Merge pull request #225 from DoubleML/s-restructure-doubleml

Restructure doubleml

Author: SvenKlaassen

.github/workflows/codeql.yml

@@ -24,18 +24,18 @@ jobs:
 
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
 
       - name: Initialize CodeQL
-        uses: github/codeql-action/init@v2
+        uses: github/codeql-action/init@v3
         with:
           languages: ${{ matrix.language }}
           queries: +security-and-quality
 
       - name: Autobuild
-        uses: github/codeql-action/autobuild@v2
+        uses: github/codeql-action/autobuild@v3
 
       - name: Perform CodeQL Analysis
-        uses: github/codeql-action/analyze@v2
+        uses: github/codeql-action/analyze@v3
         with:
           category: "/language:${{ matrix.language }}"

.github/workflows/deploy_pkg.yml

@@ -12,12 +12,12 @@ jobs:
     runs-on: ubuntu-latest
 
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         persist-credentials: false
 
     - name: Install python
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: '3.8'
 
@@ -32,7 +32,7 @@ jobs:
         pip install wheel
         python setup.py sdist bdist_wheel
 
-    - uses: actions/upload-artifact@v3
+    - uses: actions/upload-artifact@v4
       with:
         name: DoubleML-pkg
         path: dist/

.github/workflows/pytest.yml

@@ -30,11 +30,11 @@ jobs:
           - {os: 'ubuntu-latest', python-version: '3.11'}
 
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         fetch-depth: 2
     - name: Set up Python ${{ matrix.config.python-version }}
-      uses: actions/setup-python@v4
+      uses: actions/setup-python@v5
       with:
         python-version: ${{ matrix.config.python-version }}
     - name: Install OpenMP runtime for unit tests with xgboost learners

doubleml/__init__.py

@@ -1,32 +1,33 @@
 from pkg_resources import get_distribution
 
-from .double_ml_plr import DoubleMLPLR
-from .double_ml_pliv import DoubleMLPLIV
-from .double_ml_irm import DoubleMLIRM
-from .double_ml_iivm import DoubleMLIIVM
+from .plm.plr import DoubleMLPLR
+from .plm.pliv import DoubleMLPLIV
+from .irm.irm import DoubleMLIRM
+from .irm.iivm import DoubleMLIIVM
 from .double_ml_data import DoubleMLData, DoubleMLClusterData
-from .double_ml_blp import DoubleMLBLP
-from .double_ml_did import DoubleMLDID
-from .double_ml_did_cs import DoubleMLDIDCS
-from .double_ml_qte import DoubleMLQTE
-from .double_ml_pq import DoubleMLPQ
-from .double_ml_lpq import DoubleMLLPQ
-from .double_ml_cvar import DoubleMLCVAR
-from .double_ml_policytree import DoubleMLPolicyTree
+from .did.did import DoubleMLDID
+from .did.did_cs import DoubleMLDIDCS
+from .irm.qte import DoubleMLQTE
+from .irm.pq import DoubleMLPQ
+from .irm.lpq import DoubleMLLPQ
+from .irm.cvar import DoubleMLCVAR
+
+from .utils.blp import DoubleMLBLP
+from .utils.policytree import DoubleMLPolicyTree
 
 __all__ = ['DoubleMLPLR',
            'DoubleMLPLIV',
            'DoubleMLIRM',
            'DoubleMLIIVM',
            'DoubleMLData',
            'DoubleMLClusterData',
-           'DoubleMLBLP',
            'DoubleMLDID',
            'DoubleMLDIDCS',
            'DoubleMLPQ',
            'DoubleMLQTE',
            'DoubleMLLPQ',
            'DoubleMLCVAR',
+           'DoubleMLBLP',
            'DoubleMLPolicyTree']
 
 __version__ = get_distribution('doubleml').version

doubleml/did/__init__.py

@@ -0,0 +1,11 @@
+"""
+The :mod:`doubleml.did` module implements double machine learning estimates based on difference in differences models.
+"""
+
+from .did import DoubleMLDID
+from .did_cs import DoubleMLDIDCS
+
+__all__ = [
+    "DoubleMLDID",
+    "DoubleMLDIDCS",
+]

doubleml/double_ml_did.py renamed to doubleml/did/did.py

@@ -3,12 +3,12 @@
 from sklearn.utils.multiclass import type_of_target
 import warnings
 
-from .double_ml import DoubleML
-from .double_ml_data import DoubleMLData
-from .double_ml_score_mixins import LinearScoreMixin
+from ..double_ml import DoubleML
+from ..double_ml_data import DoubleMLData
+from ..double_ml_score_mixins import LinearScoreMixin
 
-from ._utils import _dml_cv_predict, _get_cond_smpls, _dml_tune, _trimm
-from ._utils_checks import _check_score, _check_trimming, _check_finite_predictions, _check_is_propensity
+from ..utils._estimation import _dml_cv_predict, _get_cond_smpls, _dml_tune, _trimm
+from ..utils._checks import _check_score, _check_trimming, _check_finite_predictions, _check_is_propensity
 
 
 class DoubleMLDID(LinearScoreMixin, DoubleML):
@@ -117,6 +117,8 @@ def __init__(self,
 
         # set stratication for resampling
         self._strata = self._dml_data.d
+        if draw_sample_splitting:
+            self.draw_sample_splitting()
 
         # check learners
         ml_g_is_classifier = self._check_learner(ml_g, 'ml_g', regressor=True, classifier=True)

doubleml/double_ml_did_cs.py renamed to doubleml/did/did_cs.py

@@ -3,12 +3,12 @@
 from sklearn.utils.multiclass import type_of_target
 import warnings
 
-from .double_ml import DoubleML
-from .double_ml_data import DoubleMLData
-from .double_ml_score_mixins import LinearScoreMixin
+from ..double_ml import DoubleML
+from ..double_ml_data import DoubleMLData
+from ..double_ml_score_mixins import LinearScoreMixin
 
-from ._utils import _dml_cv_predict, _trimm, _get_cond_smpls_2d, _dml_tune
-from ._utils_checks import _check_score, _check_trimming, _check_finite_predictions, _check_is_propensity
+from ..utils._estimation import _dml_cv_predict, _trimm, _get_cond_smpls_2d, _dml_tune
+from ..utils._checks import _check_score, _check_trimming, _check_finite_predictions, _check_is_propensity
 
 
 class DoubleMLDIDCS(LinearScoreMixin, DoubleML):
@@ -117,6 +117,8 @@ def __init__(self,
 
         # set stratication for resampling
         self._strata = self._dml_data.d.reshape(-1, 1) + 2 * self._dml_data.t.reshape(-1, 1)
+        if draw_sample_splitting:
+            self.draw_sample_splitting()
 
         # check learners
         ml_g_is_classifier = self._check_learner(ml_g, 'ml_g', regressor=True, classifier=True)

doubleml/did/tests/__init__.py

@@ -1,7 +1,7 @@
 import numpy as np
 from sklearn.base import clone
 
-from ._utils import fit_predict, fit_predict_proba, tune_grid_search
+from ...tests._utils import fit_predict, fit_predict_proba, tune_grid_search
 from ._utils_did_manual import did_dml1, did_dml2
 
 

doubleml/tests/_utils_did_cs_manual.py renamed to doubleml/did/tests/_utils_did_cs_manual.py

@@ -1,8 +1,8 @@
 import numpy as np
 from sklearn.base import clone
 
-from ._utils_boot import boot_manual, draw_weights
-from ._utils import fit_predict, fit_predict_proba, tune_grid_search
+from ...tests._utils_boot import boot_manual, draw_weights
+from ...tests._utils import fit_predict, fit_predict_proba, tune_grid_search
 
 
 def fit_did(y, x, d,

doubleml/tests/_utils_did_manual.py renamed to doubleml/did/tests/_utils_did_manual.py

