From 5202b6c1f6af283c4bc08e954db30e4e94719b61 Mon Sep 17 00:00:00 2001
From: jeremie du boisberranger <jeremiedbb@yahoo.fr>
Date: Tue, 15 Mar 2022 22:27:45 +0100
Subject: [PATCH 1/5] change default to diagram

---
 doc/modules/compose.rst                       | 17 ++++++----
 .../plot_column_transformer_mixed_types.py    |  6 ++--
 .../ensemble/plot_feature_transformation.py   |  4 ---
 examples/ensemble/plot_stack_predictors.py    |  4 ---
 .../plot_feature_selection_pipeline.py        |  4 ---
 examples/linear_model/plot_lasso_lars_ic.py   |  5 ---
 .../plot_lasso_model_selection.py             |  5 ---
 .../miscellaneous/plot_pipeline_display.py    | 34 +++++--------------
 sklearn/_config.py                            |  6 ++--
 sklearn/tests/test_base.py                    | 18 +++++-----
 sklearn/tests/test_config.py                  |  6 ++--
 11 files changed, 36 insertions(+), 73 deletions(-)

diff --git a/doc/modules/compose.rst b/doc/modules/compose.rst
index 4c54c496490f0..1ed09667eae3a 100644
--- a/doc/modules/compose.rst
+++ b/doc/modules/compose.rst
@@ -567,16 +567,19 @@ will use the column names to select the columns::
 Visualizing Composite Estimators
 ================================
 
-Estimators can be displayed with a HTML representation when shown in a
-jupyter notebook. This can be useful to diagnose or visualize a Pipeline with
-many estimators. This visualization is activated by setting the
-`display` option in :func:`~sklearn.set_config`::
+Estimators are displayed with an HTML representation when shown in a
+jupyter notebook. This is useful to diagnose or visualize a Pipeline with
+many estimators. This visualization is activated by default::
 
-  >>> from sklearn import set_config
-  >>> set_config(display='diagram')   # doctest: +SKIP
-  >>> # displays HTML representation in a jupyter context
   >>> column_trans  # doctest: +SKIP
 
+It can be deactivated by setting the `text` option in :func:`~sklearn.set_config`::
+
+  >>> from sklearn import set_config
+  >>> set_config(display='text')
+  >>> # displays text representation in a jupyter context
+  >>> column_trans
+
 An example of the HTML output can be seen in the
 **HTML representation of Pipeline** section of
 :ref:`sphx_glr_auto_examples_compose_plot_column_transformer_mixed_types.py`.
diff --git a/examples/compose/plot_column_transformer_mixed_types.py b/examples/compose/plot_column_transformer_mixed_types.py
index 307f0384ab1d2..d7503bbc39607 100644
--- a/examples/compose/plot_column_transformer_mixed_types.py
+++ b/examples/compose/plot_column_transformer_mixed_types.py
@@ -97,10 +97,7 @@
 # HTML representation of ``Pipeline`` (display diagram)
 ###############################################################################
 # When the ``Pipeline`` is printed out in a jupyter notebook an HTML
-# representation of the estimator is displayed as follows:
-from sklearn import set_config
-
-set_config(display="diagram")
+# representation of the estimator is displayed:
 clf
 
 # %%
@@ -150,6 +147,7 @@
 
 clf.fit(X_train, y_train)
 print("model score: %.3f" % clf.score(X_test, y_test))
+clf
 
 # %%
 # The resulting score is not exactly the same as the one from the previous
diff --git a/examples/ensemble/plot_feature_transformation.py b/examples/ensemble/plot_feature_transformation.py
index 53c5d9b1a233c..409396a0376b8 100644
--- a/examples/ensemble/plot_feature_transformation.py
+++ b/examples/ensemble/plot_feature_transformation.py
@@ -25,10 +25,6 @@
 #
 # License: BSD 3 clause
 
-from sklearn import set_config
-
-set_config(display="diagram")
-
 # %%
 # First, we will create a large dataset and split it into three sets:
 #
diff --git a/examples/ensemble/plot_stack_predictors.py b/examples/ensemble/plot_stack_predictors.py
index 959adfbcf4dfc..7737e91a0fdec 100644
--- a/examples/ensemble/plot_stack_predictors.py
+++ b/examples/ensemble/plot_stack_predictors.py
@@ -20,10 +20,6 @@
 #          Maria Telenczuk    <https://github.com/maikia>
 # License: BSD 3 clause
 
