[pre-commit.ci] pre-commit autoupdate (#555)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Tim Mensinger <[email protected]>

.pre-commit-config.yaml

@@ -80,7 +80,7 @@ repos:
           - --blank
         exclude: src/optimagic/optimization/algo_options.py
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.7.2
+    rev: v0.9.2
     hooks:
       # Run the linter.
       - id: ruff
@@ -97,7 +97,7 @@ repos:
           - pyi
           - jupyter
   - repo: https://github.com/executablebooks/mdformat
-    rev: 0.7.18
+    rev: 0.7.21
     hooks:
       - id: mdformat
         additional_dependencies:
@@ -109,7 +109,7 @@ repos:
           - '88'
         files: (README\.md)
   - repo: https://github.com/executablebooks/mdformat
-    rev: 0.7.18
+    rev: 0.7.21
     hooks:
       - id: mdformat
         additional_dependencies:
@@ -121,7 +121,7 @@ repos:
         files: (docs/.)
         exclude: docs/source/how_to/how_to_specify_algorithm_and_algo_options.md
   - repo: https://github.com/kynan/nbstripout
-    rev: 0.8.0
+    rev: 0.8.1
     hooks:
       - id: nbstripout
         exclude: |
@@ -132,7 +132,7 @@ repos:
         args:
           - --drop-empty-cells
   - repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v1.13.0
+    rev: v1.14.1
     hooks:
       - id: mypy
         files: src|tests

.tools/envs/testenv-linux.yml

@@ -23,7 +23,7 @@ dependencies:
   - scipy>=1.2.1  # run, tests
   - sqlalchemy  # run, tests
   - seaborn  # dev, tests
-  - mypy=1.13  # dev, tests
+  - mypy=1.14.1  # dev, tests
   - pyyaml  # dev, tests
   - jinja2  # dev, tests
   - annotated-types  # dev, tests

.tools/envs/testenv-numpy.yml

@@ -21,7 +21,7 @@ dependencies:
   - scipy>=1.2.1  # run, tests
   - sqlalchemy  # run, tests
   - seaborn  # dev, tests
-  - mypy=1.13  # dev, tests
+  - mypy=1.14.1  # dev, tests
   - pyyaml  # dev, tests
   - jinja2  # dev, tests
   - annotated-types  # dev, tests

.tools/envs/testenv-others.yml

@@ -21,7 +21,7 @@ dependencies:
   - scipy>=1.2.1  # run, tests
   - sqlalchemy  # run, tests
   - seaborn  # dev, tests
-  - mypy=1.13  # dev, tests
+  - mypy=1.14.1  # dev, tests
   - pyyaml  # dev, tests
   - jinja2  # dev, tests
   - annotated-types  # dev, tests

.tools/envs/testenv-pandas.yml

@@ -21,7 +21,7 @@ dependencies:
   - scipy>=1.2.1  # run, tests
   - sqlalchemy  # run, tests
   - seaborn  # dev, tests
-  - mypy=1.13  # dev, tests
+  - mypy=1.14.1  # dev, tests
   - pyyaml  # dev, tests
   - jinja2  # dev, tests
   - annotated-types  # dev, tests

docs/source/development/ep-01-pytrees.md

@@ -384,7 +384,7 @@ much complexity by avoiding complex pytrees as inputs and outputs at the same ti
 To see this in action, let's look at an example. We repeat the example from the JAX
 interface above with the following changes:
 
-1. The 1d numpy array in x\["a"\] is replaced by a DataFrame with `"value"` column
+1. The 1d numpy array in x["a"] is replaced by a DataFrame with `"value"` column
 1. The "d" entry in the output becomes a Series instead of a 1d numpy array.
 
 ```python
@@ -460,7 +460,7 @@ very first jacobian:
 +--------+----------+----------+----------+----------+----------+----------+----------+
 ```
 
-The indices \["j", "k", "l", "m"\] unfortunately never made it into the result because
+The indices ["j", "k", "l", "m"] unfortunately never made it into the result because
 they were only applied to elements that already came from a 2d array and thus always
 have a 3d Jacobian, i.e. the result entry `["c"][b"]` is a reshaped version of the upper
 right 2 by 4 array and the result entry `["d"]["b"]` is a reshaped version of the lower

docs/source/how_to/how_to_scaling.md

@@ -90,7 +90,7 @@ these considerations are typically tighter than for parameters that have a small
 on the objective function.
 
 Thus, a natural approach to improve the scaling of the optimization problem is to re-map
-all parameters such that the bounds are \[0, 1\] for all parameters. This has the
+all parameters such that the bounds are [0, 1] for all parameters. This has the
 additional advantage that absolute and relative convergence criteria on parameter
 changes become the same.
 
