pydiverse-common 0.3.4__tar.gz → 0.3.5__tar.gz

pydiverse_common-0.3.5/.github/scripts/check_deps.sh

@@ -0,0 +1,22 @@
+#!/usr/bin/env bash
+
+set -euo pipefail
+
+contains_dependency_all=true
+
+while read -r package version; do
+    if [[ $package == "python" ]]; then
+        continue
+    fi
+
+    dependency="${package} ${version}"
+    contains_dependency=$(yq -r ".project.dependencies | map(. == \"${dependency}\") | any" pyproject.toml)
+    if [[ $contains_dependency == "false" ]]; then
+        echo "${dependency} not found in pyproject.toml"
+        contains_dependency_all=false
+    fi
+done < <(yq -r '.dependencies | to_entries | .[] | "\(.key) \(.value)"' pixi.toml)
+
+if [[ $contains_dependency_all == "false" ]]; then
+    exit 1
+fi

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/.github/workflows/tests.yml

@@ -23,7 +23,8 @@ jobs:
         run: pixi run postinstall
       - name: pre-commit
         run: pixi run pre-commit run -a --color=always --show-diff-on-failure
-
+      - name: Check pyproject.toml dependencies
+        run: pixi run check-deps
   test:
     name: pytest
     runs-on: ${{ matrix.os }}

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pydiverse-common
-Version: 0.3.4
+Version: 0.3.5
 Summary: Common functionality shared between pydiverse libraries
 Author: QuantCo, Inc.
 Author-email: Martin Trautmann <windiana@users.sf.net>, Finn Rudolph <finn.rudolph@t-online.de>
@@ -48,7 +48,7 @@ Classifier: Topic :: Database
 Classifier: Topic :: Scientific/Engineering
 Classifier: Topic :: Software Development
 Requires-Python: <3.14,>=3.10
-Requires-Dist: python-box<8.0.0,>=7.3.2
+Requires-Dist: python-box<8,>=7.3.2
 Description-Content-Type: text/markdown
 
