Merge branch 'main' into s-notebook-CATE

Author: SvenKlaassen

.github/workflows/deploy_docu_dev.yml

@@ -18,24 +18,19 @@ jobs:
 
     steps:
     - name: Check out the repo containing the docu source
-      uses: actions/checkout@v2
+      uses: actions/checkout@v3
 
     - name: Check out the repo containing the python pkg DoubleML (dev)
-      uses: actions/checkout@v2
+      uses: actions/checkout@v3
       with:
         repository: DoubleML/doubleml-for-py
         path: doubleml-for-py
 
-    - name: Install SSH Client for deploying the docu to github pages
-      uses: webfactory/[email protected]
-      with:
-        ssh-private-key: ${{ secrets.DEPLOY_KEY }}
-
     - name: Install graphviz
       run: sudo apt-get install graphviz
 
     - name: Install python
-      uses: actions/setup-python@v2
+      uses: actions/setup-python@v4
       with:
         python-version: '3.8'
     - name: Install dependencies and the python package
@@ -69,7 +64,7 @@ jobs:
       shell: Rscript {0}
 
     - name: Cache R packages
-      uses: actions/cache@v2
+      uses: actions/cache@v3
       with:
         path: ${{ env.R_LIBS_USER }}
         key: doubleml-user-guide-dev-${{ hashFiles('.github/R-version') }}
@@ -87,13 +82,13 @@ jobs:
         make -C doc html # build docu
 
     - name: Deploy to dev
-      uses: JamesIves/github-pages-deploy-action@3.7.1
+      uses: JamesIves/github-pages-deploy-action@v4
       with:
-        REPOSITORY_NAME: DoubleML/doubleml.github.io
-        BRANCH: main
-        FOLDER: doc/_build/html
-        TARGET_FOLDER: dev
-        GIT_CONFIG_NAME: DoubleML Deploy Bot
-        GIT_CONFIG_EMAIL: [email protected]
-        CLEAN: true
-        SSH: true
+        repository-name: DoubleML/doubleml.github.io
+        branch: main
+        folder: doc/_build/html
+        target-folder: dev
+        git-config-name: DoubleML Deploy Bot
+        git-config-email: [email protected]
+        clean: true
+        ssh-key: ${{ secrets.DEPLOY_KEY }}

.github/workflows/deploy_docu_stable.yml

@@ -13,18 +13,13 @@ jobs:
 
     steps:
     - name: Check out the repo containing the docu source
-      uses: actions/checkout@v2
-
-    - name: Install SSH Client for deploying the docu to github pages
-      uses: webfactory/[email protected]
-      with:
-        ssh-private-key: ${{ secrets.DEPLOY_KEY }}
+      uses: actions/checkout@v3
 
     - name: Install graphviz
       run: sudo apt-get install graphviz
 
     - name: Install python
-      uses: actions/setup-python@v2
+      uses: actions/setup-python@v4
       with:
         python-version: '3.8'
     - name: Install dependencies and the python package
@@ -55,7 +50,7 @@ jobs:
       shell: Rscript {0}
 
     - name: Cache R packages
-      uses: actions/cache@v2
+      uses: actions/cache@v3
       with:
         path: ${{ env.R_LIBS_USER }}
         key: doubleml-user-guide-stable-${{ hashFiles('.github/R-version') }}
@@ -73,13 +68,13 @@ jobs:
         make -C doc html # build docu
 
     - name: Deploy to stable
-      uses: JamesIves/github-pages-deploy-action@3.7.1
+      uses: JamesIves/github-pages-deploy-action@v4
       with:
-        REPOSITORY_NAME: DoubleML/doubleml.github.io
-        BRANCH: main
-        FOLDER: doc/_build/html
-        TARGET_FOLDER: stable
-        GIT_CONFIG_NAME: DoubleML Deploy Bot
-        GIT_CONFIG_EMAIL: [email protected]
-        CLEAN: true
-        SSH: true
+        repository-name: DoubleML/doubleml.github.io
+        branch: main
+        folder: doc/_build/html
+        target-folder: stable
+        git-config-name: DoubleML Deploy Bot
+        git-config-email: [email protected]
+        clean: true
+        ssh-key: ${{ secrets.DEPLOY_KEY }}