@@ -0,0 +1,38 @@
+import numpy as np
+import pytest
+
+from doubleml.datasets import make_did_SZ2020
+
+
+@pytest.fixture(scope='session',
+                params=[(500, 1),
+                        (1000, 1),
+                        (1000, 2)])
+def generate_data_did(request):
+    params = request.param
+    np.random.seed(1111)
+    # setting parameters
+    n = params[0]
+    dpg = params[1]
+
+    # generating data
+    data = make_did_SZ2020(n, dgp_type=dpg, return_type='array')
+
+    return data
+
+
+@pytest.fixture(scope='session',
+                params=[(500, 1),
+                        (1000, 1),
+                        (1000, 2)])
+def generate_data_did_cs(request):
+    params = request.param
+    np.random.seed(1111)
+    # setting parameters
+    n = params[0]
+    dpg = params[1]
+
+    # generating data
+    data = make_did_SZ2020(n, dgp_type=dpg, cross_sectional_data=True, return_type='array')
+
+    return data

doubleml/did/tests/conftest.py

@@ -9,7 +9,7 @@
 
 import doubleml as dml
 
-from ._utils import draw_smpls
+from ...tests._utils import draw_smpls
 from ._utils_did_manual import fit_did, boot_did, fit_sensitivity_elements_did
 
 
@@ -122,14 +122,14 @@ def dml_did_fixture(generate_data_did, learner, score, in_sample_normalization,
 
 @pytest.mark.ci
 def test_dml_did_coef(dml_did_fixture):
-    assert math.isclose(dml_did_fixture['coef'],
+    assert math.isclose(dml_did_fixture['coef'][0],
                         dml_did_fixture['coef_manual'],
                         rel_tol=1e-9, abs_tol=1e-4)
 
 
 @pytest.mark.ci
 def test_dml_did_se(dml_did_fixture):
-    assert math.isclose(dml_did_fixture['se'],
+    assert math.isclose(dml_did_fixture['se'][0],
                         dml_did_fixture['se_manual'],
                         rel_tol=1e-9, abs_tol=1e-4)
 
@@ -189,8 +189,8 @@ def test_dml_did_experimental(generate_data_did, in_sample_normalization, learne
                                             score='experimental',
                                             in_sample_normalization=in_sample_normalization)
     dml_did_obj_with_ml_m.fit()
-    assert math.isclose(dml_did_obj_with_ml_m.coef,
-                        dml_did_obj_without_ml_m.coef,
+    assert math.isclose(dml_did_obj_with_ml_m.coef[0],
+                        dml_did_obj_without_ml_m.coef[0],
                         rel_tol=1e-9, abs_tol=1e-4)
 
     msg = ('A learner ml_m has been provided for score = "experimental" but will be ignored. '

doubleml/tests/test_did.py renamed to doubleml/did/tests/test_did.py

@@ -9,7 +9,7 @@
 
 import doubleml as dml
 
-from ._utils import draw_smpls
+from ...tests._utils import draw_smpls
 from ._utils_did_cs_manual import fit_did_cs, fit_sensitivity_elements_did_cs
 from ._utils_did_manual import boot_did
 
@@ -63,7 +63,8 @@ def dml_did_cs_fixture(generate_data_did_cs, learner, score, in_sample_normaliza
 
     np.random.seed(3141)
     n_obs = len(y)
-    all_smpls = draw_smpls(n_obs, n_folds, n_rep=1, groups=d)
+
+    all_smpls = draw_smpls(n_obs, n_folds, n_rep=1, groups=d+2*t)
     obj_dml_data = dml.DoubleMLData.from_arrays(x, y, d, t=t)
 
     np.random.seed(3141)
@@ -122,14 +123,14 @@ def dml_did_cs_fixture(generate_data_did_cs, learner, score, in_sample_normaliza
 
 @pytest.mark.ci
 def test_dml_did_cs_coef(dml_did_cs_fixture):
-    assert math.isclose(dml_did_cs_fixture['coef'],
+    assert math.isclose(dml_did_cs_fixture['coef'][0],
                         dml_did_cs_fixture['coef_manual'],
                         rel_tol=1e-9, abs_tol=1e-4)
 
 
 @pytest.mark.ci
 def test_dml_did_cs_se(dml_did_cs_fixture):
-    assert math.isclose(dml_did_cs_fixture['se'],
+    assert math.isclose(dml_did_cs_fixture['se'][0],
                         dml_did_cs_fixture['se_manual'],
                         rel_tol=1e-9, abs_tol=1e-4)
 
