Merge pull request #595 from bashtage/rls-6

DOC: Update for release 6.0

ci/install-posix.sh

@@ -17,6 +17,5 @@ eval "$CMD"
 
 if [ "${PIP_PRE}" = true ]; then
   python -m pip uninstall -y numpy pandas scipy matplotlib statsmodels xarray
-  python -m pip install -i https://pypi.anaconda.org/scientific-python-nightly-wheels/simple numpy pandas scipy matplotlib xarray --upgrade --use-deprecated=legacy-resolver
-  python -m pip install git+https://github.com/statsmodels/statsmodels.git --upgrade --no-build-isolation -v
+  python -m pip install -i https://pypi.anaconda.org/scientific-python-nightly-wheels/simple numpy pandas scipy matplotlib xarray statsmodels --upgrade --use-deprecated=legacy-resolver
 fi

doc/source/changes/6.0.rst

@@ -0,0 +1,21 @@
+Version 6.0
+------------
+
+* Increased minimums:
+
+  - Python: 3.9
+  - formulaic: 1.0.0
+  - NumPy: 1.22.3
+  - SciPy: 1.8.0
+  - pandas: 1.4.0
+  - statsmodels: 0.13.0
+
+* The key feature of this release is compatibility with NumPy 2. linearmodels
+  wheels are built using NumPy 2.0.0rc1 (or later) and can run on any version
+  of NumPy 1.22.3 or later, including NumPy 2.0.0.
+* Improved compatibility with fuure changes in pandas 3.0.0.
+
+.. note::
+
+   In order to use NumPy 2, the environment must consist of packages
+   that have been built against NumPy 2.0.0rc1 or later.

linearmodels/iv/results.py

@@ -516,7 +516,7 @@ def _out_of_sample(
             )
         pred = self.model.predict(self.params, exog=exog, endog=endog, data=data)
         if not missing:
-            pred = pred.loc[pred.notnull().all(1)]
+            pred = pred.loc[pred.notnull().all(axis=1)]
         return pred
 
     def predict(

linearmodels/panel/model.py

@@ -3055,7 +3055,7 @@ def single(z: DataFrame) -> Series:
         else:
             avg_adj_r2 = np.nan
         all_params = all_params.iloc[:, 1:-2]
-        params = np.asarray(all_params.mean(0).values[:, None], dtype=float)
+        params = np.asarray(all_params.mean(axis=0).values[:, None], dtype=float)
 
         wy = np.asarray(wy_df)
         wx = np.asarray(wx_df)

linearmodels/panel/results.py

@@ -488,7 +488,7 @@ def _out_of_sample(
             )
         pred = self.model.predict(self.params, exog=exog, data=data, context=context)
         if not missing:
-            pred = pred.loc[pred.notnull().all(1)]
+            pred = pred.loc[pred.notnull().all(axis=1)]
         return pred
 
     def predict(

linearmodels/tests/panel/test_data.py

@@ -400,13 +400,13 @@ def test_dummies(mi_df):
     data = PanelData(mi_df)
     edummy = data.dummies()
     assert edummy.shape == (77, 11)
-    assert np.all(edummy.sum(0) == 7)
+    assert np.all(edummy.sum(axis=0) == 7)
     tdummy = data.dummies(group="time")
     assert tdummy.shape == (77, 7)
-    assert np.all(tdummy.sum(0) == 11)
+    assert np.all(tdummy.sum(axis=0) == 11)
     tdummy_drop = data.dummies(group="time", drop_first=True)
     assert tdummy_drop.shape == (77, 6)
-    assert np.all(tdummy.sum(0) == 11)
+    assert np.all(tdummy.sum(axis=0) == 11)
     with pytest.raises(ValueError):
         data.dummies("unknown")
 

linearmodels/tests/panel/test_formula.py

@@ -352,4 +352,4 @@ def test_escaped_variable_name():
     )
     mod = PanelOLS.from_formula("`var a` ~ 1", data=data)
     res = mod.fit()
-    assert_allclose(res.params, data.mean(0))
+    assert_allclose(res.params, data.mean(axis=0))

requirements-dev.txt

@@ -1,7 +1,7 @@
 xarray>=0.16
 mypy>=1.3
-black[jupyter]==23.11.0
-pytest>=7.3.0
+black[jupyter]==24.4.0
+pytest>=7.3.0,<8
 isort>=5.12
 ipython
 matplotlib

requirements-test.txt

@@ -1,10 +1,10 @@
-black[jupyter]==23.11.0
+black[jupyter]==24.4.0
 coverage
 flake8
 isort
 colorama
 matplotlib
-pytest>=7.3.0
+pytest>=7.3.0,<8
 pytest-xdist
 pytest-cov
 seaborn

requirements.txt

@@ -1,11 +1,10 @@
 numpy>=1.22.3,<3
 pandas>=1.4.0
 scipy>=1.8.0
-statsmodels>=0.12.0
+statsmodels>=0.13.0
 mypy_extensions>=0.4
 Cython>=3.0.10
 pyhdfe>=0.1
 formulaic>=1.0.0
 # versioning
 setuptools_scm[toml]>=8.0.0,<9.0.0
-