rows = ax.shape[0] if ax.ndim > 1 else 1
         assert ax is not None
 
@@ -2221,9 +2221,9 @@ def pairplot_dep(
 
     # checks.
     if opts["legend"]:
-        assert len(opts["samples_labels"]) >= len(
-            samples
-        ), "Provide at least as many labels as samples."
+        assert len(opts["samples_labels"]) >= len(samples), (
+            "Provide at least as many labels as samples."
+        )
     if opts["upper"] is not None:
         opts["offdiag"] = opts["upper"]
 

sbi/analysis/sensitivity_analysis.py

@@ -250,9 +250,9 @@ def train(
                 prevent exploding gradients. Use `None` for no clipping.
         """
 
-        assert (
-            self._theta is not None and self._emergent_property is not None
-        ), "You must call .add_property() first."
+        assert self._theta is not None and self._emergent_property is not None, (
+            "You must call .add_property() first."
+        )
 
         # Get indices for permutation of the data.
         num_examples = len(self._theta)
@@ -433,9 +433,9 @@ def find_directions(
         if posterior_log_prob_as_property:
             predictions = self._posterior.potential(thetas, track_gradients=True)
         else:
-            assert (
-                self._regression_net is not None
-            ), "self._regression_net is None, you must call `.train()` first."
+            assert self._regression_net is not None, (
+                "self._regression_net is None, you must call `.train()` first."
+            )
             predictions = self._regression_net.forward(thetas)
         loss = predictions.mean()
         loss.backward()

sbi/diagnostics/lc2st.py

@@ -83,9 +83,9 @@ def __init__(
         [2] : https://github.com/sbi-dev/sbi/blob/main/sbi/utils/metrics.py
         """
 
-        assert (
-            thetas.shape[0] == xs.shape[0] == posterior_samples.shape[0]
-        ), "Number of samples must match"
+        assert thetas.shape[0] == xs.shape[0] == posterior_samples.shape[0], (
+            "Number of samples must match"
+        )
 
         # set observed data for classification
         self.theta_p = posterior_samples
@@ -283,9 +283,9 @@ def get_statistic_on_observed_data(
         Returns:
             L-C2ST statistic at `x_o`.
         """
-        assert (
-            self.trained_clfs is not None
-        ), "No trained classifiers found. Run `train_on_observed_data` first."
+        assert self.trained_clfs is not None, (
+            "No trained classifiers found. Run `train_on_observed_data` first."
+        )
         _, scores = self.get_scores(
             theta_o=theta_o,
             x_o=x_o,
@@ -372,9 +372,9 @@ def train_under_null_hypothesis(
                     joint_q_perm[:, self.theta_q.shape[1] :],
                 )
             else:
-                assert (
-                    self.null_distribution is not None
-                ), "You need to provide a null distribution"
+                assert self.null_distribution is not None, (
+                    "You need to provide a null distribution"
+                )
                 theta_p_t = self.null_distribution.sample((self.theta_p.shape[0],))
                 theta_q_t = self.null_distribution.sample((self.theta_p.shape[0],))
                 x_p_t, x_q_t = self.x_p, self.x_q
@@ -419,9 +419,9 @@ def get_statistics_under_null_hypothesis(
                     Run `train_under_null_hypothesis`."
             )
         else:
-            assert (
-                len(self.trained_clfs_null) == self.num_trials_null
-            ), "You need one classifier per trial."
+            assert len(self.trained_clfs_null) == self.num_trials_null, (
+                "You need one classifier per trial."
+            )
 
         probs_null, stats_null = [], []
         for t in tqdm(
@@ -433,9 +433,9 @@ def get_statistics_under_null_hypothesis(
             if self.permutation:
                 theta_o_t = theta_o
             else:
-                assert (
-                    self.null_distribution is not None
-                ), "You need to provide a null distribution"
+                assert self.null_distribution is not None, (
+                    "You need to provide a null distribution"
+                )
 
                 theta_o_t = self.null_distribution.sample((theta_o.shape[0],))
 

sbi/diagnostics/sbc.py

@@ -69,9 +69,9 @@ def run_sbc(
             stacklevel=2,
         )
 
-    assert (
-        thetas.shape[0] == xs.shape[0]
-    ), "Unequal number of parameters and observations."
+    assert thetas.shape[0] == xs.shape[0], (
+        "Unequal number of parameters and observations."
+    )
 
     if "sbc_batch_size" in kwargs:
         warnings.warn(

sbi/diagnostics/tarp.py

@@ -133,9 +133,9 @@ def _run_tarp(
     """
     num_posterior_samples, num_tarp_samples, _ = posterior_samples.shape
 
-    assert (
-        references.shape == thetas.shape
-    ), "references must have the same shape as thetas"
+    assert references.shape == thetas.shape, (
+        "references must have the same shape as thetas"
+    )
 
     if num_bins is None:
         num_bins = num_tarp_samples // 10

sbi/inference/abc/mcabc.py

@@ -130,9 +130,9 @@ def __call__(
         """
 
         # Exactly one of eps or quantile need to be passed.
-        assert (eps is not None) ^ (
-            quantile is not None
-        ), "Eps or quantile must be passed, but not both."
+        assert (eps is not None) ^ (quantile is not None), (
+            "Eps or quantile must be passed, but not both."
+        )
         if kde_kwargs is None:
             kde_kwargs = {}
 

sbi/inference/abc/smcabc.py

@@ -95,9 +95,9 @@ def __init__(
         )
 
         kernels = ("gaussian", "uniform")
-        assert (
-            kernel in kernels
-        ), f"Kernel '{kernel}' not supported. Choose one from {kernels}."
+        assert kernel in kernels, (
+            f"Kernel '{kernel}' not supported. Choose one from {kernels}."
+        )
         self.kernel = kernel
 
         algorithm_variants = ("A", "B", "C")
@@ -198,13 +198,13 @@ def __call__(
         if kde_kwargs is None:
             kde_kwargs = {}
         assert isinstance(epsilon_decay, float) and epsilon_decay > 0.0
-        assert not (
-            self.distance.requires_iid_data and lra
-        ), "Currently there is no support to run inference "
+        assert not (self.distance.requires_iid_data and lra), (
+            "Currently there is no support to run inference "
+        )
         "on multiple observations together with lra."
-        assert not (
-            self.distance.requires_iid_data and sass
-        ), "Currently there is no support to run inference "
+        assert not (self.distance.requires_iid_data and sass), (
+            "Currently there is no support to run inference "
+        )
         "on multiple observations together with sass."
 
         # Pilot run for SASS.
@@ -363,9 +363,9 @@ def _set_xo_and_sample_initial_population(
     ) -> Tuple[Tensor, float, Tensor, Tensor]:
         """Return particles, epsilon and distances of initial population."""
 
-        assert (
-            num_particles <= num_initial_pop
-        ), "number of initial round simulations must be greater than population size"
+        assert num_particles <= num_initial_pop, (
+            "number of initial round simulations must be greater than population size"
+        )
 
         assert (x_o.shape[0] == 1) or self.distance.requires_iid_data, (
             "Your data contain iid data-points, but the choice of "

sbi/inference/posteriors/base_posterior.py

@@ -288,9 +288,7 @@ def __repr__(self):
         return desc
 
     def __str__(self):
-        desc = (
-            f"Posterior p(θ|x) of type {self.__class__.__name__}. " f"{self._purpose}"
-        )
+        desc = f"Posterior p(θ|x) of type {self.__class__.__name__}. {self._purpose}"
         return desc
 
     def __getstate__(self) -> Dict: