Merge branch 'main' into update-contributor-guide

.pre-commit-config.yaml

@@ -57,7 +57,7 @@ repos:
       - id: yamllint
         exclude: tests/optimization/fixtures
   - repo: https://github.com/psf/black
-    rev: 23.11.0
+    rev: 24.1.1
     hooks:
       - id: black
         language_version: python3.10
@@ -79,7 +79,7 @@ repos:
           - --blank
         exclude: src/estimagic/optimization/algo_options.py
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.1.6
+    rev: v0.2.0
     hooks:
       - id: ruff
   - repo: https://github.com/nbQA-dev/nbQA

src/estimagic/batch_evaluators.py

@@ -4,6 +4,7 @@
 can be used used as batch evaluator in estimagic.
 
 """
+
 from joblib import Parallel, delayed
 
 try:

src/estimagic/benchmarking/cartis_roberts.py

@@ -14,6 +14,7 @@
 - https://vanderbei.princeton.edu/ampl/nlmodels/cute/index.html
 
 """
+
 from functools import partial
 
 import numpy as np
@@ -5128,7 +5129,7 @@ def get_start_points_methanl8():
         "start_x": [-1] * 98,
         "solution_x": None,
         "start_criterion": 3.2160e4,
-        "solution_criterion": None
+        "solution_criterion": None,
         # we found a lower minimum than Cartis and Roberts (2019) at 1651.837;
         # according to table 3 in their paper, the minimum is at 4292.197.
         # We suspect, however, that the true optimum is even lower.

src/estimagic/benchmarking/more_wild.py

@@ -20,6 +20,7 @@
     - Whether the base start vector is multiplied by a factor of ten or not (column 4).
 
 """
+
 from functools import partial
 
 import numpy as np

src/estimagic/benchmarking/run_benchmark.py

@@ -7,6 +7,7 @@
 - Add option for deterministic noise or wiggle.
 
 """
+
 import numpy as np
 
 from estimagic import batch_evaluators

src/estimagic/cli.py

@@ -1,4 +1,5 @@
 """This module comprises all CLI capabilities of estimagic."""
+
 import click
 
 from estimagic.dashboard.run_dashboard import run_dashboard

src/estimagic/dashboard/dashboard_app.py

@@ -1,4 +1,5 @@
 """Show the development of one database's criterion and parameters over time."""
+
 from functools import partial
 from pathlib import Path
 

src/estimagic/dashboard/plot_functions.py

@@ -1,4 +1,5 @@
 """Helper functions for the dashboard."""
+
 from bokeh.models import HoverTool, Legend
 from bokeh.plotting import figure
 

src/estimagic/decorators.py

@@ -23,6 +23,7 @@
 https://realpython.com/primer-on-python-decorators/
 
 """
+
 import functools
 import inspect
 import warnings

src/estimagic/differentiation/derivatives.py

@@ -747,9 +747,11 @@ def _convert_evals_to_numpy(
         evals = [val.astype(float) if not _is_scalar_nan(val) else val for val in evals]
     else:
         evals = [
-            np.array(tree_leaves(val, registry=registry), dtype=np.float64)
-            if not _is_scalar_nan(val)
-            else val
+            (
+                np.array(tree_leaves(val, registry=registry), dtype=np.float64)
+                if not _is_scalar_nan(val)
+                else val
+            )
             for val in evals
         ]
 

src/estimagic/differentiation/finite_differences.py

@@ -8,6 +8,7 @@
 warnings or errors for that case.
 
 """
+
 from typing import NamedTuple
 
 import numpy as np

src/estimagic/estimation/estimate_msm.py

@@ -1,4 +1,5 @@
 """Do a method of simlated moments estimation."""
+
 import functools
 import warnings
 from collections.abc import Callable

src/estimagic/examples/logit.py

@@ -1,4 +1,5 @@
 """Likelihood functions and derivatives of a logit model."""
+
 import numpy as np
 import pandas as pd
 

src/estimagic/examples/numdiff_functions.py

@@ -8,6 +8,7 @@
 We pickled them so we don't need statsmodels as a dependency.
 
 """
+
 import numpy as np
 from scipy.stats import norm
 

src/estimagic/inference/ml_covs.py

@@ -1,4 +1,5 @@
 """Functions for inferences in maximum likelihood models."""
+
 import numpy as np
 import pandas as pd
 

src/estimagic/logging/read_from_database.py

@@ -9,6 +9,7 @@
 ``read_log.py`` instead.
 