@@ -189,8 +190,8 @@ def test_dml_did_cs_experimental(generate_data_did_cs, in_sample_normalization,
                                               score='experimental',
                                               in_sample_normalization=in_sample_normalization)
     dml_did_obj_with_ml_m.fit()
-    assert math.isclose(dml_did_obj_with_ml_m.coef,
-                        dml_did_obj_without_ml_m.coef,
+    assert math.isclose(dml_did_obj_with_ml_m.coef[0],
+                        dml_did_obj_without_ml_m.coef[0],
                         rel_tol=1e-9, abs_tol=1e-4)
 
     msg = ('A learner ml_m has been provided for score = "experimental" but will be ignored. '

doubleml/tests/test_did_cs.py renamed to doubleml/did/tests/test_did_cs.py

@@ -9,7 +9,7 @@
 
 import doubleml as dml
 
-from ._utils import draw_smpls
+from ...tests._utils import draw_smpls
 from ._utils_did_manual import boot_did
 from ._utils_did_cs_manual import fit_did_cs, tune_nuisance_did_cs
 
@@ -77,15 +77,22 @@ def dml_did_cs_fixture(generate_data_did_cs, learner_g, learner_m, score, in_sam
     ml_g = clone(learner_g)
     ml_m = clone(learner_m)
 
+    n_obs = len(y)
+    all_smpls = draw_smpls(n_obs, n_folds, n_rep=1, groups=d+2*t)
+
     np.random.seed(3141)
     obj_dml_data = dml.DoubleMLData.from_arrays(x, y, d, t=t)
     dml_did_cs_obj = dml.DoubleMLDIDCS(obj_dml_data,
                                        ml_g, ml_m,
                                        n_folds,
                                        score=score,
                                        in_sample_normalization=in_sample_normalization,
-                                       dml_procedure=dml_procedure)
+                                       dml_procedure=dml_procedure,
+                                       draw_sample_splitting=False)
+    # synchronize the sample splitting
+    dml_did_cs_obj.set_sample_splitting(all_smpls=all_smpls)
 
+    np.random.seed(3141)
     # tune hyperparameters
     tune_res = dml_did_cs_obj.tune(par_grid, tune_on_folds=tune_on_folds,
                                    n_folds_tune=n_folds_tune,
@@ -95,8 +102,6 @@ def dml_did_cs_fixture(generate_data_did_cs, learner_g, learner_m, score, in_sam
     dml_did_cs_obj.fit()
 
     np.random.seed(3141)
-    n_obs = len(y)
-    all_smpls = draw_smpls(n_obs, n_folds)
     smpls = all_smpls[0]
 
     if tune_on_folds:
@@ -152,14 +157,14 @@ def dml_did_cs_fixture(generate_data_did_cs, learner_g, learner_m, score, in_sam
 
 @pytest.mark.ci
 def test_dml_did_cs_coef(dml_did_cs_fixture):
-    assert math.isclose(dml_did_cs_fixture['coef'],
+    assert math.isclose(dml_did_cs_fixture['coef'][0],
                         dml_did_cs_fixture['coef_manual'],
                         rel_tol=1e-9, abs_tol=1e-4)
 
 
 @pytest.mark.ci
 def test_dml_did_cs_se(dml_did_cs_fixture):
-    assert math.isclose(dml_did_cs_fixture['se'],
+    assert math.isclose(dml_did_cs_fixture['se'][0],
                         dml_did_cs_fixture['se_manual'],
                         rel_tol=1e-9, abs_tol=1e-4)
 

doubleml/tests/test_did_cs_tune.py renamed to doubleml/did/tests/test_did_cs_tune.py

@@ -5,7 +5,7 @@
 from doubleml import DoubleMLDID
 from doubleml.datasets import make_did_SZ2020
 from doubleml.utils import DMLDummyRegressor, DMLDummyClassifier
-from ._utils import draw_smpls
+from ...tests._utils import draw_smpls
 
 
 @pytest.fixture(scope="module", params=["observational", "experimental"])