@@ -164,7 +164,7 @@ to a strictly non-zero number for the `"start_values"` and `"gradient"` approach
 `"bounds"` approach avoids division by exact zeros by construction. The
 `"clipping_value"` can still be used to avoid extreme upscaling of parameters with very
 tight bounds. However, this means that the bounds of the re-scaled problem are not
-exactly \[0, 1\] for all parameters.
+exactly [0, 1] for all parameters.
 
 (scaling-default-values)=
 

environment.yml

@@ -31,7 +31,7 @@ dependencies:
   - sphinx-panels  # docs
   - sphinxcontrib-bibtex  # docs
   - seaborn  # dev, tests
-  - mypy=1.13  # dev, tests
+  - mypy=1.14.1  # dev, tests
   - pyyaml  # dev, tests
   - jinja2  # dev, tests
   - furo  # dev, docs

src/optimagic/differentiation/derivatives.py

@@ -328,7 +328,7 @@ def first_derivative(
         f0 = np.array(f0, dtype=np.float64)
 
     # convert the raw evaluations to numpy arrays
-    raw_evals = _convert_evals_to_numpy(
+    raw_evals_arr = _convert_evals_to_numpy(
         raw_evals=raw_evals,
         unpacker=unpacker,
         registry=registry,
@@ -337,9 +337,9 @@ def first_derivative(
     )
 
     # apply finite difference formulae
-    evals_data = np.array(raw_evals).reshape(2, n_steps, len(x), -1)
-    evals_data = np.transpose(evals_data, axes=(0, 1, 3, 2))
-    evals = Evals(pos=evals_data[0], neg=evals_data[1])
+    evals_data = np.array(raw_evals_arr).reshape(2, n_steps, len(x), -1)
+    evals_data_transposed = np.transpose(evals_data, axes=(0, 1, 3, 2))
+    evals = Evals(pos=evals_data_transposed[0], neg=evals_data_transposed[1])
 
     jac_candidates = {}
     for m in ["forward", "backward", "central"]:

src/optimagic/differentiation/generate_steps.py

@@ -90,9 +90,9 @@ def generate_steps(
     )
     min_steps = base_steps if min_steps is None else min_steps
 
-    assert (
-        bounds.upper - bounds.lower >= 2 * min_steps
-    ).all(), "min_steps is too large to fit into bounds."
+    assert (bounds.upper - bounds.lower >= 2 * min_steps).all(), (
+        "min_steps is too large to fit into bounds."
+    )
 
     upper_step_bounds = bounds.upper - x
     lower_step_bounds = bounds.lower - x

src/optimagic/differentiation/richardson_extrapolation.py

@@ -52,13 +52,13 @@ def richardson_extrapolation(sequence, steps, method="central", num_terms=None):
     n_steps = steps.shape[0]
     num_terms = n_steps if num_terms is None else num_terms
 
-    assert (
-        seq_len == n_steps
-    ), "Length of ``steps`` must coincide with length of ``sequence``."
+    assert seq_len == n_steps, (
+        "Length of ``steps`` must coincide with length of ``sequence``."
+    )
     assert num_terms > 0, "``num_terms`` must be greater than zero."
-    assert (
-        seq_len - 1 >= num_terms
-    ), "``num_terms`` cannot be greater than ``seq_len`` - 1."
+    assert seq_len - 1 >= num_terms, (
+        "``num_terms`` cannot be greater than ``seq_len`` - 1."
+    )
 
     step_ratio = _compute_step_ratio(steps)
     order, exponentiation_step = _get_order_and_exponentiation_step(method)

src/optimagic/examples/criterion_functions.py

@@ -114,7 +114,7 @@ def rosenbrock_gradient(params: PyTree) -> PyTree:
     l4 = np.delete(x, [0])
     l4 = np.append(l4, 0)
     l5 = np.full((len(x) - 1), 2)
-    l5 = np.append(l5, 0)
+    l5 = np.append(l5, 0)  # type: ignore[assignment]
     flat = 100 * (4 * (l1**3) + 2 * l2 - 2 * (l3**2) - 4 * (l4 * x)) + 2 * l1 - l5
     return _unflatten_gradient(flat, params)
 

src/optimagic/logging/logger.py

@@ -187,7 +187,10 @@ def _build_history_dataframe(self) -> pd.DataFrame:
 
         times = np.array(history["time"])
         times -= times[0]
-        history["time"] = times.tolist()
+        # For numpy arrays with ndim = 0, tolist() returns a scalar, which violates the
+        # type hinting list[Any] from above. As history["time"] is always a list, this
+        # case is safe to ignore.
+        history["time"] = times.tolist()  # type: ignore[assignment]
 
         df = pd.DataFrame(history)
         df = df.merge(

src/optimagic/optimization/multistart.py

@@ -313,9 +313,11 @@ def run_explorations(
 
     """
     internal_problem = internal_problem.with_step_id(step_id)
-    x_list = list(sample)
+    x_list: list[NDArray[np.float64]] = list(sample)
 