-from sklearn import set_config
-
-set_config(display="diagram")
-
 # %%
 # Download the dataset
 ##############################################################################
diff --git a/examples/feature_selection/plot_feature_selection_pipeline.py b/examples/feature_selection/plot_feature_selection_pipeline.py
index 22a8bfc8dd5c3..7a9a621aad5ee 100644
--- a/examples/feature_selection/plot_feature_selection_pipeline.py
+++ b/examples/feature_selection/plot_feature_selection_pipeline.py
@@ -10,10 +10,6 @@
 
 """
 
-from sklearn import set_config
-
-set_config(display="diagram")
-
 # %%
 # We will start by generating a binary classification dataset. Subsequently, we
 # will divide the dataset into two subsets.
diff --git a/examples/linear_model/plot_lasso_lars_ic.py b/examples/linear_model/plot_lasso_lars_ic.py
index 2f5392696ecc9..31a15b2ba582c 100644
--- a/examples/linear_model/plot_lasso_lars_ic.py
+++ b/examples/linear_model/plot_lasso_lars_ic.py
@@ -28,11 +28,6 @@
 #         Guillaume Lemaitre
 # License: BSD 3 clause
 
-# %%
-import sklearn
-
-sklearn.set_config(display="diagram")
-
 # %%
 # We will use the diabetes dataset.
 from sklearn.datasets import load_diabetes
diff --git a/examples/linear_model/plot_lasso_model_selection.py b/examples/linear_model/plot_lasso_model_selection.py
index 7cc05055b22d9..bf2111e32b427 100644
--- a/examples/linear_model/plot_lasso_model_selection.py
+++ b/examples/linear_model/plot_lasso_model_selection.py
@@ -19,11 +19,6 @@
 #         Guillaume Lemaitre
 # License: BSD 3 clause
 
-# %%
-import sklearn
-
-sklearn.set_config(display="diagram")
-
 # %%
 # Dataset
 # -------
diff --git a/examples/miscellaneous/plot_pipeline_display.py b/examples/miscellaneous/plot_pipeline_display.py
index 0ba638afb0593..664a04db4c7c9 100755
--- a/examples/miscellaneous/plot_pipeline_display.py
+++ b/examples/miscellaneous/plot_pipeline_display.py
@@ -3,9 +3,9 @@
 Displaying Pipelines
 =================================================================
 
-The default configuration for displaying a pipeline is `'text'` where
-`set_config(display='text')`.  To visualize the diagram in Jupyter Notebook,
-use `set_config(display='diagram')` and then output the pipeline object.
+The default configuration for displaying a pipeline in a Jupyter Notebook is
+`'diagram'` where `set_config(display='diagram')`. To deactivate HTML representation,
+use `set_config(display='text')`.
 
 To see more detailed steps in the visualization of the pipeline, click on the
 steps in the pipeline.
@@ -31,15 +31,15 @@
 pipe = Pipeline(steps)
 
 # %%
-# To view the text pipeline, the default is `display='text'`.
-set_config(display="text")
-pipe
-
-# %%
-# To visualize the diagram, change `display='diagram'`.
+# To visualize the diagram, the default is `display='diagram'`.
 set_config(display="diagram")
 pipe  # click on the diagram below to see the details of each step
 
+# %%
+# To view the text pipeline, change to `display='text'`.
+set_config(display="text")
+pipe
+
 # %%
 # Displaying a Pipeline Chaining Multiple Preprocessing Steps & Classifier
 ################################################################################
@@ -60,10 +60,6 @@
     ("classifier", LogisticRegression(C=2.0)),
 ]
 pipe = Pipeline(steps)
-
-# %%
-# To visualize the diagram, change to display='diagram'
-set_config(display="diagram")
 pipe  # click on the diagram below to see the details of each step
 
 # %%
@@ -81,10 +77,6 @@
 
 steps = [("reduce_dim", PCA(n_components=4)), ("classifier", SVC(kernel="linear"))]
 pipe = Pipeline(steps)
-
-# %%
-# To visualize the diagram, change to `display='diagram'`.
-set_config(display="diagram")
 pipe  # click on the diagram below to see the details of each step
 
 # %%
@@ -129,10 +121,6 @@
 )
 
 pipe = make_pipeline(preprocessor, LogisticRegression(max_iter=500))
-
-# %%
-# To visualize the diagram, change to `display='diagram'`
-set_config(display="diagram")
 pipe  # click on the diagram below to see the details of each step
 