 # pydiverse.common

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/pixi.lock

@@ -1180,6 +1180,7 @@ environments:
       - conda: https://conda.anaconda.org/conda-forge/noarch/docformatter-1.7.7-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/exceptiongroup-1.3.0-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/filelock-3.18.0-pyhd8ed1ab_0.conda
+      - conda: https://conda.anaconda.org/conda-forge/linux-64/go-yq-4.45.4-hdd9d35a_0.conda
       - conda: https://conda.anaconda.org/conda-forge/linux-64/icu-75.1-he02047a_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/identify-2.6.12-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/importlib_resources-6.5.2-pyhd8ed1ab_0.conda
@@ -1276,6 +1277,7 @@ environments:
       - conda: https://conda.anaconda.org/conda-forge/noarch/docformatter-1.7.7-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/exceptiongroup-1.3.0-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/filelock-3.18.0-pyhd8ed1ab_0.conda
+      - conda: https://conda.anaconda.org/conda-forge/linux-aarch64/go-yq-4.45.4-h86ecc28_0.conda
       - conda: https://conda.anaconda.org/conda-forge/linux-aarch64/icu-75.1-hf9b3779_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/identify-2.6.12-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/importlib_resources-6.5.2-pyhd8ed1ab_0.conda
@@ -1370,6 +1372,7 @@ environments:
       - conda: https://conda.anaconda.org/conda-forge/noarch/docformatter-1.7.7-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/exceptiongroup-1.3.0-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/filelock-3.18.0-pyhd8ed1ab_0.conda
+      - conda: https://conda.anaconda.org/conda-forge/osx-64/go-yq-4.45.4-h6e16a3a_0.conda
       - conda: https://conda.anaconda.org/conda-forge/osx-64/icu-75.1-h120a0e1_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/identify-2.6.12-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/importlib_resources-6.5.2-pyhd8ed1ab_0.conda
@@ -1458,6 +1461,7 @@ environments:
       - conda: https://conda.anaconda.org/conda-forge/noarch/docformatter-1.7.7-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/exceptiongroup-1.3.0-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/filelock-3.18.0-pyhd8ed1ab_0.conda
+      - conda: https://conda.anaconda.org/conda-forge/osx-arm64/go-yq-4.45.4-h5505292_0.conda
       - conda: https://conda.anaconda.org/conda-forge/osx-arm64/icu-75.1-hfee45f7_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/identify-2.6.12-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/importlib_resources-6.5.2-pyhd8ed1ab_0.conda
@@ -1548,6 +1552,7 @@ environments:
       - conda: https://conda.anaconda.org/conda-forge/noarch/docformatter-1.7.7-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/exceptiongroup-1.3.0-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/filelock-3.18.0-pyhd8ed1ab_0.conda
+      - conda: https://conda.anaconda.org/conda-forge/win-64/go-yq-4.45.4-h2466b09_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/identify-2.6.12-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/importlib_resources-6.5.2-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/iniconfig-2.0.0-pyhd8ed1ab_1.conda
@@ -7609,6 +7614,57 @@ packages:
   license_family: BSD
   size: 112215
   timestamp: 1718284365403
+- conda: https://conda.anaconda.org/conda-forge/linux-64/go-yq-4.45.4-hdd9d35a_0.conda
+  sha256: 710d614cd37077a1b8cb473d9d18b52e7ad4167a40a480f57afdeb42e658f318
+  md5: 0c3da887d37e71224d0077fbb26d1bc5
+  depends:
+  - __glibc >=2.17
+  - __glibc >=2.17,<3.0.a0
+  - libgcc >=13
+  license: MIT
+  license_family: MIT
+  size: 3962140
+  timestamp: 1746984672469
+- conda: https://conda.anaconda.org/conda-forge/linux-aarch64/go-yq-4.45.4-h86ecc28_0.conda
+  sha256: 5f7f6482aaa817807e43ec5e69967d2c2862687d4b683994b0f170521742c5a9
+  md5: 550e6f0253305a62e99e113974502f14
+  depends:
+  - libgcc >=13
+  license: MIT
+  license_family: MIT
+  size: 3574358
+  timestamp: 1746984625484
+- conda: https://conda.anaconda.org/conda-forge/osx-64/go-yq-4.45.4-h6e16a3a_0.conda
+  sha256: 446b4cf3173b27ba3bad51b57592e0ffd3345ffbae60e04d184b9e70c6172082
+  md5: 34bec47340c1e15660de53e6ac2256ab
+  depends:
+  - __osx >=10.13
+  constrains:
+  - __osx>=10.12
+  license: MIT
+  license_family: MIT
+  size: 3854128
+  timestamp: 1746984683074
+- conda: https://conda.anaconda.org/conda-forge/osx-arm64/go-yq-4.45.4-h5505292_0.conda
+  sha256: 5749df601e7039b023481b3fede7cef428e05620f230e016103cd78373bde04f
+  md5: 9e123f64b7084c63ae3c167e68efcfa7
+  depends:
+  - __osx >=11.0
+  license: MIT
+  license_family: MIT
+  size: 3556744
+  timestamp: 1746984719572
+- conda: https://conda.anaconda.org/conda-forge/win-64/go-yq-4.45.4-h2466b09_0.conda
+  sha256: 7032ee8e93fb4f4af7a793df7d2166d226ae8a0b43580a5b4fa759b19af0e5e2
+  md5: 571bc8198620709a5bcaa824fc467b57
+  depends:
+  - ucrt >=10.0.20348.0
+  - vc >=14.2,<15
+  - vc14_runtime >=14.29.30139
+  license: MIT
+  license_family: MIT
+  size: 8291388
+  timestamp: 1746984722615
 - conda: https://conda.anaconda.org/conda-forge/linux-64/greenlet-3.2.3-py310hf71b8c6_0.conda
   sha256: 9d02287ff70b5734ee7a23acde4012df62f851e787aad4219d4259b591aeb054
   md5: 8455091aed2d84aa239354f67cbdfe74