-    raw_values = np.array(internal_problem.exploration_fun(x_list, n_cores=n_cores))
+    raw_values = np.asarray(
+        internal_problem.exploration_fun(x_list, n_cores=n_cores), dtype=np.float64
+    )
 
     is_valid = np.isfinite(raw_values)
 

src/optimagic/optimization/optimize_result.py

@@ -173,8 +173,7 @@ def __repr__(self) -> str:
 
         if self.start_fun is not None and self.fun is not None:
             improvement = (
-                f"The value of criterion improved from {self.start_fun} to "
-                f"{self.fun}."
+                f"The value of criterion improved from {self.start_fun} to {self.fun}."
             )
         else:
             improvement = None

src/optimagic/optimizers/nag_optimizers.py

@@ -1051,8 +1051,9 @@ def _build_options_dict(user_input, default_options):
     invalid = [x for x in user_input if x not in full_options]
     if len(invalid) > 0:
         raise ValueError(
-            f"You specified illegal options {', '.join(invalid)}. Allowed are: "
-            ", ".join(full_options.keys())
+            f"You specified illegal options {', '.join(invalid)}. Allowed are: , ".join(
+                full_options.keys()
+            )
         )
     full_options.update(user_input)
     return full_options

src/optimagic/optimizers/scipy_optimizers.py

@@ -799,11 +799,7 @@ def _solve_internal_problem(
             )
         raw_res = scipy.optimize.brute(
             func=problem.fun,
-            ranges=tuple(
-                map(
-                    tuple, np.column_stack((problem.bounds.lower, problem.bounds.upper))
-                )
-            ),
+            ranges=tuple(zip(problem.bounds.lower, problem.bounds.upper, strict=True)),
             Ns=self.n_grid_points,
             full_output=True,
             finish=self.polishing_function,

src/optimagic/parameters/consolidate_constraints.py

@@ -205,9 +205,9 @@ def _consolidate_fixes_with_equality_constraints(
     for eq in equality_constraints:
         if np.isfinite(fixed_value[eq["index"]]).any():
             valcounts = _unique_values(fixed_value[eq["index"]])
-            assert (
-                len(valcounts) == 1
-            ), "Equality constrained parameters cannot be fixed to different values."
+            assert len(valcounts) == 1, (
+                "Equality constrained parameters cannot be fixed to different values."
+            )
             fixed_value[eq["index"]] = valcounts[0]
 
     return fixed_value

tests/estimagic/test_bootstrap.py

@@ -123,16 +123,16 @@ def test_bootstrap_existing_outcomes(setup):
     result = bootstrap(
         data=setup["df"],
         outcome=_outcome_func,
-        n_draws=2,
+        n_draws=3,
     )
-    assert len(result.outcomes) == 2
+    assert len(result.outcomes) == 3
     result = bootstrap(
         outcome=_outcome_func,
         data=setup["df"],
         existing_result=result,
-        n_draws=1,
+        n_draws=2,
     )
-    assert len(result.outcomes) == 1
+    assert len(result.outcomes) == 2
 
 
 def test_bootstrap_from_outcomes(setup, expected):

tests/estimagic/test_estimation_table.py

@@ -395,7 +395,7 @@ def test_get_model_names():
 def test_get_default_column_names_and_groups():
     model_names = ["a_name", "a_name", "(3)", "(4)", "third_name"]
     res_names, res_groups = _get_default_column_names_and_groups(model_names)
-    exp_names = [f"({i+1})" for i in range(len(model_names))]
+    exp_names = [f"({i + 1})" for i in range(len(model_names))]
     exp_groups = ["a_name", "a_name", "(3)", "(4)", "third_name"]
     assert res_names == exp_names
     assert res_groups == exp_groups

tests/optimagic/optimization/test_with_constraints.py

@@ -244,7 +244,10 @@ def test_three_independent_constraints():
     )
     expected = np.array([0] * 4 + [4, 5] + [0] + [7.5] * 2 + [0])
 
-    aaae(res.params, expected, decimal=4)
+    # TODO: Increase precision back to decimal=4. The reduced precision is likely due
+    # to the re-written L-BFGS-B algorithm in SciPy 1.15.
+    # See https://github.com/optimagic-dev/optimagic/issues/556.
+    aaae(res.params, expected, decimal=3)
 
 
 INVALID_CONSTRAINT_COMBIS = [