 # %%
@@ -189,8 +177,4 @@
 }
 
 grid_search = GridSearchCV(pipe, param_grid=param_grid, n_jobs=1)
-
-# %%
-# To visualize the diagram, change to `display='diagram'`.
-set_config(display="diagram")
 grid_search  # click on the diagram below to see the details of each step
diff --git a/sklearn/_config.py b/sklearn/_config.py
index 6248025b05aa0..c865b879dbea3 100644
--- a/sklearn/_config.py
+++ b/sklearn/_config.py
@@ -8,7 +8,7 @@
     "assume_finite": bool(os.environ.get("SKLEARN_ASSUME_FINITE", False)),
     "working_memory": int(os.environ.get("SKLEARN_WORKING_MEMORY", 1024)),
     "print_changed_only": True,
-    "display": "text",
+    "display": "diagram",
     "pairwise_dist_chunk_size": int(
         os.environ.get("SKLEARN_PAIRWISE_DIST_CHUNK_SIZE", 256)
     ),
@@ -85,7 +85,7 @@ def set_config(
     display : {'text', 'diagram'}, default=None
         If 'diagram', estimators will be displayed as a diagram in a Jupyter
         lab or notebook context. If 'text', estimators will be displayed as
-        text. Default is 'text'.
+        text. Default is 'diagram'.
 
         .. versionadded:: 0.23
 
@@ -173,7 +173,7 @@ def config_context(
         If 'diagram', estimators will be displayed as a diagram in a Jupyter
         lab or notebook context. If 'text', estimators will be displayed as
         text. If None, the existing value won't change.
-        The default value is 'text'.
+        The default value is 'diagram'.
 
         .. versionadded:: 0.23
 
diff --git a/sklearn/tests/test_base.py b/sklearn/tests/test_base.py
index 2608b77622e9a..aedefe29dff0c 100644
--- a/sklearn/tests/test_base.py
+++ b/sklearn/tests/test_base.py
@@ -539,25 +539,25 @@ def test_repr_mimebundle_():
     tree = DecisionTreeClassifier()
     output = tree._repr_mimebundle_()
     assert "text/plain" in output
-    assert "text/html" not in output
+    assert "text/html" in output
 
-    with config_context(display="diagram"):
+    with config_context(display="text"):
         output = tree._repr_mimebundle_()
         assert "text/plain" in output
-        assert "text/html" in output
+        assert "text/html" not in output
 
 
 def test_repr_html_wraps():
     # Checks the display configuration flag controls the html output
     tree = DecisionTreeClassifier()
-    msg = "_repr_html_ is only defined when"
-    with pytest.raises(AttributeError, match=msg):
-        output = tree._repr_html_()
 
-    with config_context(display="diagram"):
-        output = tree._repr_html_()
-        assert "<style>" in output
+    output = tree._repr_html_()
+    assert "<style>" in output
 
+    with config_context(display="text"):
+        msg = "_repr_html_ is only defined when"
+        with pytest.raises(AttributeError, match=msg):
+            output = tree._repr_html_()
 
 def test_n_features_in_validation():
     """Check that `_check_n_features` validates data when reset=False"""
diff --git a/sklearn/tests/test_config.py b/sklearn/tests/test_config.py
index c0bddf8cbab0e..86496d6bd45cf 100644
--- a/sklearn/tests/test_config.py
+++ b/sklearn/tests/test_config.py
@@ -13,7 +13,7 @@ def test_config_context():
         "assume_finite": False,
         "working_memory": 1024,
         "print_changed_only": True,
-        "display": "text",
+        "display": "diagram",
         "pairwise_dist_chunk_size": 256,
         "enable_cython_pairwise_dist": True,
     }
@@ -27,7 +27,7 @@ def test_config_context():
             "assume_finite": True,
             "working_memory": 1024,
             "print_changed_only": True,
-            "display": "text",
+            "display": "diagram",
             "pairwise_dist_chunk_size": 256,
             "enable_cython_pairwise_dist": True,
         }
@@ -58,7 +58,7 @@ def test_config_context():
         "assume_finite": False,
         "working_memory": 1024,
         "print_changed_only": True,
-        "display": "text",
+        "display": "diagram",
         "pairwise_dist_chunk_size": 256,
         "enable_cython_pairwise_dist": True,
     }