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/pixi.toml

@@ -41,7 +41,7 @@ Sphinx = ">=7.3.7"
 furo = ">=2023.5.20"
 sphinxcontrib-apidoc = ">=0.3.0"
 myst-parser = ">=2.0.0"
-sphinx-autosummary-accessors = "2023.4.0"
+sphinx-autosummary-accessors = ">=2023.4.0"
 sphinx-click = ">=4.4.0"
 sphinx-autobuild = ">=2021.3.14"
 sphinx-copybutton = ">=0.5.2"
@@ -63,6 +63,7 @@ taplo = "*"
 pre-commit-hooks = "*"
 typos = "*"
 conda-deny = "*"
+go-yq = "*"
 
 [feature.lint.tasks]
 #commitit-install = "commitit install"
@@ -70,6 +71,7 @@ conda-deny = "*"
 pre-commit-install = "pre-commit install"
 pre-commit-run = "pre-commit run -a"
 conda-deny = "conda-deny check"
+check-deps = "sh -c .github/scripts/check_deps.sh"
 
 [feature.py310.dependencies]
 python = "3.10.*"

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "pydiverse-common"
-version = "0.3.4"
+version = "0.3.5"
 description = "Common functionality shared between pydiverse libraries"
 authors = [
   { name = "QuantCo, Inc." },
@@ -25,7 +25,7 @@ classifiers = [
   "Topic :: Software Development",
 ]
 requires-python = ">=3.10,<3.14"
-dependencies = ["python-box>=7.3.2,<8.0.0"]
+dependencies = ["python-box >=7.3.2,<8"]
 
 [tool.hatch.build.targets.wheel]
 packages = ["src/pydiverse"]

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/src/pydiverse/common/__init__.py

@@ -7,6 +7,7 @@ from .dtypes import (
     Decimal,
     Dtype,
     Duration,
+    Enum,
     Float,
     Float32,
     Float64,
@@ -51,5 +52,6 @@ __all__ = [
     "UInt32",
     "UInt64",
     "List",
+    "Enum",
     "PandasBackend",
 ]

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/src/pydiverse/common/dtypes.py

@@ -1,9 +1,9 @@
 # Copyright (c) QuantCo and pydiverse contributors 2025-2025
 # SPDX-License-Identifier: BSD-3-Clause
-from enum import Enum
+import enum
 
 
-class PandasBackend(str, Enum):
+class PandasBackend(str, enum.Enum):
     NUMPY = "numpy"
     ARROW = "arrow"
 
@@ -141,6 +141,9 @@ class Dtype:
         # we don't know any decimal/time/null dtypes in pandas if column is not
         # arrow backed
 
+        if pandas_type.name == "category":
+            return Enum(*pandas_type.categories.to_list())
+
         raise TypeError
 
     @staticmethod
@@ -204,6 +207,8 @@ class Dtype:
 
         if isinstance(polars_type, pl.List):
             return List(Dtype.from_polars(polars_type.inner))
+        if isinstance(polars_type, pl.Enum):
+            return Enum(*polars_type.categories)
 
         return {
             pl.Int64: Int64(),
@@ -224,7 +229,6 @@ class Dtype:
             pl.Date: Date(),
             pl.Null: NullType(),
             pl.Duration: Duration(),
-            pl.Enum: String(),
         }[polars_type.base_type()]
 
     def to_sql(self):
@@ -293,6 +297,9 @@ class Dtype:
         if isinstance(self, List):
             raise TypeError("pandas doesn't have a native list dtype")
 
+        if isinstance(self, Enum):
+            return pd.CategoricalDtype(self.categories)
+
         return {
             Int(): pd.Int64Dtype(),  # we default to 64 bit
             Int8(): pd.Int8Dtype(),
@@ -319,6 +326,9 @@ class Dtype:
         """Convert this Dtype to a PyArrow type."""
         import pyarrow as pa
 