.github/workflows/test_build_docu_dev.yml

@@ -32,18 +32,18 @@ jobs:
 
     steps:
     - name: Check out the repo containing the docu source
-      uses: actions/checkout@v2
+      uses: actions/checkout@v3
 
     - name: Check out the repo containing the python pkg DoubleML (dev)
       if: ${{ github.event_name != 'workflow_dispatch' }}
-      uses: actions/checkout@v2
+      uses: actions/checkout@v3
       with:
         repository: DoubleML/doubleml-for-py
         path: doubleml-for-py
 
     - name: Check out the repo containing the python pkg DoubleML (dev)
       if: ${{ github.event_name == 'workflow_dispatch' }}
-      uses: actions/checkout@v2
+      uses: actions/checkout@v3
       with:
         repository: DoubleML/doubleml-for-py
         path: doubleml-for-py
@@ -53,7 +53,7 @@ jobs:
       run: sudo apt-get install graphviz
 
     - name: Install python
-      uses: actions/setup-python@v2
+      uses: actions/setup-python@v4
       with:
         python-version: '3.8'
     - name: Install dependencies and the python package
@@ -87,7 +87,7 @@ jobs:
       shell: Rscript {0}
 
     - name: Cache R packages
-      uses: actions/cache@v2
+      uses: actions/cache@v3
       with:
         path: ${{ env.R_LIBS_USER }}
         key: doubleml-test-build-dev-${{ hashFiles('.github/R-version') }}
@@ -119,7 +119,7 @@ jobs:
         make -C doc linkcheck
 
     - name: Upload html artifacts
-      uses: actions/upload-artifact@v2
+      uses: actions/upload-artifact@v3
       with:
         name: build_html
         path: doc/_build/html/

.github/workflows/test_build_docu_released.yml

@@ -21,13 +21,13 @@ jobs:
 
     steps:
     - name: Check out the repo containing the docu source
-      uses: actions/checkout@v2
+      uses: actions/checkout@v3
 
     - name: Install graphviz
       run: sudo apt-get install graphviz
 
     - name: Install python
-      uses: actions/setup-python@v2
+      uses: actions/setup-python@v4
       with:
         python-version: '3.8'
     - name: Install dependencies and the python package
@@ -58,7 +58,7 @@ jobs:
       shell: Rscript {0}
 
     - name: Cache R packages
-      uses: actions/cache@v2
+      uses: actions/cache@v3
       with:
         path: ${{ env.R_LIBS_USER }}
         key: doubleml-test-build-stable-${{ hashFiles('.github/R-version') }}
@@ -80,7 +80,7 @@ jobs:
         make -C doc linkcheck
 
     - name: Upload html artifacts
-      uses: actions/upload-artifact@v2
+      uses: actions/upload-artifact@v3
       with:
         name: build_html
         path: doc/_build/html/

doc/guide/basics.rst

@@ -224,7 +224,7 @@ other half of observations indexed with :math:`i \in I`
         }
 
         g_nonorth = ggplot(data.frame(theta_rescaled=(theta_nonorth - alpha)/se_nonorth)) +
-                        geom_histogram(aes(y=..density.., x=theta_rescaled, colour = "Non-orthogonal ML", fill="Non-orthogonal ML"),
+                        geom_histogram(aes(y=after_stat(density), x=theta_rescaled, colour = "Non-orthogonal ML", fill="Non-orthogonal ML"),
                                        bins = 30, alpha = 0.3) +
                         geom_vline(aes(xintercept = 0), col = "black") +
                         suppressWarnings(geom_function(fun = dnorm, aes(colour = "N(0, 1)", fill="N(0, 1)"))) +
@@ -348,7 +348,7 @@ orthogonalized regressor :math:`V = D - m(X)`. We then use the final estimate
         }
 
