perf: default integer arrays to int32 for ~25% memory reduction

linopy/common.py

@@ -8,7 +8,6 @@
 from __future__ import annotations
 
 import operator
-import os
 from collections.abc import Callable, Generator, Hashable, Iterable, Sequence
 from functools import partial, reduce, wraps
 from pathlib import Path
@@ -18,7 +17,7 @@
 import numpy as np
 import pandas as pd
 import polars as pl
-from numpy import arange, signedinteger
+from numpy import signedinteger
 from xarray import DataArray, Dataset, apply_ufunc, broadcast
 from xarray import align as xr_align
 from xarray.core import dtypes, indexing
@@ -27,6 +26,7 @@
 
 from linopy.config import options
 from linopy.constants import (
+    DEFAULT_LABEL_DTYPE,
     HELPER_DIMS,
     SIGNS,
     SIGNS_alternative,
@@ -333,11 +333,9 @@ def infer_schema_polars(ds: Dataset) -> dict[Hashable, pl.DataType]:
         dict: A dictionary mapping column names to their corresponding Polars data types.
     """
     schema = {}
-    np_major_version = int(np.__version__.split(".")[0])
-    use_int32 = os.name == "nt" and np_major_version < 2
     for name, array in ds.items():
         if np.issubdtype(array.dtype, np.integer):
-            schema[name] = pl.Int32 if use_int32 else pl.Int64
+            schema[name] = pl.Int32 if array.dtype.itemsize <= 4 else pl.Int64
         elif np.issubdtype(array.dtype, np.floating):
             schema[name] = pl.Float64  # type: ignore
         elif np.issubdtype(array.dtype, np.bool_):
@@ -462,7 +460,7 @@ def save_join(*dataarrays: DataArray, integer_dtype: bool = False) -> Dataset:
         )
         arrs = xr_align(*dataarrays, join="outer")
         if integer_dtype:
-            arrs = tuple([ds.fillna(-1).astype(int) for ds in arrs])
+            arrs = tuple([ds.fillna(-1).astype(DEFAULT_LABEL_DTYPE) for ds in arrs])
     return Dataset({ds.name: ds for ds in arrs})
 
 
@@ -523,7 +521,7 @@ def fill_missing_coords(
     # Fill in missing integer coordinates
     for dim in ds.dims:
         if dim not in ds.coords and dim not in skip_dims:
-            ds.coords[dim] = arange(ds.sizes[dim])
+            ds.coords[dim] = np.arange(ds.sizes[dim], dtype=DEFAULT_LABEL_DTYPE)
 
     return ds
 

linopy/constants.py

@@ -33,6 +33,8 @@
     short_LESS_EQUAL: LESS_EQUAL,
 }
 
+DEFAULT_LABEL_DTYPE = np.int32
+
 TERM_DIM = "_term"
 STACKED_TERM_DIM = "_stacked_term"
 GROUPED_TERM_DIM = "_grouped_term"

linopy/constraints.py

@@ -56,6 +56,7 @@
 )
 from linopy.config import options
 from linopy.constants import (
+    DEFAULT_LABEL_DTYPE,
     EQUAL,
     GREATER_EQUAL,
     HELPER_DIMS,
@@ -1071,7 +1072,10 @@ def flat(self) -> pd.DataFrame:
             return pd.DataFrame(columns=["coeffs", "vars", "labels", "key"])
         df = pd.concat(dfs, ignore_index=True)
         unique_labels = df.labels.unique()
-        map_labels = pd.Series(np.arange(len(unique_labels)), index=unique_labels)
+        map_labels = pd.Series(
+            np.arange(len(unique_labels), dtype=DEFAULT_LABEL_DTYPE),
+            index=unique_labels,
+        )
         df["key"] = df.labels.map(map_labels)
         return df
 

linopy/expressions.py

@@ -68,6 +68,7 @@
 from linopy.config import options
 from linopy.constants import (
     CV_DIM,
+    DEFAULT_LABEL_DTYPE,
     EQUAL,
     FACTOR_DIM,
     GREATER_EQUAL,
@@ -279,7 +280,9 @@ def sum(self, use_fallback: bool = False, **kwargs: Any) -> LinearExpression:
 
         def func(ds: Dataset) -> Dataset:
             ds = LinearExpression._sum(ds, str(self.groupby._group_dim))
-            ds = ds.assign_coords({TERM_DIM: np.arange(len(ds._term))})
+            ds = ds.assign_coords(
+                {TERM_DIM: np.arange(len(ds._term), dtype=DEFAULT_LABEL_DTYPE)}
+            )
             return ds
 
         return self.map(func, **kwargs, shortcut=True)
@@ -360,7 +363,9 @@ def __init__(self, data: Dataset | Any | None, model: Model) -> None:
             )
 
         if np.issubdtype(data.vars, np.floating):
-            data = assign_multiindex_safe(data, vars=data.vars.fillna(-1).astype(int))
+            data = assign_multiindex_safe(
+                data, vars=data.vars.fillna(-1).astype(DEFAULT_LABEL_DTYPE)
+            )
         if not np.issubdtype(data.coeffs, np.floating):
             data["coeffs"].values = data.coeffs.values.astype(float)
 