+        if isinstance(self, Enum):
+            return pa.string()
+
         return {
             Int(): pa.int64(),  # we default to 64 bit
             Int8(): pa.int8(),
@@ -463,3 +473,36 @@ class List(Dtype):
         import pyarrow as pa
 
         return pa.list_(self.inner.to_arrow())
+
+
+class Enum(String):
+    def __init__(self, *categories: str):
+        if not all(isinstance(c, str) for c in categories):
+            raise TypeError("arguments for `Enum` must have type `str`")
+        self.categories = list(categories)
+
+    def __eq__(self, rhs):
+        return isinstance(rhs, Enum) and self.categories == rhs.categories
+
+    def __repr__(self) -> str:
+        return f"Enum[{', '.join(repr(c) for c in self.categories)}]"
+
+    def __hash__(self):
+        return hash(tuple(self.categories))
+
+    def to_polars(self):
+        import polars as pl
+
+        return pl.Enum(self.categories)
+
+    def to_sql(self):
+        import sqlalchemy as sqa
+
+        return sqa.String()
+
+    def to_arrow(self):
+        import pyarrow as pa
+
+        # There is also pa.dictionary(), which seems to be kind of similar to an enum.
+        # Maybe it is better to convert to this.
+        return pa.string()

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/tests/dtypes/test_dtype_pandas.py

@@ -21,9 +21,13 @@ from pydiverse.common import (
     Bool,
     Date,
     Datetime,
+    Decimal,
     Dtype,
+    Enum,
+    Float,
     Float32,
     Float64,
+    Int,
     Int8,
     Int16,
     Int32,
@@ -154,8 +158,11 @@ def test_dtype_to_pandas_pyarrow():
 def test_all_types_numpy(type_):
     if type_ is pdc.List:
         type_obj = type_(pdc.Int64())
+    elif type_ is pdc.Enum:
+        type_obj = type_("a", "b", "c")
     else:
         type_obj = type_()
+
     if type_ is pdc.NullType:
         with pytest.raises(TypeError, match="pandas doesn't have a native null dtype"):
             type_obj.to_pandas(PandasBackend.NUMPY)
@@ -166,14 +173,17 @@ def test_all_types_numpy(type_):
         with pytest.raises(TypeError, match="pandas doesn't have a native list dtype"):
             type_obj.to_pandas(PandasBackend.NUMPY)
     else:
+        acceptance_map = {
+            Float: Float64(),
+            Int: Int64(),
+            Decimal: Float64(),
+            Date: Datetime(),
+        }
+
         dst_type = type_obj.to_pandas(PandasBackend.NUMPY)
         back_type = Dtype.from_pandas(dst_type)
-        if type_ is pdc.Decimal:
-            assert isinstance(back_type, pdc.Float64)
-        elif type_ is pdc.Date:
-            assert isinstance(back_type, pdc.Datetime)
-        else:
-            assert isinstance(back_type, type_)
+
+        assert back_type == acceptance_map.get(type_, type_obj)
 
 
 @pytest.mark.skipif(np is None, reason="requires pandas, numpy, and pyarrow")
@@ -184,8 +194,17 @@ def test_all_types_numpy(type_):
 def test_all_types_arrow(type_):
     if type_ is pdc.List:
         type_obj = type_(pdc.Int64())
+    elif type_ is pdc.Enum:
+        type_obj = type_("a", "b", "c")
     else:
         type_obj = type_()
+
+    acceptance_map = {
+        Enum: String(),
+        Float: Float64(),
+        Int: Int64(),
+    }
+
     dst_type = type_obj.to_pandas(PandasBackend.ARROW)
     back_type = Dtype.from_pandas(dst_type)
-    assert isinstance(back_type, type_)
+    assert back_type == acceptance_map.get(type_, type_obj)

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/tests/dtypes/test_dtype_polars.py

