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_copy_task_service, get_task_service from backend.app.db.task_repository import TaskRepository from backend.app.main import app from backend.app.security.path_guard import PathGuard from backend.app.services.copy_task_service import CopyTaskService from backend.app.services.task_service import TaskService from backend.app.tasks_runner import TaskRunner from backend.app.fs.filesystem_adapter import FilesystemAdapter class FailingFilesystemAdapter(FilesystemAdapter): def copy_file(self, source: str, destination: str, on_progress: callable | None = None) -> None: raise OSError("forced copy failure") def copy_directory(self, source: str, destination: str) -> None: raise OSError("forced copy failure") class BlockingCopyFilesystemAdapter(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) return super().copy_file(source=source, destination=destination, on_progress=on_progress) class CopyApiGoldenTest(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) copy_service = CopyTaskService(path_guard=path_guard, repository=self.repo, runner=runner) task_service = TaskService(repository=self.repo) async def _override_copy_service() -> CopyTaskService: return copy_service async def _override_task_service() -> TaskService: return task_service app.dependency_overrides[get_copy_task_service] = _override_copy_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_copy_success_create_task_shape(self) -> None: src = self.root / "source.txt" src.write_text("hello", encoding="utf-8") response = self._request( "POST", "/api/files/copy", {"source": "storage1/source.txt", "destination": "storage1/copy.txt"}, ) self.assertEqual(response.status_code, 202) body = response.json() self.assertIn("task_id", body) self.assertEqual(body["status"], "queued") detail = self._wait_task(body["task_id"]) self.assertEqual(detail["status"], "completed") self.assertEqual(detail["total_bytes"], 5) self.assertEqual(detail["done_bytes"], 5) self.assertTrue((self.root / "copy.txt").exists()) self.assertEqual((self.root / "copy.txt").read_text(encoding="utf-8"), "hello") def test_copy_batch_multi_file_success(self) -> None: (self.root / "a.txt").write_text("A", encoding="utf-8") (self.root / "b.txt").write_text("B", encoding="utf-8") (self.root / "dest").mkdir() response = self._request( "POST", "/api/files/copy", { "sources": ["storage1/a.txt", "storage1/b.txt"], "destination_base": "storage1/dest", }, ) 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((self.root / "dest" / "a.txt").read_text(encoding="utf-8"), "A") self.assertEqual((self.root / "dest" / "b.txt").read_text(encoding="utf-8"), "B") def test_copy_single_directory_success(self) -> None: src = self.root / "photos" (src / "nested").mkdir(parents=True) (src / "cover.jpg").write_text("img", encoding="utf-8") (src / "nested" / "a.txt").write_text("nested", encoding="utf-8") response = self._request( "POST", "/api/files/copy", {"source": "storage1/photos", "destination": "storage1/photos-copy"}, ) self.assertEqual(response.status_code, 202) detail = self._wait_task(response.json()["task_id"]) self.assertEqual(detail["status"], "completed") self.assertEqual(detail["done_items"], 1) self.assertEqual(detail["total_items"], 1) self.assertTrue((self.root / "photos-copy").is_dir()) self.assertEqual((self.root / "photos-copy" / "cover.jpg").read_text(encoding="utf-8"), "img") self.assertEqual((self.root / "photos-copy" / "nested" / "a.txt").read_text(encoding="utf-8"), "nested") def test_copy_batch_multi_directory_success(self) -> None: (self.root / "dir1" / "sub").mkdir(parents=True) (self.root / "dir2").mkdir() (self.root / "dir1" / "sub" / "a.txt").write_text("A", encoding="utf-8") (self.root / "dir2" / "b.txt").write_text("B", encoding="utf-8") (self.root / "dest").mkdir() response = self._request( "POST", "/api/files/copy", { "sources": ["storage1/dir1", "storage1/dir2"], "destination_base": "storage1/dest", }, ) 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((self.root / "dest" / "dir1" / "sub" / "a.txt").read_text(encoding="utf-8"), "A") self.assertEqual((self.root / "dest" / "dir2" / "b.txt").read_text(encoding="utf-8"), "B") def test_copy_batch_mixed_file_and_directory_success(self) -> None: (self.root / "file.txt").write_text("F", encoding="utf-8") (self.root / "docs" / "nested").mkdir(parents=True) (self.root / "docs" / "nested" / "note.txt").write_text("N", encoding="utf-8") (self.root / "dest").mkdir() response = self._request( "POST", "/api/files/copy", { "sources": ["storage1/file.txt", "storage1/docs"], "destination_base": "storage1/dest", }, ) 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((self.root / "dest" / "file.txt").read_text(encoding="utf-8"), "F") self.assertEqual((self.root / "dest" / "docs" / "nested" / "note.txt").read_text(encoding="utf-8"), "N") def test_copy_batch_cancelled_after_current_file_finishes(self) -> None: blocking_fs = BlockingCopyFilesystemAdapter() 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") (self.root / "dest").mkdir() response = self._request( "POST", "/api/files/copy", { "sources": ["storage1/a.txt", "storage1/b.txt"], "destination_base": "storage1/dest", }, ) 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"], str(self.root / "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 / "dest" / "a.txt").exists()) self.assertFalse((self.root / "dest" / "b.txt").exists()) def test_copy_source_not_found(self) -> None: response = self._request( "POST", "/api/files/copy", {"source": "storage1/missing.txt", "destination": "storage1/out.txt"}, ) self.assertEqual(response.status_code, 404) self.assertEqual( response.json(), { "error": { "code": "path_not_found", "message": "Requested path was not found", "details": {"path": "storage1/missing.txt"}, } }, ) def test_copy_destination_exists_already_exists(self) -> None: (self.root / "source.txt").write_text("x", encoding="utf-8") (self.root / "exists.txt").write_text("y", encoding="utf-8") response = self._request( "POST", "/api/files/copy", {"source": "storage1/source.txt", "destination": "storage1/exists.txt"}, ) self.assertEqual(response.status_code, 409) self.assertEqual( response.json(), { "error": { "code": "already_exists", "message": "Target path already exists", "details": {"path": "storage1/exists.txt"}, } }, ) def test_copy_directory_destination_exists_already_exists(self) -> None: (self.root / "src").mkdir() (self.root / "src" / "a.txt").write_text("x", encoding="utf-8") (self.root / "exists").mkdir() response = self._request( "POST", "/api/files/copy", {"source": "storage1/src", "destination": "storage1/exists"}, ) self.assertEqual(response.status_code, 409) self.assertEqual(response.json()["error"]["code"], "already_exists") def test_copy_batch_destination_exists_already_exists(self) -> None: (self.root / "a.txt").write_text("A", encoding="utf-8") (self.root / "dest").mkdir() (self.root / "dest" / "a.txt").write_text("exists", encoding="utf-8") (self.root / "b.txt").write_text("B", encoding="utf-8") response = self._request( "POST", "/api/files/copy", { "sources": ["storage1/a.txt", "storage1/b.txt"], "destination_base": "storage1/dest", }, ) self.assertEqual(response.status_code, 409) self.assertEqual(response.json()["error"]["code"], "already_exists") def test_copy_traversal_source(self) -> None: response = self._request( "POST", "/api/files/copy", {"source": "storage1/../etc/passwd", "destination": "storage1/out.txt"}, ) self.assertEqual(response.status_code, 403) self.assertEqual(response.json()["error"]["code"], "path_traversal_detected") def test_copy_traversal_destination(self) -> None: (self.root / "source.txt").write_text("x", encoding="utf-8") response = self._request( "POST", "/api/files/copy", {"source": "storage1/source.txt", "destination": "storage1/../etc/out.txt"}, ) self.assertEqual(response.status_code, 403) self.assertEqual(response.json()["error"]["code"], "path_traversal_detected") def test_copy_invalid_root_alias(self) -> None: (self.root / "source.txt").write_text("x", encoding="utf-8") response = self._request( "POST", "/api/files/copy", {"source": "storage1/source.txt", "destination": "unknown/out.txt"}, ) self.assertEqual(response.status_code, 403) self.assertEqual(response.json()["error"]["code"], "invalid_root_alias") def test_copy_destination_inside_directory_source_blocked(self) -> None: (self.root / "src").mkdir() (self.root / "src" / "a.txt").write_text("x", encoding="utf-8") response = self._request( "POST", "/api/files/copy", {"source": "storage1/src", "destination": "storage1/src/child"}, ) self.assertEqual(response.status_code, 400) self.assertEqual(response.json()["error"]["code"], "invalid_request") def test_copy_source_symlink_rejected(self) -> None: target = self.root / "real.txt" target.write_text("x", encoding="utf-8") link = self.root / "link.txt" link.symlink_to(target) response = self._request( "POST", "/api/files/copy", {"source": "storage1/link.txt", "destination": "storage1/out.txt"}, ) self.assertEqual(response.status_code, 409) self.assertEqual(response.json()["error"]["code"], "type_conflict") def test_copy_runtime_io_error_failed_task_shape(self) -> None: src = self.root / "source.txt" src.write_text("hello", encoding="utf-8") path_guard = PathGuard({"storage1": str(self.root), "storage2": str(self.root)}) self._set_services(path_guard=path_guard, filesystem=FailingFilesystemAdapter()) response = self._request( "POST", "/api/files/copy", {"source": "storage1/source.txt", "destination": "storage1/copy.txt"}, ) self.assertEqual(response.status_code, 202) task_id = response.json()["task_id"] detail = self._wait_task(task_id) self.assertEqual(detail["status"], "failed") self.assertEqual(detail["error_code"], "io_error") self.assertEqual(detail["failed_item"], str(src)) if __name__ == "__main__": unittest.main()