refactor(dependency): use get_packages_from_distribution

Author: mkniewallner

python/deptry/dependency.py

@@ -1,15 +1,12 @@
 from __future__ import annotations
 
 import logging
-import re
-from contextlib import suppress
 from typing import TYPE_CHECKING
 
-from deptry.compat import importlib_metadata
+from deptry.distribution import get_packages_from_distribution
 
 if TYPE_CHECKING:
     from collections.abc import Sequence
-    from importlib.metadata import Distribution
     from pathlib import Path
 
 
@@ -22,7 +19,6 @@ class Dependency:
         name (str): The name of the dependency.
         definition_file (Path): The path to the file defining the dependency, e.g. 'pyproject.toml'.
           and that can be used to create a variant of the package with a set of extra functionalities.
-        found (bool): Indicates if the dependency has been found in the environment.
         top_levels (set[str]): The top-level module names associated with the dependency.
     """
 
@@ -32,16 +28,11 @@ def __init__(
         definition_file: Path,
         module_names: Sequence[str] | None = None,
     ) -> None:
-        distribution = self.find_distribution(name)
-
         self.name = name
         self.definition_file = definition_file
-        self.found = distribution is not None
-        self.top_levels = self._get_top_levels(name, distribution, module_names)
+        self.top_levels = self._get_top_levels(name, module_names)
 
-    def _get_top_levels(
-        self, name: str, distribution: Distribution | None, module_names: Sequence[str] | None
-    ) -> set[str]:
+    def _get_top_levels(self, name: str, module_names: Sequence[str] | None) -> set[str]:
         """
         Get the top-level module names for a dependency. They are searched for in the following order:
                 1. If `module_names` is defined, simply use those as the top-level modules.
@@ -50,22 +41,16 @@ def _get_top_levels(
 
         Args:
             name: The name of the dependency.
-            distribution: The metadata distribution of the package.
             module_names: If this is given, use these as the top-level modules instead of
                 searching for them in the metadata.
         """
         if module_names is not None:
             return set(module_names)
 
-        if distribution is not None:
-            with suppress(FileNotFoundError):
-                return self._get_top_level_module_names_from_top_level_txt(distribution)
-
-            with suppress(FileNotFoundError):
-                return self._get_top_level_module_names_from_record_file(distribution)
+        if distributions := get_packages_from_distribution(self.name):
+            return distributions
 
-        # No metadata or other configuration has been found. As a fallback
-        # we'll guess the name.
+        # No metadata or other configuration has been found. As a fallback we'll guess the name.
         module_name = name.replace("-", "_").lower()
         logging.warning(
             "Assuming the corresponding module name of package %r is %r. Install the package or configure a"
@@ -80,56 +65,3 @@ def __repr__(self) -> str:
 
     def __str__(self) -> str:
         return f"Dependency '{self.name}' with top-levels: {self.top_levels}."
-
-    @staticmethod
-    def find_distribution(name: str) -> Distribution | None:
-        try:
-            return importlib_metadata.distribution(name)
-        except importlib_metadata.PackageNotFoundError:
-            return None
-
-    @staticmethod
-    def _get_top_level_module_names_from_top_level_txt(distribution: Distribution) -> set[str]:
-        """
-        top-level.txt is a metadata file added by setuptools that looks as follows:
-
-        610faff656c4cfcbb4a3__mypyc
-        _black_version
-        black
-        blackd
-        blib2to3
-
-        This function extracts these names, if a top-level.txt file exists.
-        """
-        metadata_top_levels = distribution.read_text("top_level.txt")
-        if metadata_top_levels is None:
-            raise FileNotFoundError("top_level.txt")
-
-        return {x for x in metadata_top_levels.splitlines() if x}
-
-    @staticmethod
-    def _get_top_level_module_names_from_record_file(distribution: Distribution) -> set[str]:
-        """
-        Get the top-level module names from the RECORD file, whose contents usually look as follows:
-
-            ...
-            ../../../bin/black,sha256=<HASH>,247
-            __pycache__/_black_version.cpython-311.pyc,,
-            _black_version.py,sha256=<HASH>,19
-            black/trans.cpython-39-darwin.so,sha256=<HASH>
-            black/trans.py,sha256=<HASH>
-            blackd/__init__.py,sha256=<HASH>
-            blackd/__main__.py,sha256=<HASH>
-            ...
-
-        So if no file top-level.txt is provided, we can try and extract top-levels from this file, in
-        this case _black_version, black, and blackd.
-        """
-        metadata_records = distribution.read_text("RECORD")
-
-        if metadata_records is None:
-            raise FileNotFoundError("RECORD")
-
-        matches = re.finditer(r"^(?!__)([a-zA-Z0-9-_]+)(?:/|\.py,)", metadata_records, re.MULTILINE)
-
-        return {x.group(1) for x in matches}