@@ -9,8 +9,10 @@ from pydiverse.common import (
     Date,
     Datetime,
     Dtype,
+    Float,
     Float32,
     Float64,
+    Int,
     Int8,
     Int16,
     Int32,
@@ -95,6 +97,12 @@ def test_all_types(type_):
         type_obj = type_(pdc.Int64())
     else:
         type_obj = type_()
+
+    acceptance_map = {
+        Float: Float64(),
+        Int: Int64(),
+    }
+
     dst_type = type_obj.to_polars()
     back_type = Dtype.from_polars(dst_type)
-    assert isinstance(back_type, type_)
+    assert back_type == acceptance_map.get(type_, type_obj)

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/tests/dtypes/test_dtype_pyarrow.py

@@ -15,8 +15,11 @@ from pydiverse.common import (
     Date,
     Datetime,
     Dtype,
+    Enum,
+    Float,
     Float32,
     Float64,
+    Int,
     Int8,
     Int16,
     Int32,
@@ -92,7 +95,7 @@ def test_dtype_to_pyarrow():
     assert_conversion(Datetime(), pa.timestamp("us"))
 
 
-@pytest.mark.skipif(pa is None, reason="requires pandas, numpy, and pyarrow")
+@pytest.mark.skipif(pa is None, reason="requires pyarrow")
 @pytest.mark.parametrize(
     "type_",
     ALL_TYPES,
@@ -100,8 +103,18 @@ def test_dtype_to_pyarrow():
 def test_all_types(type_):
     if type_ is pdc.List:
         type_obj = type_(pdc.Int64())
+    elif type_ is pdc.Enum:
+        type_obj = type_("a", "b", "c")
     else:
         type_obj = type_()
+
+    acceptance_map = {
+        Enum: String(),
+        Float: Float64(),
+        Int: Int64(),
+    }
+
     dst_type = type_obj.to_arrow()
     back_type = Dtype.from_arrow(dst_type)
-    assert isinstance(back_type, type_)
+
+    assert back_type == acceptance_map.get(type_, type_obj)

{pydiverse_common-0.3.4 → pydiverse_common-0.3.5}/tests/dtypes/test_dtype_sqlalchemy.py

@@ -9,8 +9,11 @@ from pydiverse.common import (
     Datetime,
     Decimal,
     Dtype,
+    Enum,
+    Float,
     Float32,
     Float64,
+    Int,
     Int8,
     Int16,
     Int32,
@@ -82,7 +85,7 @@ def test_dtype_to_sqlalchemy():
     assert_conversion(Datetime(), sa.DateTime)
 
 
-@pytest.mark.skipif(sa is None, reason="requires polars")
+@pytest.mark.skipif(sa is None, reason="requires sqlalchemy")
 @pytest.mark.parametrize(
     "type_",
     ALL_TYPES,
@@ -90,19 +93,25 @@ def test_dtype_to_sqlalchemy():
 def test_all_types(type_):
     if type_ is pdc.List:
         type_obj = type_(pdc.Int64())
+    elif type_ is pdc.Enum:
+        type_obj = type_("a", "b", "c")
     else:
         type_obj = type_()
     dst_type = type_obj.to_sql()
     back_type = Dtype.from_sql(dst_type)
     acceptance_map = {
         # SQL is a bit less strict about integer precisions
-        Int8: Int16,
-        UInt8: Int16,
-        UInt16: Int32,
-        UInt32: Int64,
-        UInt64: Int64,
+        Int8: Int16(),
+        UInt8: Int16(),
+        UInt16: Int32(),
+        UInt32: Int64(),
+        UInt64: Int64(),
         # we intentionally fetch Decimal as Float since Decimal is more a relational
         # database thing
-        Decimal: Float64,
+        Decimal: Float64(),
+        Float: Float64(),
+        Int: Int64(),
+        # there is no Enum
+        Enum: String(),
     }
-    assert isinstance(back_type, acceptance_map.get(type_, type_))
+    assert back_type == acceptance_map.get(type_, type_obj)