From a3d11205c8532b912e85d62a5aa8f51e1d0661ff Mon Sep 17 00:00:00 2001
From: jeremie du boisberranger <jeremiedbb@yahoo.fr>
Date: Tue, 15 Mar 2022 22:38:35 +0100
Subject: [PATCH 2/5] what's new

---
 doc/whats_new/v1.1.rst     | 5 +++++
 sklearn/tests/test_base.py | 1 +
 2 files changed, 6 insertions(+)

diff --git a/doc/whats_new/v1.1.rst b/doc/whats_new/v1.1.rst
index 03b1490a062eb..9c569ba45f0ac 100644
--- a/doc/whats_new/v1.1.rst
+++ b/doc/whats_new/v1.1.rst
@@ -875,6 +875,11 @@ Changelog
 - |Fix| :func:`utils.estimator_html_repr` has an improved visualization for nested
   meta-estimators. :pr:`21310` by `Thomas Fan`_.
 
+- |API| Rich html representation of estimators is now enabled by default in Jupyter
+  notebooks. It can be deactivated by setting `display='text'` in
+  :func:`~sklearn.set_config`.
+  :pr:`22856` by `Jérémie du Boisberranger <jeremiedbb>`.
+
 Code and Documentation Contributors
 -----------------------------------
 
diff --git a/sklearn/tests/test_base.py b/sklearn/tests/test_base.py
index aedefe29dff0c..88a065fe79657 100644
--- a/sklearn/tests/test_base.py
+++ b/sklearn/tests/test_base.py
@@ -559,6 +559,7 @@ def test_repr_html_wraps():
         with pytest.raises(AttributeError, match=msg):
             output = tree._repr_html_()
 
+
 def test_n_features_in_validation():
     """Check that `_check_n_features` validates data when reset=False"""
     est = MyEstimator()

From 25e87e66071079b3f4a6900500e20abe6d75afaa Mon Sep 17 00:00:00 2001
From: jeremie du boisberranger <jeremiedbb@yahoo.fr>
Date: Tue, 15 Mar 2022 23:38:35 +0100
Subject: [PATCH 3/5] fix doc test

---
 doc/modules/compose.rst | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/doc/modules/compose.rst b/doc/modules/compose.rst
index 1ed09667eae3a..432423929c7b9 100644
--- a/doc/modules/compose.rst
+++ b/doc/modules/compose.rst
@@ -576,9 +576,9 @@ many estimators. This visualization is activated by default::
 It can be deactivated by setting the `text` option in :func:`~sklearn.set_config`::
 
   >>> from sklearn import set_config
-  >>> set_config(display='text')
+  >>> set_config(display='text')  # doctest: +SKIP
   >>> # displays text representation in a jupyter context
-  >>> column_trans
+  >>> column_trans  # doctest: +SKIP
 
 An example of the HTML output can be seen in the
 **HTML representation of Pipeline** section of

From 24cdd6663c5405085d7bceada4841fe71c08ed65 Mon Sep 17 00:00:00 2001
From: jeremie du boisberranger <jeremiedbb@yahoo.fr>
Date: Wed, 16 Mar 2022 11:25:44 +0100
Subject: [PATCH 4/5] iter

---
 .../compose/plot_column_transformer_mixed_types.py    | 11 +++++++----
 examples/miscellaneous/plot_pipeline_display.py       |  8 ++++----
 sklearn/utils/tests/test_pprint.py                    |  2 +-
 3 files changed, 12 insertions(+), 9 deletions(-)

diff --git a/examples/compose/plot_column_transformer_mixed_types.py b/examples/compose/plot_column_transformer_mixed_types.py
index d7503bbc39607..afc0b07f5d126 100644
--- a/examples/compose/plot_column_transformer_mixed_types.py
+++ b/examples/compose/plot_column_transformer_mixed_types.py
@@ -28,6 +28,7 @@
 #
 # License: BSD 3 clause
 
+# %%
 import numpy as np
 
 from sklearn.compose import ColumnTransformer
@@ -40,6 +41,7 @@
 
 np.random.seed(0)
 
+# %%
 # Load data from https://www.openml.org/d/40945
 X, y = fetch_openml("titanic", version=1, as_frame=True, return_X_y=True)
 
@@ -49,7 +51,7 @@
 
 # %%
 # Use ``ColumnTransformer`` by selecting column by names
-###############################################################################
+#
 # We will train our classifier with the following features:
 #
 # Numeric Features:
@@ -82,6 +84,7 @@
     ]
 )
 