         g_nosplit = ggplot(data.frame(theta_rescaled=(theta_orth_nosplit - alpha)/se_orth_nosplit), aes(x = theta_rescaled)) +
-                        geom_histogram(aes(y=..density.., x=theta_rescaled, colour = "Double ML (no sample splitting)", fill="Double ML (no sample splitting)"),
+                        geom_histogram(aes(y=after_stat(density), x=theta_rescaled, colour = "Double ML (no sample splitting)", fill="Double ML (no sample splitting)"),
                                        bins = 30, alpha = 0.3) +
                         geom_vline(aes(xintercept = 0), col = "black") +
                         suppressWarnings(geom_function(fun = dnorm, aes(colour = "N(0, 1)", fill="N(0, 1)"))) +
@@ -452,7 +452,7 @@ Cross-fitting performs well empirically because the entire sample can be used fo
         }
 
         g_dml = ggplot(data.frame(theta_rescaled=(theta_dml - alpha)/se_dml), aes(x = theta_rescaled)) +
-                        geom_histogram(aes(y=..density.., x=theta_rescaled, colour = "Double ML with cross-fitting", fill="Double ML with cross-fitting"),
+                        geom_histogram(aes(y=after_stat(density), x=theta_rescaled, colour = "Double ML with cross-fitting", fill="Double ML with cross-fitting"),
                                        bins = 30, alpha = 0.3) +
                         geom_vline(aes(xintercept = 0), col = "black") +
                         suppressWarnings(geom_function(fun = dnorm, aes(colour = "N(0, 1)", fill="N(0, 1)"))) +
@@ -525,11 +525,11 @@ The third term :math:`c^*` vanishes in probability if sample splitting is applie
         g_all = ggplot(data.frame(t_nonorth=(theta_nonorth - alpha)/se_nonorth,
                                   t_orth_nosplit=(theta_orth_nosplit - alpha)/se_orth_nosplit,
                                   t_dml=(theta_dml - alpha)/se_dml)) +
-                        geom_histogram(aes(x = t_nonorth, y=..density.., colour = "Non-orthogonal ML", fill="Non-orthogonal ML"),
+                        geom_histogram(aes(x = t_nonorth, y=after_stat(density), colour = "Non-orthogonal ML", fill="Non-orthogonal ML"),
                                            bins = 30, alpha = 0.3) +
-                        geom_histogram(aes(x = t_orth_nosplit, y=..density.., colour = "Double ML (no sample splitting)", fill="Double ML (no sample splitting)"),
+                        geom_histogram(aes(x = t_orth_nosplit, y=after_stat(density), colour = "Double ML (no sample splitting)", fill="Double ML (no sample splitting)"),
                                            bins = 30, alpha = 0.3) +
-                        geom_histogram(aes(x = t_dml, y=..density.., colour = "Double ML with cross-fitting", fill="Double ML with cross-fitting"),
+                        geom_histogram(aes(x = t_dml, y=after_stat(density), colour = "Double ML with cross-fitting", fill="Double ML with cross-fitting"),
                                            bins = 30, alpha = 0.3) +
                         geom_vline(aes(xintercept = 0), col = "black") +
                         suppressWarnings(geom_function(fun = dnorm, aes(colour = "N(0, 1)", fill="N(0, 1)"))) +
@@ -643,7 +643,7 @@ estimate :math:`\theta_0` without sample splitting. Again we observe a bias from
         }
 
         g_nosplit = ggplot(data.frame(theta_rescaled=(theta_orth_po_nosplit - alpha)/se_orth_po_nosplit), aes(x = theta_rescaled)) +
-                        geom_histogram(aes(y=..density.., x=theta_rescaled, colour = "Double ML (no sample splitting)", fill="Double ML (no sample splitting)"),
+                        geom_histogram(aes(y=after_stat(density), x=theta_rescaled, colour = "Double ML (no sample splitting)", fill="Double ML (no sample splitting)"),
                                        bins = 30, alpha = 0.3) +
                         geom_vline(aes(xintercept = 0), col = "black") +
                         suppressWarnings(geom_function(fun = dnorm, aes(colour = "N(0, 1)", fill="N(0, 1)"))) +
@@ -735,7 +735,7 @@ Using sample splitting, overcomes the bias induced by overfitting.
         }
 
         g_dml = ggplot(data.frame(theta_rescaled=(theta_dml_po - alpha)/se_dml_po), aes(x = theta_rescaled)) +
-                        geom_histogram(aes(y=..density.., x=theta_rescaled, colour = "Double ML with cross-fitting", fill="Double ML with cross-fitting"),
+                        geom_histogram(aes(y=after_stat(density), x=theta_rescaled, colour = "Double ML with cross-fitting", fill="Double ML with cross-fitting"),
                                        bins = 30, alpha = 0.3) +
                         geom_vline(aes(xintercept = 0), col = "black") +
                         suppressWarnings(geom_function(fun = dnorm, aes(colour = "N(0, 1)", fill="N(0, 1)"))) +
@@ -780,11 +780,11 @@ Using sample splitting, overcomes the bias induced by overfitting.
         g_all = ggplot(data.frame(t_nonorth=(theta_nonorth - alpha)/se_nonorth,
                                   t_orth_nosplit=(theta_orth_po_nosplit - alpha)/se_orth_po_nosplit,
                                   t_dml=(theta_dml_po - alpha)/se_dml_po)) +
-                        geom_histogram(aes(x = t_nonorth, y=..density.., colour = "Non-orthogonal ML", fill="Non-orthogonal ML"),
+                        geom_histogram(aes(x = t_nonorth, y=after_stat(density), colour = "Non-orthogonal ML", fill="Non-orthogonal ML"),
                                            bins = 30, alpha = 0.3) +
-                        geom_histogram(aes(x = t_orth_nosplit, y=..density.., colour = "Double ML (no sample splitting)", fill="Double ML (no sample splitting)"),
+                        geom_histogram(aes(x = t_orth_nosplit, y=after_stat(density), colour = "Double ML (no sample splitting)", fill="Double ML (no sample splitting)"),
                                            bins = 30, alpha = 0.3) +
-                        geom_histogram(aes(x = t_dml, y=..density.., colour = "Double ML with cross-fitting", fill="Double ML with cross-fitting"),
+                        geom_histogram(aes(x = t_dml, y=after_stat(density), colour = "Double ML with cross-fitting", fill="Double ML with cross-fitting"),
                                            bins = 30, alpha = 0.3) +
                         geom_vline(aes(xintercept = 0), col = "black") +
                         suppressWarnings(geom_function(fun = dnorm, aes(colour = "N(0, 1)", fill="N(0, 1)"))) +

