Fix pytask clean when git is not installed. (#276)

Author: tobiasraabe

docs/source/changes.md

@@ -5,10 +5,16 @@ chronological order. Releases follow [semantic versioning](https://semver.org/)
 releases are available on [PyPI](https://pypi.org/project/pytask) and
 [Anaconda.org](https://anaconda.org/conda-forge/pytask).
 
-## 0.2.2 - 2022-xx-xx
+## 0.2.3 - 2022-xx-xx
+
+- {pull}`276` fixes `pytask clean` when git is not installed. Fixes {issue}`275`.
+
+## 0.2.2 - 2022-05-14
 
 - {pull}`267` fixes the info under the live execution table to show the total number of
   tasks also for pytask-parallel.
+- {pull}`273` reworks `pytask clean` so that it ignores files tracked by git. Resolves
+  {issue}`146`.
 
 ## 0.2.1 - 2022-04-28
 

src/_pytask/clean.py

@@ -20,6 +20,7 @@
 from _pytask.exceptions import CollectionError
 from _pytask.git import get_all_files
 from _pytask.git import get_root
+from _pytask.git import is_git_installed
 from _pytask.nodes import Task
 from _pytask.outcomes import ExitCode
 from _pytask.path import find_common_ancestor
@@ -206,12 +207,15 @@ def _collect_all_paths_known_to_pytask(session: Session) -> set[Path]:
     known_paths.add(session.config["database_filename"])
 
     # Add files tracked by git.
-    git_root = get_root(session.config["root"])
-    if git_root is not None:
-        paths_known_by_git = get_all_files(session.config["root"])
-        absolute_paths_known_by_git = [git_root.joinpath(p) for p in paths_known_by_git]
-        known_paths.update(absolute_paths_known_by_git)
-        known_paths.add(git_root / ".git")
+    if is_git_installed():
+        git_root = get_root(session.config["root"])
+        if git_root is not None:
+            paths_known_by_git = get_all_files(session.config["root"])
+            absolute_paths_known_by_git = [
+                git_root.joinpath(p) for p in paths_known_by_git
+            ]
+            known_paths.update(absolute_paths_known_by_git)
+            known_paths.add(git_root / ".git")
 
     return known_paths
 

src/_pytask/git.py

@@ -1,12 +1,18 @@
 """This module contains all functions related to git."""
 from __future__ import annotations
 
+import shutil
 import subprocess
 from os import PathLike
 from pathlib import Path
 from typing import Any
 
 
+def is_git_installed() -> bool:
+    """Check if git is installed."""
+    return shutil.which("git") is not None
+
+
 def cmd_output(*cmd: str, **kwargs: Any) -> tuple[int, str, str]:
     """Execute a command and capture the output."""
     r = subprocess.run(cmd, capture_output=True, **kwargs)
@@ -48,6 +54,6 @@ def get_root(cwd: PathLike[str] | None) -> Path | None:
         _, stdout, _ = cmd_output("git", "rev-parse", "--show-cdup", cwd=cwd)
         root = Path(cwd) / stdout.strip()
     except subprocess.CalledProcessError:
-        # Either git is not installed or user is not in git repo.
+        # User is not in git repo.
         root = None
     return root

tests/test_clean.py

@@ -221,3 +221,31 @@ def test_dont_remove_files_tracked_by_git(runner, git_project):
     assert "tracked.txt" not in result.output
     assert "in_tracked.txt" not in result.output
     assert ".git" not in result.output
+
+
+@pytest.mark.end_to_end
+def test_clean_git_files_if_git_is_not_installed(monkeypatch, runner, git_project):
+    monkeypatch.setattr(
+        "_pytask.clean.is_git_installed", lambda *x: False  # noqa: U100
+    )
+
+    result = runner.invoke(cli, ["clean", git_project.as_posix()])
+
+    assert result.exit_code == ExitCode.OK
+    assert "tracked.txt" in result.output
+    assert "in_tracked.txt" not in result.output
+    assert ".git" not in result.output
+
+
+@pytest.mark.end_to_end
+def test_clean_git_files_if_git_is_installed_but_git_root_is_not_found(
+    monkeypatch, runner, git_project
+):
+    monkeypatch.setattr("_pytask.clean.get_root", lambda x: None)  # noqa: U100
+
+    result = runner.invoke(cli, ["clean", git_project.as_posix()])
+
+    assert result.exit_code == ExitCode.OK
+    assert "tracked.txt" in result.output
+    assert "in_tracked.txt" not in result.output
+    assert ".git" not in result.output

tests/test_git.py

@@ -0,0 +1,13 @@
+from __future__ import annotations
+
+import pytest
+from _pytask.git import is_git_installed
+
+
+@pytest.mark.parametrize("mock_return, expected", [(True, True), (None, False)])
+def test_is_git_installed(monkeypatch, mock_return, expected):
+    monkeypatch.setattr(
+        "_pytask.git.shutil.which", lambda *x: mock_return  # noqa: U100
+    )
+    result = is_git_installed()
+    assert result is expected