tests/unit/test_dependency.py

@@ -1,6 +1,5 @@
 from __future__ import annotations
 
-from importlib.metadata import PackageNotFoundError
 from pathlib import Path
 from unittest.mock import patch
 
@@ -21,80 +20,38 @@ def test_create_default_top_level_if_metadata_not_found() -> None:
     assert dependency.top_levels == {"foo_bar"}
 
 
-def test_read_top_level_from_top_level_txt() -> None:
+def test_get_top_levels_from_distribution() -> None:
     """
-    Read the top-levels.txt file
+    Get the packages from distribution.
     """
 
-    class MockDistribution:
-        def __init__(self) -> None:
-            pass
-
-        def read_text(self, file_name: str) -> str:
-            return "foo\nbar"
-
-    with patch("deptry.dependency.metadata.distribution") as mock:
-        mock.return_value = MockDistribution()
+    with patch("deptry.dependency.get_packages_from_distribution", return_value={"foo", "bar"}):
         dependency = Dependency("Foo-bar", Path("pyproject.toml"))
 
     assert dependency.name == "Foo-bar"
     assert dependency.definition_file == Path("pyproject.toml")
     assert dependency.top_levels == {"foo", "bar"}
 
 
-def test_read_top_level_from_record() -> None:
-    """
-    Verify that if top-level.txt not found, an attempt is made to extract top-level module names from
-    the metadata RECORD file.
-    """
-
-    class MockDistribution:
-        def __init__(self) -> None:
-            pass
-
-        def read_text(self, file_name: str) -> str | None:
-            if file_name == "RECORD":
-                return """\
-../../../bin/black,sha256=<HASH>,247
-__pycache__/_black_version.cpython-311.pyc,,
-_black_version.py,sha256=<HASH>,19
-black/trans.cpython-39-darwin.so,sha256=<HASH>
-black/trans.py,sha256=<HASH>
-blackd/__init__.py,sha256=<HASH>
-blackd/__main__.py,sha256=<HASH>
-                """
-            return None
-
-    with patch("deptry.dependency.metadata.distribution") as mock:
-        mock.return_value = MockDistribution()
-        dependency = Dependency("Foo-bar", Path("pyproject.toml"))
-
-    assert dependency.name == "Foo-bar"
-    assert dependency.definition_file == Path("pyproject.toml")
-    assert dependency.top_levels == {"_black_version", "black", "blackd"}
-
-
-def test_read_top_level_from_predefined() -> None:
+def test_get_top_levels_from_predefined() -> None:
     """
-    Verify that if there are predefined top-level module names it takes
-    precedence over other lookup methods.
+    Verify that if there are predefined top-level module names it take precedence over other lookup methods.
     """
-    with patch("deptry.dependency.metadata.distribution") as mock:
+    with patch("deptry.dependency.get_packages_from_distribution") as mock:
         dependency = Dependency("Foo-bar", Path("pyproject.toml"), module_names=["foo"])
 
     assert dependency.name == "Foo-bar"
     assert dependency.definition_file == Path("pyproject.toml")
     assert dependency.top_levels == {"foo"}
-    mock.return_value.read_text.assert_not_called()
+    mock.assert_not_called()
 
 
-def test_not_predefined_and_not_installed() -> None:
+def test_get_top_levels_fallback() -> None:
     """
     Use the fallback option of translating the package name.
     """
 
-    with patch("deptry.dependency.metadata.distribution") as mock:
-        mock.side_effect = PackageNotFoundError
+    with patch("deptry.dependency.get_packages_from_distribution", return_value=None):
         dependency = Dependency("Foo-bar", Path("pyproject.toml"))
 
     assert dependency.name == "Foo-bar"