feat: allow to include files in wheel .data directory (#182)

docs/build_config.md

@@ -195,6 +195,29 @@ If neither `includes` and `excludes` is specified, the backend can determine the
     `*.pyc`, `__pycache__/` and `build/` are always excluded.
 
 
+### Wheel data files
+
+You can include additional files that are not normally installed inside site-packages directory, with `tool.pdm.build.wheel-data` table:
+
+```toml
+[tool.pdm.build.wheel-data]
+# Install all files under scripts/ to the $prefix/bin directory
+scripts = ["scripts/*"]
+# Install all files under include/ to the $prefix/include directory recursively, keeping the directory structure
+include = [{path = "include/**/*.h", relative-to = "include/"}]
+```
+
+The key is the name of the install scheme, and should be one of `scripts`, `purelib`, `platlib`, `include`, `platinclude` and `data`.
+And each value should be a list of items, which may contain the following attributes:
+
+- `path`: The path pattern to match the files to be included.
+- `relative-to`: if specified, the relative paths of the matched files will be calculated based on this directory,
+otherwise the files will be flattened and installed directly under the scheme directory.
+
+In both attributes, you can use `${BUILD_DIR}` to refer to the build directory.
+
+These files will be packaged into the `{name}-{version}.data/{scheme}` directory in the wheel distribution.
+
 ## Local build hooks
 
 You can specify a custom script to be executed before the build process, which can be used to generate files or modify the metadata.

src/pdm/backend/config.py

@@ -4,7 +4,7 @@
 import os
 import sys
 from pathlib import Path
-from typing import Any, TypeVar
+from typing import TYPE_CHECKING, Any, TypeVar
 
 from pdm.backend._vendor import tomli_w
 from pdm.backend._vendor.pyproject_metadata import ConfigurationError, StandardMetadata
@@ -19,6 +19,14 @@
 
 T = TypeVar("T")
 
+if TYPE_CHECKING:
+    from typing import TypedDict
+
+    DataSpecDict = TypedDict(
+        "DataSpecDict", {"path": str, "relative-to": str}, total=False
+    )
+    DataSpec = DataSpecDict | str
+
 
 class Config:
     """The project config object for pdm backend.
@@ -257,3 +265,8 @@ def editable_backend(self) -> str:
         - path: the legacy .pth file method(default)
         """
         return self.get("editable-backend", "path")
+
+    @property
+    def wheel_data(self) -> dict[str, list[DataSpec]]:
+        """The wheel data configuration"""
+        return self.get("wheel-data", {})

src/pdm/backend/hooks/base.py

@@ -2,7 +2,7 @@
 
 import dataclasses
 from pathlib import Path
-from typing import TYPE_CHECKING, Any
+from typing import TYPE_CHECKING, Any, Iterable
 
 from pdm.backend.config import Config
 
@@ -60,6 +60,13 @@ def ensure_build_dir(self) -> Path:
             (self.build_dir / ".gitignore").write_text("*\n")
         return self.build_dir
 
+    def expand_paths(self, path: str) -> Iterable[Path]:
+        plib_path = Path(path)
+        if plib_path.parts and plib_path.parts[0] == "${BUILD_DIR}":
+            return self.build_dir.glob(Path(*plib_path.parts[1:]).as_posix())
+
+        return self.root.glob(path)
+
 
 class BuildHookInterface(Protocol):
     """The interface definition for build hooks.

src/pdm/backend/wheel.py

@@ -28,6 +28,10 @@
     to_filename,
 )
 
+SCHEME_NAMES = frozenset(
+    ["purelib", "platlib", "include", "platinclude", "scripts", "data"]
+)
+
 if sys.version_info < (3, 8):
     from importlib_metadata import version as get_version
 else:
@@ -72,6 +76,13 @@ def __init__(
         super().__init__(location, config_settings)
         self.__tag: str | None = None
 
+    def scheme_path(self, name: str, relative: str) -> str:
+        if name not in SCHEME_NAMES:
+            raise ValueError(
+                f"Unknown scheme name {name!r}, must be one of {SCHEME_NAMES}"
+            )
+        return f"{self.name_version}.data/{name}/{relative}"
+
     def _get_platform_tags(self) -> tuple[str | None, str | None, str | None]:
         python_tag: str | None = None
         py_limited_api: str | None = None
@@ -118,6 +129,22 @@ def get_files(self, context: Context) -> Iterable[tuple[str, Path]]:
                 relpath = relpath[len(package_dir) + 1 :]
             yield relpath, path
         yield from self._get_metadata_files(context)
+        yield from self._get_wheel_data(context)
+
+    def _get_wheel_data(self, context: Context) -> Iterable[tuple[str, Path]]:
+        for name, paths in context.config.build_config.wheel_data.items():
+            for path in paths:
+                relative_to: str | None = None
+                if not isinstance(path, str):
+                    relative_to = path.get("relative-to")
+                    path = path["path"]
+                for child in context.expand_paths(path):
+                    relpath = (
+                        child.relative_to(relative_to).as_posix()
+                        if relative_to
+                        else child.name
+                    )
+                    yield self.scheme_path(name, relpath), child
 
     def build_artifact(
         self, context: Context, files: Iterable[tuple[str, Path]]

tests/fixtures/projects/demo-package-include/pyproject.toml

@@ -28,3 +28,7 @@ includes = [
 excludes = [
     "my_package/*.json"
 ]
+source-includes = ["scripts/"]
+
+[tool.pdm.build.wheel-data]
+scripts = ["scripts/*"]

tests/fixtures/projects/demo-package-include/scripts/my_script.sh

@@ -0,0 +1,3 @@
+#!/bin/bash
+
+echo "Hello world"

tests/test_api.py

@@ -84,17 +84,23 @@ def test_build_package_include(tmp_path: Path) -> None:
         assert wheel_name == "demo_package-0.1.0-py3-none-any.whl"
 
         tar_names = get_tarball_names(tmp_path / sdist_name)
-        zip_names = get_wheel_names(tmp_path / wheel_name)
 
         assert "demo_package-0.1.0/my_package/__init__.py" in tar_names
         assert "demo_package-0.1.0/my_package/data.json" not in tar_names
         assert "demo_package-0.1.0/requirements.txt" in tar_names
         assert "demo_package-0.1.0/data_out.json" in tar_names
 
-        assert "my_package/__init__.py" in zip_names
-        assert "my_package/data.json" not in zip_names
-        assert "requirements.txt" in zip_names
-        assert "data_out.json" in zip_names
+        with zipfile.ZipFile(tmp_path / wheel_name) as zf:
+            zip_names = zf.namelist()
+            assert "my_package/__init__.py" in zip_names
+            assert "my_package/data.json" not in zip_names
+            assert "requirements.txt" in zip_names
+            assert "data_out.json" in zip_names
+            assert "demo_package-0.1.0.data/scripts/my_script.sh" in zip_names
+            if os.name != "nt":
+                info = zf.getinfo("demo_package-0.1.0.data/scripts/my_script.sh")
+                filemode = info.external_attr >> 16
+                assert filemode & 0o111
 
 
 def test_namespace_package_by_include(tmp_path: Path) -> None: