from __future__ import annotations import asyncio import sys import tempfile import threading import time import unittest from pathlib import Path import httpx sys.path.insert(0, str(Path(__file__).resolve().parents[3])) from backend.app.dependencies import get_duplicate_task_service, get_task_service from backend.app.db.task_repository import TaskRepository from backend.app.fs.filesystem_adapter import FilesystemAdapter from backend.app.main import app from backend.app.security.path_guard import PathGuard from backend.app.services.duplicate_task_service import DuplicateTaskService from backend.app.services.task_service import TaskService from backend.app.tasks_runner import TaskRunner class FailOnSecondCopyFilesystemAdapter(FilesystemAdapter): def __init__(self) -> None: super().__init__() self._copy_calls = 0 def copy_file(self, source: str, destination: str, on_progress: callable | None = None) -> None: self._copy_calls += 1 if self._copy_calls == 2: raise OSError("forced duplicate failure") super().copy_file(source=source, destination=destination, on_progress=on_progress) class BlockingDuplicateFilesystemAdapter(FilesystemAdapter): def __init__(self) -> None: super().__init__() self.entered = threading.Event() self.release = threading.Event() def copy_file(self, source: str, destination: str, on_progress: callable | None = None) -> None: self.entered.set() self.release.wait(timeout=2.0) super().copy_file(source=source, destination=destination, on_progress=on_progress) class DuplicateApiGoldenTest(unittest.TestCase): def setUp(self) -> None: self.temp_dir = tempfile.TemporaryDirectory() self.root = Path(self.temp_dir.name) / "root" self.root.mkdir(parents=True, exist_ok=True) self.repo = TaskRepository(str(Path(self.temp_dir.name) / "tasks.db")) path_guard = PathGuard({"storage1": str(self.root), "storage2": str(self.root)}) self._set_services(path_guard=path_guard, filesystem=FilesystemAdapter()) def tearDown(self) -> None: app.dependency_overrides.clear() self.temp_dir.cleanup() def _set_services(self, path_guard: PathGuard, filesystem: FilesystemAdapter) -> None: runner = TaskRunner(repository=self.repo, filesystem=filesystem) duplicate_service = DuplicateTaskService(path_guard=path_guard, repository=self.repo, runner=runner) task_service = TaskService(repository=self.repo) async def _override_duplicate_service() -> DuplicateTaskService: return duplicate_service async def _override_task_service() -> TaskService: return task_service app.dependency_overrides[get_duplicate_task_service] = _override_duplicate_service app.dependency_overrides[get_task_service] = _override_task_service def _request(self, method: str, url: str, payload: dict | None = None) -> httpx.Response: async def _run() -> httpx.Response: transport = httpx.ASGITransport(app=app) async with httpx.AsyncClient(transport=transport, base_url="http://testserver") as client: if method == "POST": return await client.post(url, json=payload) return await client.get(url) return asyncio.run(_run()) def _wait_task(self, task_id: str, timeout_s: float = 2.0) -> dict: deadline = time.time() + timeout_s while time.time() < deadline: response = self._request("GET", f"/api/tasks/{task_id}") body = response.json() if body["status"] in {"completed", "failed", "cancelled"}: return body time.sleep(0.02) self.fail("task did not reach terminal state in time") def _wait_for_status(self, task_id: str, statuses: set[str], timeout_s: float = 2.0) -> dict: deadline = time.time() + timeout_s while time.time() < deadline: response = self._request("GET", f"/api/tasks/{task_id}") body = response.json() if body["status"] in statuses: return body time.sleep(0.02) self.fail(f"task did not reach one of {sorted(statuses)} in time") def test_duplicate_single_file_success(self) -> None: (self.root / "note.txt").write_text("hello", encoding="utf-8") response = self._request("POST", "/api/files/duplicate", {"paths": ["storage1/note.txt"]}) self.assertEqual(response.status_code, 202) detail = self._wait_task(response.json()["task_id"]) self.assertEqual(detail["operation"], "duplicate") self.assertEqual(detail["status"], "completed") self.assertEqual(detail["done_items"], 1) self.assertEqual(detail["total_items"], 1) self.assertEqual(detail["source"], "storage1/note.txt") self.assertEqual(detail["destination"], "storage1/note copy.txt") self.assertEqual((self.root / "note copy.txt").read_text(encoding="utf-8"), "hello") def test_duplicate_single_directory_success(self) -> None: (self.root / "Folder" / "nested").mkdir(parents=True) (self.root / "Folder" / "alpha.txt").write_text("A", encoding="utf-8") (self.root / "Folder" / "nested" / "beta.txt").write_text("B", encoding="utf-8") response = self._request("POST", "/api/files/duplicate", {"paths": ["storage1/Folder"]}) self.assertEqual(response.status_code, 202) detail = self._wait_task(response.json()["task_id"]) self.assertEqual(detail["status"], "completed") self.assertEqual(detail["done_items"], 2) self.assertEqual(detail["total_items"], 2) self.assertTrue((self.root / "Folder copy").is_dir()) self.assertEqual((self.root / "Folder copy" / "alpha.txt").read_text(encoding="utf-8"), "A") self.assertEqual((self.root / "Folder copy" / "nested" / "beta.txt").read_text(encoding="utf-8"), "B") def test_duplicate_multi_select_success(self) -> None: (self.root / "a.txt").write_text("A", encoding="utf-8") (self.root / "docs" / "nested").mkdir(parents=True) (self.root / "docs" / "nested" / "b.txt").write_text("B", encoding="utf-8") response = self._request( "POST", "/api/files/duplicate", {"paths": ["storage1/a.txt", "storage1/docs"]}, ) self.assertEqual(response.status_code, 202) detail = self._wait_task(response.json()["task_id"]) self.assertEqual(detail["status"], "completed") self.assertEqual(detail["done_items"], 2) self.assertEqual(detail["total_items"], 2) self.assertEqual(detail["source"], "2 items") self.assertEqual(detail["destination"], "same directory") self.assertEqual((self.root / "a copy.txt").read_text(encoding="utf-8"), "A") self.assertEqual((self.root / "docs copy" / "nested" / "b.txt").read_text(encoding="utf-8"), "B") def test_duplicate_multi_select_cancelled_after_current_item_finishes(self) -> None: blocking_fs = BlockingDuplicateFilesystemAdapter() path_guard = PathGuard({"storage1": str(self.root), "storage2": str(self.root)}) self._set_services(path_guard=path_guard, filesystem=blocking_fs) (self.root / "a.txt").write_text("A", encoding="utf-8") (self.root / "b.txt").write_text("B", encoding="utf-8") response = self._request( "POST", "/api/files/duplicate", {"paths": ["storage1/a.txt", "storage1/b.txt"]}, ) task_id = response.json()["task_id"] self.assertTrue(blocking_fs.entered.wait(timeout=2.0)) running = self._wait_for_status(task_id, {"running"}) self.assertEqual(running["current_item"], "a.txt") cancel_response = self._request("POST", f"/api/tasks/{task_id}/cancel") self.assertEqual(cancel_response.status_code, 200) self.assertEqual(cancel_response.json()["status"], "cancelling") blocking_fs.release.set() detail = self._wait_task(task_id) self.assertEqual(detail["status"], "cancelled") self.assertEqual(detail["done_items"], 1) self.assertEqual(detail["total_items"], 2) self.assertTrue((self.root / "a copy.txt").exists()) self.assertFalse((self.root / "b copy.txt").exists()) def test_duplicate_collision_resolution_for_files_and_directories(self) -> None: (self.root / "report.txt").write_text("R", encoding="utf-8") (self.root / "report copy.txt").write_text("existing", encoding="utf-8") (self.root / "report copy 2.txt").write_text("existing", encoding="utf-8") (self.root / "Album").mkdir() (self.root / "Album copy").mkdir() (self.root / "Album copy 2").mkdir() file_response = self._request("POST", "/api/files/duplicate", {"paths": ["storage1/report.txt"]}) dir_response = self._request("POST", "/api/files/duplicate", {"paths": ["storage1/Album"]}) self.assertEqual(self._wait_task(file_response.json()["task_id"])["destination"], "storage1/report copy 3.txt") self.assertEqual(self._wait_task(dir_response.json()["task_id"])["destination"], "storage1/Album copy 3") self.assertTrue((self.root / "report copy 3.txt").exists()) self.assertTrue((self.root / "Album copy 3").is_dir()) def test_duplicate_skips_top_level_macos_sidecar(self) -> None: (self.root / "._note.txt").write_text("sidecar", encoding="utf-8") response = self._request("POST", "/api/files/duplicate", {"paths": ["storage1/._note.txt"]}) self.assertEqual(response.status_code, 202) detail = self._wait_task(response.json()["task_id"]) self.assertEqual(detail["status"], "completed") self.assertEqual(detail["done_items"], 0) self.assertEqual(detail["total_items"], 0) self.assertFalse((self.root / "._note copy.txt").exists()) def test_duplicate_skips_macos_sidecars_inside_directory_tree(self) -> None: (self.root / "docs" / "nested").mkdir(parents=True) (self.root / "docs" / "keep.txt").write_text("keep", encoding="utf-8") (self.root / "docs" / "._skip.txt").write_text("skip", encoding="utf-8") (self.root / "docs" / "nested" / "._nested.txt").write_text("skip", encoding="utf-8") (self.root / "docs" / "nested" / "real.txt").write_text("real", encoding="utf-8") response = self._request("POST", "/api/files/duplicate", {"paths": ["storage1/docs"]}) self.assertEqual(response.status_code, 202) self._wait_task(response.json()["task_id"]) self.assertTrue((self.root / "docs copy" / "keep.txt").exists()) self.assertTrue((self.root / "docs copy" / "nested" / "real.txt").exists()) self.assertFalse((self.root / "docs copy" / "._skip.txt").exists()) self.assertFalse((self.root / "docs copy" / "nested" / "._nested.txt").exists()) def test_duplicate_failure_removes_partial_directory_artifact(self) -> None: (self.root / "docs").mkdir() (self.root / "docs" / "a.txt").write_text("A", encoding="utf-8") (self.root / "docs" / "b.txt").write_text("B", encoding="utf-8") path_guard = PathGuard({"storage1": str(self.root), "storage2": str(self.root)}) self._set_services(path_guard=path_guard, filesystem=FailOnSecondCopyFilesystemAdapter()) response = self._request("POST", "/api/files/duplicate", {"paths": ["storage1/docs"]}) self.assertEqual(response.status_code, 202) detail = self._wait_task(response.json()["task_id"]) self.assertEqual(detail["status"], "failed") self.assertEqual(detail["error_code"], "io_error") self.assertFalse((self.root / "docs copy").exists()) def test_duplicate_path_safety_stays_intact(self) -> None: response = self._request("POST", "/api/files/duplicate", {"paths": ["storage1/../escape.txt"]}) self.assertEqual(response.status_code, 403) self.assertEqual(response.json()["error"]["code"], "path_traversal_detected") if __name__ == "__main__": unittest.main()