@@ -1137,7 +1142,7 @@ def sanitize(self: GenericExpression) -> GenericExpression:
         linopy.LinearExpression
         """
         if not np.issubdtype(self.vars.dtype, np.integer):
-            return self.assign(vars=self.vars.fillna(-1).astype(int))
+            return self.assign(vars=self.vars.fillna(-1).astype(DEFAULT_LABEL_DTYPE))
 
         return self
 
@@ -1541,12 +1546,12 @@ def _simplify_row(vars_row: np.ndarray, coeffs_row: np.ndarray) -> np.ndarray:
         # Combined has dimensions (.., CV_DIM, TERM_DIM)
 
         # Drop terms where all vars are -1 (i.e., empty terms across all coordinates)
-        vars = combined.isel({CV_DIM: 0}).astype(int)
+        vars = combined.isel({CV_DIM: 0}).astype(DEFAULT_LABEL_DTYPE)
         non_empty_terms = (vars != -1).any(dim=[d for d in vars.dims if d != TERM_DIM])
         combined = combined.isel({TERM_DIM: non_empty_terms})
 
         # Extract vars and coeffs from the combined result
-        vars = combined.isel({CV_DIM: 0}).astype(int)
+        vars = combined.isel({CV_DIM: 0}).astype(DEFAULT_LABEL_DTYPE)
         coeffs = combined.isel({CV_DIM: 1})
 
         # Create new dataset with simplified data

linopy/model.py

@@ -35,6 +35,7 @@
     to_path,
 )
 from linopy.constants import (
+    DEFAULT_LABEL_DTYPE,
     GREATER_EQUAL,
     HELPER_DIMS,
     LESS_EQUAL,
@@ -534,7 +535,14 @@ def add_variables(
 
         start = self._xCounter
         end = start + data.labels.size
-        data.labels.values = np.arange(start, end).reshape(data.labels.shape)
+        if end > np.iinfo(DEFAULT_LABEL_DTYPE).max:
+            raise ValueError(
+                f"Number of labels ({end}) exceeds the maximum value for "
+                f"{DEFAULT_LABEL_DTYPE.__name__} ({np.iinfo(DEFAULT_LABEL_DTYPE).max}). "
+            )
+        data.labels.values = np.arange(start, end, dtype=DEFAULT_LABEL_DTYPE).reshape(
+            data.labels.shape
+        )
         self._xCounter += data.labels.size
 
         if mask is not None:
@@ -713,7 +721,14 @@ def add_constraints(
 
         start = self._cCounter
         end = start + data.labels.size
-        data.labels.values = np.arange(start, end).reshape(data.labels.shape)
+        if end > np.iinfo(DEFAULT_LABEL_DTYPE).max:
+            raise ValueError(
+                f"Number of labels ({end}) exceeds the maximum value for "
+                f"{DEFAULT_LABEL_DTYPE.__name__} ({np.iinfo(DEFAULT_LABEL_DTYPE).max}). "
+            )
+        data.labels.values = np.arange(start, end, dtype=DEFAULT_LABEL_DTYPE).reshape(
+            data.labels.shape
+        )
         self._cCounter += data.labels.size
 
         if mask is not None:

linopy/variables.py

@@ -52,7 +52,7 @@
     to_polars,
 )
 from linopy.config import options
-from linopy.constants import HELPER_DIMS, TERM_DIM
+from linopy.constants import DEFAULT_LABEL_DTYPE, HELPER_DIMS, TERM_DIM
 from linopy.solver_capabilities import SolverFeature, solver_supports
 from linopy.types import (
     ConstantLike,
@@ -1066,7 +1066,9 @@ def ffill(self, dim: str, limit: None = None) -> Variable:
             .map(DataArray.ffill, dim=dim, limit=limit)
             .fillna(self._fill_value)
         )
-        return self.assign_multiindex_safe(labels=data.labels.astype(int))
+        return self.assign_multiindex_safe(
+            labels=data.labels.astype(DEFAULT_LABEL_DTYPE)
+        )
 
     def bfill(self, dim: str, limit: None = None) -> Variable:
         """
@@ -1093,7 +1095,7 @@ def bfill(self, dim: str, limit: None = None) -> Variable:
             .map(DataArray.bfill, dim=dim, limit=limit)
             .fillna(self._fill_value)
         )
-        return self.assign(labels=data.labels.astype(int))
+        return self.assign(labels=data.labels.astype(DEFAULT_LABEL_DTYPE))
 