+# %%
 # Append classifier to preprocessing pipeline.
 # Now we have a full prediction pipeline.
 clf = Pipeline(
@@ -95,14 +98,14 @@
 
 # %%
 # HTML representation of ``Pipeline`` (display diagram)
-###############################################################################
+#
 # When the ``Pipeline`` is printed out in a jupyter notebook an HTML
 # representation of the estimator is displayed:
 clf
 
 # %%
 # Use ``ColumnTransformer`` by selecting column by data types
-###############################################################################
+#
 # When dealing with a cleaned dataset, the preprocessing can be automatic by
 # using the data types of the column to decide whether to treat a column as a
 # numerical or categorical feature.
@@ -162,7 +165,7 @@
 
 # %%
 # Using the prediction pipeline in a grid search
-##############################################################################
+#
 # Grid search can also be performed on the different preprocessing steps
 # defined in the ``ColumnTransformer`` object, together with the classifier's
 # hyperparameters as part of the ``Pipeline``.
diff --git a/examples/miscellaneous/plot_pipeline_display.py b/examples/miscellaneous/plot_pipeline_display.py
index 664a04db4c7c9..3e785551132b5 100755
--- a/examples/miscellaneous/plot_pipeline_display.py
+++ b/examples/miscellaneous/plot_pipeline_display.py
@@ -40,6 +40,10 @@
 set_config(display="text")
 pipe
 
+# %%
+# Put back the default display
+set_config(display="diagram")
+
 # %%
 # Displaying a Pipeline Chaining Multiple Preprocessing Steps & Classifier
 ################################################################################
@@ -52,7 +56,6 @@
 from sklearn.pipeline import Pipeline
 from sklearn.preprocessing import StandardScaler, PolynomialFeatures
 from sklearn.linear_model import LogisticRegression
-from sklearn import set_config
 
 steps = [
     ("standard_scaler", StandardScaler()),
@@ -73,7 +76,6 @@
 from sklearn.pipeline import Pipeline
 from sklearn.svm import SVC
 from sklearn.decomposition import PCA
-from sklearn import set_config
 
 steps = [("reduce_dim", PCA(n_components=4)), ("classifier", SVC(kernel="linear"))]
 pipe = Pipeline(steps)
@@ -94,7 +96,6 @@
 from sklearn.compose import ColumnTransformer
 from sklearn.preprocessing import OneHotEncoder, StandardScaler
 from sklearn.linear_model import LogisticRegression
-from sklearn import set_config
 
 numeric_preprocessor = Pipeline(
     steps=[
@@ -139,7 +140,6 @@
 from sklearn.preprocessing import OneHotEncoder, StandardScaler
 from sklearn.ensemble import RandomForestClassifier
 from sklearn.model_selection import GridSearchCV
-from sklearn import set_config
 
 numeric_preprocessor = Pipeline(
     steps=[
diff --git a/sklearn/utils/tests/test_pprint.py b/sklearn/utils/tests/test_pprint.py
index d4c93779eb110..aa1e2e03841e9 100644
--- a/sklearn/utils/tests/test_pprint.py
+++ b/sklearn/utils/tests/test_pprint.py
@@ -8,7 +8,7 @@
 from sklearn.pipeline import make_pipeline
 from sklearn.base import BaseEstimator, TransformerMixin
 from sklearn.feature_selection import SelectKBest, chi2
-from sklearn import set_config, config_context
+from sklearn import config_context
 
 
 # Ignore flake8 (lots of line too long issues)

From 2a313408d779eb9ffd076eb19e391d5934038563 Mon Sep 17 00:00:00 2001
From: jeremie du boisberranger <jeremiedbb@yahoo.fr>
Date: Fri, 18 Mar 2022 15:35:04 +0100
Subject: [PATCH 5/5] address comments

---
 doc/modules/compose.rst | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/doc/modules/compose.rst b/doc/modules/compose.rst
index 432423929c7b9..2a2b007783f27 100644
--- a/doc/modules/compose.rst
+++ b/doc/modules/compose.rst
@@ -573,7 +573,8 @@ many estimators. This visualization is activated by default::
 
   >>> column_trans  # doctest: +SKIP
 
-It can be deactivated by setting the `text` option in :func:`~sklearn.set_config`::
+It can be deactivated by setting the `display` option in :func:`~sklearn.set_config`
+to 'text'::
 
   >>> from sklearn import set_config
   >>> set_config(display='text')  # doctest: +SKIP