 """
+
 import traceback
 import warnings
 

src/estimagic/logging/read_log.py

@@ -7,6 +7,7 @@
 path_or_database. Otherwise, the functions may be very slow.
 
 """
+
 from dataclasses import dataclass
 from pathlib import Path
 from typing import Union

src/estimagic/optimization/bhhh.py

@@ -1,4 +1,5 @@
 """Implement Berndt-Hall-Hall-Hausman (BHHH) algorithm."""
+
 import numpy as np
 
 from estimagic.decorators import mark_minimizer

src/estimagic/optimization/cyipopt_optimizers.py

@@ -1,4 +1,5 @@
 """Implement cyipopt's Interior Point Optimizer."""
+
 import numpy as np
 
 from estimagic.config import IS_CYIPOPT_INSTALLED

src/estimagic/optimization/fides_optimizers.py

@@ -1,4 +1,5 @@
 """Implement the fides optimizer."""
+
 import logging
 
 import numpy as np

src/estimagic/optimization/nag_optimizers.py

@@ -9,6 +9,7 @@
   and algorithm.
 
 """
+
 import warnings
 
 import numpy as np

src/estimagic/optimization/neldermead.py

@@ -1,4 +1,5 @@
 """Implementation of parallelosation of Nelder-Mead algorithm."""
+
 import numpy as np
 
 from estimagic.batch_evaluators import process_batch_evaluator

src/estimagic/optimization/nlopt_optimizers.py

@@ -3,6 +3,7 @@
 The documentation is heavily based on (nlopt documentation)[nlopt.readthedocs.io].
 
 """
+
 import numpy as np
 
 from estimagic.config import IS_NLOPT_INSTALLED

src/estimagic/optimization/pounders.py

@@ -1,4 +1,5 @@
 """Implement the POUNDERS algorithm."""
+
 import warnings
 
 import numpy as np

src/estimagic/optimization/pounders_auxiliary.py

@@ -1,4 +1,5 @@
 """Auxiliary functions for the pounders algorithm."""
+
 from typing import NamedTuple, Union
 
 import numpy as np
@@ -15,9 +16,9 @@
 class ResidualModel(NamedTuple):
     intercepts: Union[np.ndarray, None] = None  # shape (n_residuals,)
     linear_terms: Union[np.ndarray, None] = None  # shape (n_residuals, n_params)
-    square_terms: Union[
-        np.ndarray, None
-    ] = None  # shape (n_residuals, n_params, n_params)
+    square_terms: Union[np.ndarray, None] = (
+        None  # shape (n_residuals, n_params, n_params)
+    )
 
 
 class MainModel(NamedTuple):

src/estimagic/optimization/pounders_history.py

@@ -1,4 +1,5 @@
 """History class for pounders and similar optimizers."""
+
 import numpy as np
 
 

src/estimagic/optimization/pygmo_optimizers.py

@@ -1,4 +1,5 @@
 """Implement pygmo optimizers."""
+
 import contextlib
 import warnings
 

src/estimagic/optimization/scipy_optimizers.py

@@ -34,6 +34,7 @@
     terminate if it returned True.
 
 """
+
 import functools
 
 import numpy as np

src/estimagic/optimization/simopt_optimizers.py

@@ -4,6 +4,7 @@
 However,     for the moment they are not supported.
 
 """
+
 import numpy as np
 
 from estimagic.config import IS_SIMOPT_INSTALLED

src/estimagic/optimization/subsolvers/_conjugate_gradient.py

@@ -1,4 +1,5 @@
 """Implementation of the Conjugate Gradient algorithm."""
+
 import numpy as np
 
 

src/estimagic/optimization/subsolvers/_steihaug_toint.py

@@ -1,4 +1,5 @@
 """Implementation of the Steihaug-Toint Conjugate Gradient algorithm."""
+
 import numpy as np
 
 

src/estimagic/optimization/subsolvers/_trsbox.py

@@ -1,4 +1,5 @@
 """Implementation of the quadratic trustregion solver TRSBOX."""
+
 import numpy as np
 
 

src/estimagic/optimization/subsolvers/bntr.py

@@ -1,4 +1,5 @@
 """Auxiliary functions for the quadratic BNTR trust-region subsolver."""
+
 from functools import reduce
 from typing import NamedTuple, Union
 