     def sanitize(self) -> Variable:
         """
@@ -1104,7 +1106,9 @@ def sanitize(self) -> Variable:
         linopy.Variable
         """
         if issubdtype(self.labels.dtype, floating):
-            return self.assign(labels=self.labels.fillna(-1).astype(int))
+            return self.assign(
+                labels=self.labels.fillna(-1).astype(DEFAULT_LABEL_DTYPE)
+            )
         return self
 
     def equals(self, other: Variable) -> bool:
@@ -1525,7 +1529,10 @@ def flat(self) -> pd.DataFrame:
         """
         df = pd.concat([self[k].flat for k in self], ignore_index=True)
         unique_labels = df.labels.unique()
-        map_labels = pd.Series(np.arange(len(unique_labels)), index=unique_labels)
+        map_labels = pd.Series(
+            np.arange(len(unique_labels), dtype=DEFAULT_LABEL_DTYPE),
+            index=unique_labels,
+        )
         df["key"] = df.labels.map(map_labels)
         return df
 

test/test_constraints.py

@@ -34,9 +34,11 @@ def test_constraint_assignment() -> None:
         assert "con0" in getattr(m.constraints, attr)
 
     assert m.constraints.labels.con0.shape == (10, 10)
-    assert m.constraints.labels.con0.dtype == int
+    assert np.issubdtype(m.constraints.labels.con0.dtype, np.integer)
     assert m.constraints.coeffs.con0.dtype in (int, float)
-    assert m.constraints.vars.con0.dtype in (int, float)
+    assert np.issubdtype(m.constraints.vars.con0.dtype, np.integer) or np.issubdtype(
+        m.constraints.vars.con0.dtype, np.floating
+    )
     assert m.constraints.rhs.con0.dtype in (int, float)
 
     assert_conequal(m.constraints.con0, con0)
@@ -88,9 +90,11 @@ def test_anonymous_constraint_assignment() -> None:
         assert "con0" in getattr(m.constraints, attr)
 
     assert m.constraints.labels.con0.shape == (10, 10)
-    assert m.constraints.labels.con0.dtype == int
+    assert np.issubdtype(m.constraints.labels.con0.dtype, np.integer)
     assert m.constraints.coeffs.con0.dtype in (int, float)
-    assert m.constraints.vars.con0.dtype in (int, float)
+    assert np.issubdtype(m.constraints.vars.con0.dtype, np.integer) or np.issubdtype(
+        m.constraints.vars.con0.dtype, np.floating
+    )
     assert m.constraints.rhs.con0.dtype in (int, float)
 
 

test/test_dtypes.py

@@ -0,0 +1,56 @@
+"""Tests for int32 default label dtype."""
+
+import numpy as np
+import pytest
+
+from linopy import Model
+from linopy.constants import DEFAULT_LABEL_DTYPE
+
+
+def test_default_label_dtype_is_int32():
+    assert DEFAULT_LABEL_DTYPE == np.int32
+
+
+def test_variable_labels_are_int32():
+    m = Model()
+    x = m.add_variables(lower=0, upper=10, coords=[range(5)], name="x")
+    assert x.labels.dtype == np.int32
+
+
+def test_constraint_labels_are_int32():
+    m = Model()
+    x = m.add_variables(lower=0, upper=10, coords=[range(5)], name="x")
+    m.add_constraints(x >= 1, name="c")
+    assert m.constraints["c"].labels.dtype == np.int32
+
+
+def test_expression_vars_are_int32():
+    m = Model()
+    x = m.add_variables(lower=0, upper=10, coords=[range(5)], name="x")
+    expr = 2 * x + 1
+    assert expr.vars.dtype == np.int32
+
+
+def test_solve_with_int32_labels():
+    m = Model()
+    x = m.add_variables(lower=0, upper=10, name="x")
+    y = m.add_variables(lower=0, upper=10, name="y")
+    m.add_constraints(x + y <= 15, name="c1")
+    m.add_objective(x + 2 * y, sense="max")
+    m.solve("highs")
+    assert m.objective.value == pytest.approx(25.0)
+
+
+def test_overflow_guard_variables():
+    m = Model()
+    m._xCounter = np.iinfo(np.int32).max - 1
+    with pytest.raises(ValueError, match="exceeds the maximum"):
+        m.add_variables(lower=0, upper=1, coords=[range(5)], name="x")
+
+
+def test_overflow_guard_constraints():
+    m = Model()
+    x = m.add_variables(lower=0, upper=1, coords=[range(5)], name="x")
+    m._cCounter = np.iinfo(np.int32).max - 1
+    with pytest.raises(ValueError, match="exceeds the maximum"):
+        m.add_constraints(x >= 0, name="c")