MNT improve the convergence warning message for LogisticRegression (scikit-learn#15665)

ogrisel · adrinjalali · commit 97c8086c2d18 · 2019-11-25T12:50:39.000+01:00
diff --git a/sklearn/linear_model/_logistic.py b/sklearn/linear_model/_logistic.py
@@ -38,6 +38,12 @@
 from ..metrics import get_scorer
 
 
+_LOGISTIC_SOLVER_CONVERGENCE_MSG = (
+    "Please also refer to the documentation for alternative solver options:\n"
+    "    https://scikit-learn.org/stable/modules/linear_model.html"
+    "#logistic-regression")
+
+
 # .. some helper functions for logistic_regression_path ..
 def _intercept_dot(w, X, y):
     """Computes y * np.dot(X, w).
@@ -928,7 +934,9 @@ def _logistic_regression_path(X, y, pos_class=None, Cs=10, fit_intercept=True,
                 args=(X, target, 1. / C, sample_weight),
                 options={"iprint": iprint, "gtol": tol, "maxiter": max_iter}
             )
-            n_iter_i = _check_optimize_result(solver, opt_res, max_iter)
+            n_iter_i = _check_optimize_result(
+                solver, opt_res, max_iter,
+                extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)
             w0, loss = opt_res.x, opt_res.fun
         elif solver == 'newton-cg':
             args = (X, target, 1. / C, sample_weight)
diff --git a/sklearn/linear_model/tests/test_logistic.py b/sklearn/linear_model/tests/test_logistic.py
@@ -31,7 +31,6 @@
 from sklearn.utils._testing import skip_if_no_parallel
 
 from sklearn.exceptions import ConvergenceWarning
-from sklearn.exceptions import ChangedBehaviorWarning
 from sklearn.linear_model._logistic import (
     LogisticRegression,
     logistic_regression_path,
@@ -391,13 +390,20 @@ def test_logistic_regression_path_convergence_fail():
     y = [1] * 100 + [-1] * 100
     Cs = [1e3]
 
-    msg = (r"lbfgs failed to converge.+Increase the number of iterations or "
-           r"scale the data")
-
-    with pytest.warns(ConvergenceWarning, match=msg):
+    # Check that the convergence message points to both a model agnostic
+    # advice (scaling the data) and to the logistic regression specific
+    # documentation that includes hints on the solver configuration.
+    with pytest.warns(ConvergenceWarning) as record:
         _logistic_regression_path(
             X, y, Cs=Cs, tol=0., max_iter=1, random_state=0, verbose=0)
 
+    assert len(record) == 1
+    warn_msg = record[0].message.args[0]
+    assert "lbfgs failed to converge" in warn_msg
+    assert "Increase the number of iterations" in warn_msg
+    assert "scale the data" in warn_msg
+    assert "linear_model.html#logistic-regression" in warn_msg
+
 
 def test_liblinear_dual_random_state():
     # random_state is relevant for liblinear solver only if dual=True
diff --git a/sklearn/utils/optimize.py b/sklearn/utils/optimize.py
@@ -213,7 +213,8 @@ def _newton_cg(grad_hess, func, grad, x0, args=(), tol=1e-4,
     return xk, k
 
 
-def _check_optimize_result(solver, result, max_iter=None):
+def _check_optimize_result(solver, result, max_iter=None,
+                           extra_warning_msg=None):
     """Check the OptimizeResult for successful convergence
 
     Parameters
@@ -233,12 +234,16 @@ def _check_optimize_result(solver, result, max_iter=None):
     # handle both scipy and scikit-learn solver names
     if solver == "lbfgs":
         if result.status != 0:
-            warnings.warn("{} failed to converge (status={}): {}. "
-                          "Increase the number of iterations or scale the "
-                          "data as shown in https://scikit-learn.org/stable/"
-                          "modules/preprocessing.html"
-                          .format(solver, result.status, result.message),
-                          ConvergenceWarning, stacklevel=2)
+            warning_msg = (
+                "{} failed to converge (status={}):\n{}.\n\n"
+                "Increase the number of iterations (max_iter) "
+                "or scale the data as shown in:\n"
+                "    https://scikit-learn.org/stable/modules/"
+                "preprocessing.html."
+            ).format(solver, result.status, result.message.decode("latin1"))
+            if extra_warning_msg is not None:
+                warning_msg += "\n" + extra_warning_msg
+            warnings.warn(warning_msg, ConvergenceWarning, stacklevel=2)
         if max_iter is not None:
             # In scipy <= 1.0.0, nit may exceed maxiter for lbfgs.
             # See https://github.com/scipy/scipy/issues/7854