@@ -583,8 +584,7 @@ def _perform_gradient_descent_step(
         square_terms = x_inactive.T @ hessian_inactive @ x_inactive
 
         predicted_reduction = trustregion_radius * (
-            gradient_norm
-            - 0.5 * trustregion_radius * square_terms / (gradient_norm**2)
+            gradient_norm - 0.5 * trustregion_radius * square_terms / (gradient_norm**2)
         )
         actual_reduction = f_candidate_initial - f_candidate
 

src/estimagic/optimization/subsolvers/gqtpar.py

@@ -1,4 +1,5 @@
 """Auxiliary functions for the quadratic GQTPAR trust-region subsolver."""
+
 from typing import NamedTuple, Union
 
 import numpy as np
@@ -356,9 +357,7 @@ def _update_lambdas_when_factorization_unsuccessful(
     )
     v_norm = np.linalg.norm(v)
 
-    lambda_lower_bound = max(
-        lambdas.lower_bound, lambdas.candidate + delta / v_norm**2
-    )
+    lambda_lower_bound = max(lambdas.lower_bound, lambdas.candidate + delta / v_norm**2)
     lambda_new_candidate = _get_new_lambda_candidate(
         lower_bound=lambda_lower_bound, upper_bound=lambdas.upper_bound
     )

src/estimagic/optimization/subsolvers/linear_subsolvers.py

@@ -1,4 +1,5 @@
 """Collection of linear trust-region subsolvers."""
+
 from typing import NamedTuple, Union
 
 import numpy as np
@@ -326,9 +327,7 @@ def _get_distance_to_trustregion_boundary(
     else:
         distance_to_boundary = (
             np.sqrt(
-                np.maximum(
-                    0, g_dot_x**2 + g_sumsq * (trustregion_radius**2 - x_sumsq)
-                )
+                np.maximum(0, g_dot_x**2 + g_sumsq * (trustregion_radius**2 - x_sumsq))
             )
             - g_dot_x
         ) / g_sumsq

src/estimagic/optimization/tao_optimizers.py

@@ -1,4 +1,5 @@
 """This module implements the POUNDERs algorithm."""
+
 import contextlib
 import functools
 
@@ -257,9 +258,9 @@ def _process_pounders_results(residuals_out, tao):
         "n_derivative_evaluations": None,
         "n_iterations": None,
         "success": bool(convergence_code >= 0),
-        "reached_convergence_criterion": convergence_reason
-        if convergence_code >= 0
-        else None,
+        "reached_convergence_criterion": (
+            convergence_reason if convergence_code >= 0 else None
+        ),
         "message": convergence_reason,
         # Further results.
         "solution_criterion_values": residuals_out.array,

src/estimagic/optimization/tiktak.py

@@ -10,6 +10,7 @@
 `GitHub Repository <https://github.com/amckay/TikTak>`_)
 
 """
+
 import warnings
 from functools import partial
 

src/estimagic/parameters/block_trees.py

@@ -1,4 +1,5 @@
 """Functions to convert between array and block-tree representations of a matrix."""
+
 import numpy as np
 import pandas as pd
 from pybaum import tree_flatten, tree_unflatten

src/estimagic/parameters/check_constraints.py

@@ -3,6 +3,7 @@
 See the module docstring of process_constraints for naming conventions.
 
 """
+
 from functools import partial
 
 import numpy as np

src/estimagic/parameters/consolidate_constraints.py

@@ -6,6 +6,7 @@
 Check the module docstring of process_constraints for naming conventions.
 
 """
+
 import numpy as np
 import pandas as pd
 

src/estimagic/parameters/conversion.py

@@ -1,4 +1,5 @@
 """Aggregate the multiple parameter and function output conversions into on."""
+
 from typing import NamedTuple
 
 import numpy as np

src/estimagic/parameters/kernel_transformations.py

@@ -28,6 +28,7 @@
    https://google.github.io/styleguide/pyguide.html
 
 """
+
 import numpy as np
 
 from estimagic.utilities import (

src/estimagic/parameters/process_constraints.py

@@ -14,6 +14,7 @@
 only be done after consolidation.
 
 """
+
 import numpy as np
 import pandas as pd
 

src/estimagic/parameters/space_conversion.py

@@ -32,6 +32,7 @@
 n_internal the length of the internal parameter vector.
 
 """
+
 from functools import partial
 from typing import NamedTuple
 

src/estimagic/parameters/tree_registry.py

@@ -1,4 +1,5 @@
 """Wrapper around pybaum get_registry to tailor it to estimagic."""
+
 from functools import partial
 from itertools import product