MAINT: Add NumPy 2 Compat

ci/azure_template_posix.yml

@@ -64,6 +64,9 @@ jobs:
         python.version: '3.11'
         XXHASH: true
         LM_TEST_COPY_ON_WRITE: 1
+      python312_pre:
+        python.version: '3.12'
+        PIP_PRE: true
     maxParallel: 10
 
   steps:

ci/install-posix.sh

@@ -13,3 +13,9 @@ if [[ -n ${FORMULAIC} ]]; then CMD="$CMD formulaic~=${FORMULAIC}"; fi
 if [[ -n ${XXHASH} ]]; then CMD="$CMD xxhash"; fi
 echo "$CMD"
 eval "$CMD"
+
+if [ "${PIP_PRE}" = true ]; then
+  python -m pip uninstall -y numpy pandas scipy matplotlib statsmodels
+  python -m pip install -i https://pypi.anaconda.org/scientific-python-nightly-wheels/simple numpy pandas scipy matplotlib --upgrade --use-deprecated=legacy-resolver
+  python -m pip install git+https://github.com/statsmodels/statsmodels.git --upgrade --no-build-isolation -v
+fi

linearmodels/compat/pandas.py

@@ -0,0 +1,9 @@
+from packaging.version import parse
+import pandas as pd
+
+PANDAS_VERSION = parse(pd.__version__)
+PD_GTE_21 = not (PANDAS_VERSION <= parse("2.0.99"))
+PD_LT_22 = PANDAS_VERSION <= parse("2.1.99")
+ANNUAL_FREQ = "A-DEC" if PD_LT_22 else "YE-DEC"
+
+__all__ = ["ANNUAL_FREQ", "PD_GTE_21", "PD_LT_22"]

linearmodels/iv/results.py

@@ -764,7 +764,11 @@ def diagnostics(self) -> DataFrame:
         ]
         out_df = out_df[cols]
         for col in out_df:
-            out_df[col] = to_numeric(out_df[col], errors="ignore")
+            try:
+                out_df[col] = to_numeric(out_df[col])
+            except ValueError:
+                # If an error is raised, ignore and keep the column
+                pass
 
         return out_df
 

linearmodels/panel/data.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+from linearmodels.compat.pandas import PD_GTE_21
+
 from collections.abc import Hashable, Sequence
 from itertools import product
 from typing import Literal, Union, cast, overload
@@ -227,7 +229,8 @@ def __init__(
                 if copy:
                     self._frame = self._frame.copy()
             else:
-                self._frame = DataFrame({var_name: x.T.stack(dropna=False)})
+                options = {"future_stack": True} if PD_GTE_21 else {"dropna": False}
+                self._frame = DataFrame({var_name: x.T.stack(**options)})
         elif isinstance(x, np.ndarray):
             if x.ndim not in (2, 3):
                 raise ValueError("2 or 3-d array required for numpy input")
@@ -298,6 +301,8 @@ def drop(self, locs: Series | BoolArray) -> None:
             Boolean array indicating observations to drop with reference to
             the dataframe view of the data
         """
+        if isinstance(locs, Series):
+            locs = np.asarray(locs)
         self._frame = self._frame.loc[~locs.ravel()]
         self._frame = self._minimize_multiindex(self._frame)
         # Reset panel and shape after a drop

linearmodels/panel/utility.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+from linearmodels.compat.pandas import ANNUAL_FREQ
+
 from collections import defaultdict
 from typing import NamedTuple, TypeVar, cast
 
@@ -613,7 +615,7 @@ def generate_panel_data(
         x.flat[locs] = np.nan
 
     entities = [f"firm{i}" for i in range(n)]
-    time = [dt for dt in date_range("1-1-1900", periods=t, freq="A-DEC")]
+    time = [dt for dt in date_range("1-1-1900", periods=t, freq=ANNUAL_FREQ)]
     var_names = [f"x{i}" for i in range(k)]
     if const:
         var_names[1:] = var_names[:-1]

linearmodels/shared/hypotheses.py

@@ -117,13 +117,13 @@ class InvalidTestStatistic(WaldTestStatistic):
 
     def __init__(self, reason: str, *, name: str | None = None) -> None:
         self._reason = reason
-        super().__init__(np.NaN, "", df=1, df_denom=1, name=name)
+        super().__init__(np.nan, "", df=1, df_denom=1, name=name)
         self.dist_name = "None"
 
     @property
     def pval(self) -> float:
-        """Always returns np.NaN"""
-        return np.NaN
+        """Always returns np.nan"""
+        return np.nan
 
     @property
     def critical_values(self) -> None:
@@ -158,13 +158,13 @@ def __init__(self, *, reason: str | None = None, name: str | None = None):
         if reason is None:
             self._reason = "Test is not applicable to model specification"
 
-        super().__init__(np.NaN, "", df=1, df_denom=1, name=name)
+        super().__init__(np.nan, "", df=1, df_denom=1, name=name)
         self.dist_name = "None"
 
     @property
     def pval(self) -> float:
-        """Always returns np.NaN"""
-        return np.NaN
+        """Always returns np.nan"""
+        return np.nan
 
     @property
     def critical_values(self) -> None:

linearmodels/tests/panel/_utility.py

@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+from linearmodels.compat.pandas import ANNUAL_FREQ
+
 from typing import Literal
 
 import numpy as np
@@ -121,7 +123,7 @@ def generate_data(
         return AttrDict(y=y, x=x, w=w, c=c, vc1=vc1, vc2=vc2)
 
     entities = ["firm" + str(i) for i in range(n)]
-    time = date_range("1-1-1900", periods=t, freq="A-DEC")
+    time = date_range("1-1-1900", periods=t, freq=ANNUAL_FREQ)
     var_names = ["x" + str(i) for i in range(k)]
     # y = DataFrame(y, index=time, columns=entities)
     y_df = panel_to_frame(

linearmodels/tests/shared/test_utility.py

@@ -1,3 +1,5 @@
+from linearmodels.compat.pandas import ANNUAL_FREQ
+
 import pickle
 import random
 import string
@@ -224,7 +226,7 @@ def test_panel_to_midf():
             ],
         )
     )
-    times = pd.date_range("1999-12-31", freq="A-DEC", periods=7)
+    times = pd.date_range("1999-12-31", freq=ANNUAL_FREQ, periods=7)
     var_names = [f"x.{i}" for i in range(1, 4)]
     df3 = panel_to_frame(x, var_names, times, entities, True)
     mi = pd.MultiIndex.from_product([times, entities])

pyproject.toml

@@ -4,7 +4,7 @@ requires = [
   "wheel",
   "setuptools_scm[toml]>=7,<8",
   "oldest-supported-numpy",
-  "numpy; python_version>='3.12'",
+  "numpy; python_version>='3.13'",
   "cython>=0.29.34"
 ]
 build-backend = "setuptools.build_meta"