doc/guide/learners.rst

@@ -189,7 +189,7 @@ The minimum requirement for a learner to be used for nuisance models in the :ref
       To specify a specific learner in :ref:`DoubleML <doubleml_package>` users can pass objects of the class
       `Learner <https://mlr3.mlr-org.com/reference/Learner.html>`_. A fast way to construct these objects is to use the
       `mlr3 <https://mlr3.mlr-org.com/>`_  function `lrn() <https://mlr3.mlr-org.com/reference/mlr_sugar.html>`_.
-      An introduction to learners in `mlr3 <https://mlr3.mlr-org.com/>`_  is provided in the `chapter on learners of the mlr3 book <https://mlr3book.mlr-org.com/basics.html#learners>`_.
+      An introduction to learners in `mlr3 <https://mlr3.mlr-org.com/>`_  is provided in the `chapter on learners of the mlr3 book <https://mlr3book.mlr-org.com/basics.html#sec-learners>`_.
     * It is also possible to pass learners that have been constructed from a pipeline with the `mlr3pipelines <https://mlr3pipelines.mlr-org.com/>`_
       package.
     * The models `DoubleML::DoubleMLIRM <https://docs.doubleml.org/r/stable/reference/DoubleMLIRM.html>`_ and
@@ -420,7 +420,7 @@ The entries in the list specify options during parameter tuning with `mlr3tuning
       `Tuner <https://mlr3tuning.mlr-org.com/reference/Tuner.html>`_ and specifies the tuning algorithm.
       Alternatively, ``algorithm`` can be a ``character()`` that is used as an argument in the wrapper
       `mlr3tuning <https://mlr3tuning.mlr-org.com/>`_ call
-      `tnr(algorithm) <https://mlr3tuning.mlr-org.com/reference/mlr_sugar.html>`_.
+      `tnr(algorithm) <https://mlr3tuning.mlr-org.com/reference/tnr.html>`_.
       `The corresponding chapter in the mlr3book <https://mlr3book.mlr-org.com/optimization.html#sec-model-tuning>`_ illustrates
       how the `Tuner <https://mlr3tuning.mlr-org.com/reference/Tuner.html>`_ class supports grid search, random search,
       generalized simulated annealing and non-linear optimization.

doc/intro/install.rst

@@ -320,7 +320,7 @@ After setting up python and pip as described above use
 
 .. code-block:: Bash
 
-    $ pip install -U DoubleML-0.5.0-py3-none-any.whl
+    $ pip install -U DoubleML-0.5.1-py3-none-any.whl
 
 Python: Building the package from source
 ----------------------------------------

doc/release/release.rst

@@ -5,6 +5,23 @@ Release notes
 
 .. tabbed:: Python
 
+    **DoubleML 0.5.2**
+    - Fix / adapted unit tests which failed in the release of 0.5.1 to conda-forge
+      `#172 <https://github.com/DoubleML/doubleml-for-py/pull/172>`_
+
+    **DoubleML 0.5.1**
+
+    - Store estimated models for nuisance parameters
+      `#159 <https://github.com/DoubleML/doubleml-for-py/pull/159>`_
+    - Bug fix: Overwrite for tune method (introduced for depreciation warning) did not return the tune result
+      `#160 <https://github.com/DoubleML/doubleml-for-py/pull/160>`_
+      `#162 <https://github.com/DoubleML/doubleml-for-py/issues/162>`_
+    - Maintenance
+      `#166 <https://github.com/DoubleML/doubleml-for-py/pull/166>`_
+      `#167 <https://github.com/DoubleML/doubleml-for-py/pull/167>`_
+      `#168 <https://github.com/DoubleML/doubleml-for-py/pull/168>`_
+      `#170 <https://github.com/DoubleML/doubleml-for-py/pull/170>`_
+
     **DoubleML 0.5.0**
 
     - Implement a new score function ``score = 'IV-type'`` for the PLIV model (for details see
@@ -153,6 +170,19 @@ Release notes
 
 .. tabbed:: R
 
+    **DoubleML 0.5.2**
+
+    - Store estimated models for nuisance parameters
+      `#169 <https://github.com/DoubleML/doubleml-for-r/pull/169>`_
+    - New maintainer of the CRAN package DoubleML `@PhilippBach <https://github.com/PhilippBach>`_
+    - Maintenance
+      `#170 <https://github.com/DoubleML/doubleml-for-r/pull/170>`_
+      `#173 <https://github.com/DoubleML/doubleml-for-r/pull/173>`_
+      `#174 <https://github.com/DoubleML/doubleml-for-r/pull/174>`_
+      `#177 <https://github.com/DoubleML/doubleml-for-r/pull/177>`_
+      `#178 <https://github.com/DoubleML/doubleml-for-r/pull/178>`_
+
+
     **DoubleML 0.5.1**
 
     - Fix a CRAN issue (html checks) by regenerating ``.Rd``-files with the newest version of ``roxygen2``.