feat(python_file_finder): improve ignore handling (#908)

* test(functional): more tests for gitignore

* test(python_file_finder): add tests for gitignore

* feat(python_file_finder): only use gitignore files when using git

* test(functional): use `.ignore` in non-git projects

src/python_file_finder.rs

@@ -61,9 +61,8 @@ fn build_walker(
 
     walk_builder
         .types(build_types(ignore_notebooks).unwrap())
+        .standard_filters(use_git_ignore)
         .hidden(false)
-        .git_ignore(use_git_ignore)
-        .require_git(false)
         .filter_entry(move |entry| entry_satisfies_predicate(entry, re.as_ref()))
         .build()
 }

tests/fixtures/project_with_gitignore/.gitignore

@@ -1 +1,3 @@
+/build
 foobar.py
+/src/barfoo.py

tests/fixtures/project_with_gitignore/src/.gitignore

@@ -0,0 +1,2 @@
+/baz.py
+/src/bar.py

tests/fixtures/project_with_gitignore/src/barfoo.py

@@ -0,0 +1 @@
+import hello

tests/fixtures/project_with_gitignore/src/baz.py

@@ -0,0 +1 @@
+import hej

tests/fixtures/project_with_src_directory/.ignore

@@ -0,0 +1 @@
+src/this_file_is_ignored.py

tests/functional/cli/test_cli_gitignore.py

@@ -14,9 +14,13 @@
 
 
 @pytest.mark.xdist_group(name=Project.GITIGNORE)
-def test_cli_gitignore_is_used(pip_venv_factory: PipVenvFactory) -> None:
+def test_cli_gitignore_used(pip_venv_factory: PipVenvFactory) -> None:
     with pip_venv_factory(Project.GITIGNORE) as virtual_env:
         issue_report = f"{uuid.uuid4()}.json"
+
+        # Simulate the fact that the project is a git repository.
+        Path(".git").mkdir(exist_ok=True)
+
         result = virtual_env.run(f"deptry . -o {issue_report}")
 
         assert result.returncode == 1
@@ -61,9 +65,64 @@ def test_cli_gitignore_is_used(pip_venv_factory: PipVenvFactory) -> None:
 
 
 @pytest.mark.xdist_group(name=Project.GITIGNORE)
-def test_cli_gitignore_is_not_used(pip_venv_factory: PipVenvFactory) -> None:
+def test_cli_gitignore_used_for_non_root_directory(pip_venv_factory: PipVenvFactory) -> None:
     with pip_venv_factory(Project.GITIGNORE) as virtual_env:
         issue_report = f"{uuid.uuid4()}.json"
+
+        # Simulate the fact that the project is a git repository.
+        Path(".git").mkdir(exist_ok=True)
+
+        result = virtual_env.run(f"deptry src -o {issue_report}")
+
+        assert result.returncode == 1
+        assert get_issues_report(Path(issue_report)) == [
+            {
+                "error": {
+                    "code": "DEP002",
+                    "message": "'requests' defined as a dependency but not used in the codebase",
+                },
+                "module": "requests",
+                "location": {
+                    "file": str(Path("pyproject.toml")),
+                    "line": None,
+                    "column": None,
+                },
+            },
+            {
+                "error": {
+                    "code": "DEP002",
+                    "message": "'mypy' defined as a dependency but not used in the codebase",
+                },
+                "module": "mypy",
+                "location": {
+                    "file": str(Path("pyproject.toml")),
+                    "line": None,
+                    "column": None,
+                },
+            },
+            {
+                "error": {
+                    "code": "DEP002",
+                    "message": "'pytest' defined as a dependency but not used in the codebase",
+                },
+                "module": "pytest",
+                "location": {
+                    "file": str(Path("pyproject.toml")),
+                    "line": None,
+                    "column": None,
+                },
+            },
+        ]
+
+
+@pytest.mark.xdist_group(name=Project.GITIGNORE)
+def test_cli_gitignore_not_used_when_using_exclude(pip_venv_factory: PipVenvFactory) -> None:
+    with pip_venv_factory(Project.GITIGNORE) as virtual_env:
+        issue_report = f"{uuid.uuid4()}.json"
+
+        # Simulate the fact that the project is a git repository.
+        Path(".git").mkdir(exist_ok=True)
+
         result = virtual_env.run(f"deptry . --exclude build/|src/bar.py -o {issue_report}")
 
         assert result.returncode == 1
@@ -104,4 +163,28 @@ def test_cli_gitignore_is_not_used(pip_venv_factory: PipVenvFactory) -> None:
                     "column": None,
                 },
             },
+            {
+                "error": {
+                    "code": "DEP001",
+                    "message": "'hello' imported but missing from the dependency definitions",
+                },
+                "module": "hello",
+                "location": {
+                    "file": str(Path("src/barfoo.py")),
+                    "line": 1,
+                    "column": 8,
+                },
+            },
+            {
+                "error": {
+                    "code": "DEP001",
+                    "message": "'hej' imported but missing from the dependency definitions",
+                },
+                "module": "hej",
+                "location": {
+                    "file": str(Path("src/baz.py")),
+                    "line": 1,
+                    "column": 8,
+                },
+            },
         ]

tests/unit/test_python_file_finder.py

@@ -174,3 +174,110 @@ def test_duplicates_are_removed(tmp_path: Path) -> None:
         files = get_all_python_files_in((Path(), Path()), exclude=(), extend_exclude=(), using_default_exclude=False)
 
         assert sorted(files) == [Path("dir/subdir/file1.py")]
+
+
+def test_gitignore_used_in_git_project(tmp_path: Path) -> None:
+    """Test that gitignore files are respected when project uses git."""
+    git_project_path = tmp_path / "git_project"
+    git_project_path.mkdir()
+
+    # Simulate the presence of a gitignore outside the git project, that should not be used.
+    with (tmp_path / ".gitignore").open("w") as f:
+        f.write("*")
+
+    with run_within_dir(tmp_path / git_project_path):
+        # Simulate the fact that the project is a git repository.
+        Path(".git").mkdir()
+
+        create_files([
+            Path("file1.py"),
+            Path("file2.py"),
+            Path("file3.py"),
+            Path("dir1/file.py"),
+            Path("dir2/file2.py"),
+            Path("dir3/file3.py"),
+            Path("dir3/file4.py"),
+        ])
+
+        with Path(".gitignore").open("w") as f:
+            f.write("""/file1.py
+/dir1/file.py
+file2.py""")
+
+        with Path("dir3/.gitignore").open("w") as f:
+            f.write("/file3.py")
+
+        files = get_all_python_files_in((Path(),), exclude=(), extend_exclude=(), using_default_exclude=True)
+
+        assert sorted(files) == [
+            Path("dir3/file4.py"),
+            Path("file3.py"),
+        ]
+
+
+def test_gitignore_ignored_when_not_using_default_exclude(tmp_path: Path) -> None:
+    """Test that gitignore files are ignored when project uses git but does not use default exclude."""
+    git_project_path = tmp_path / "git_project"
+    git_project_path.mkdir()
+
+    # Simulate the presence of a gitignore outside the git project, that should not be used.
+    with (tmp_path / ".gitignore").open("w") as f:
+        f.write("*")
+
+    with run_within_dir(tmp_path / git_project_path):
+        # Simulate the fact that the project is a git repository.
+        Path(".git").mkdir()
+
+        create_files([
+            Path("file1.py"),
+            Path("file2.py"),
+            Path("file3.py"),
+            Path("dir1/file.py"),
+            Path("dir2/file2.py"),
+            Path("dir3/file3.py"),
+            Path("dir3/file4.py"),
+        ])
+
+        with Path(".gitignore").open("w") as f:
+            f.write("""/file1.py
+/dir1/file.py
+file2.py""")
+
+        with Path("dir3/.gitignore").open("w") as f:
+            f.write("/file3.py")
+
+        files = get_all_python_files_in(
+            (Path(),),
+            exclude=("file3.py",),
+            extend_exclude=(),
+            using_default_exclude=False,
+        )
+
+        assert sorted(files) == [
+            Path("dir1/file.py"),
+            Path("dir2/file2.py"),
+            Path("dir3/file3.py"),
+            Path("dir3/file4.py"),
+            Path("file1.py"),
+            Path("file2.py"),
+        ]
+
+
+def test_gitignore_ignored_in_non_git_project(tmp_path: Path) -> None:
+    """Test that gitignore files are ignored when project does not use git."""
+    with run_within_dir(tmp_path):
+        create_files([
+            Path("file1.py"),
+            Path("file2.py"),
+        ])
+
+        # This will be ignored, since project is not a git project.
+        with Path(".gitignore").open("w") as f:
+            f.write("/file1.py")
+
+        files = get_all_python_files_in((Path(),), exclude=(), extend_exclude=(), using_default_exclude=True)
+
+        assert sorted(files) == [
+            Path("file1.py"),
+            Path("file2.py"),
+        ]