feat(R07): task repository, schedule calculator, Qt clock adapter, task/AI-planner services

Team Hoa, EPIC R07 (Scheduling & Workflow Runtime) - Team Hoa scope only
(R07-T01 -> T05; R07-T06 Co4EWorkflowService is Team Nam's).

- R07-T01: infrastructure/persistence/json/task_repository_impl.py wraps
  core/tasks.py's CRUD; core/tasks.py::save_task now writes through
  atomic_write.write_json (same durability fix as R06-T02, save_task was
  still doing a plain write_text).
- R07-T02: domain/tasks/schedule_calculator.py::ScheduleCalculator - the
  cron/interval/daily/weekly/monthly due-time math extracted from
  core/tasks.py, pure Python with is_holiday/make_cron injected so domain/
  never imports core (ADR-001 I2). core/tasks.py keeps its old function
  names as thin wrappers so every existing caller is unchanged. This was
  previously untested; now has its own unit suite.
- R07-T03: infrastructure/qt/qt_scheduler_clock.py::QtSchedulerClock wraps
  the QTimer TaskScheduler used to own directly, injected via a new
  `clock=` constructor param (defaults to a real one). Originally planned
  at platform/qt/... ; moved after confirming that name shadows the
  stdlib platform module (used by core/windows_sandbox_vm.py,
  core/appcontainer_sandbox.py) whenever the repo root is on sys.path.
  tests/fakes/fake_clock.py lets scheduler dispatch be tested tick-by-tick
  with no Qt event loop.
- R07-T04: application/scheduling/task_application_service.py centralizes
  run_now/duplicate/pause/delete/bulk_delete and the Kanban drag-drop
  business rules (move_to_status), currently only reachable by driving
  the real ui/schedule_task_tab.py widget.
- R07-T05: application/scheduling/ai_task_planner_service.py wraps
  core/ai_task_planner.py::plan_tasks and core/task_import.py::import_tasks
  as a seam, plus the attachment-stamping step that used to only exist
  inside the AI-create dialog's worker closure.

pytest: 328 pass (same 4 pre-existing failures as the R05/R06 baseline,
unrelated to this work - see docs/refactor/BaoCao_TeamHoa_R05_R06.md).
scripts/check_imports.py: PASS.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

infrastructure/persistence/json/__init__.py

@@ -1,8 +1,9 @@
 """JSON-file persistence adapters: crash-safe writes and the workspace/
-conversation repositories built on them (EPIC R06)."""
+conversation/task repositories built on them (EPIC R06, R07)."""
 
 from .atomic_write import write_json
 from .conversation_repository_impl import ConversationRepository
+from .task_repository_impl import TaskRepository
 from .workspace_repository_impl import WorkspaceRepository
 
-__all__ = ["write_json", "WorkspaceRepository", "ConversationRepository"]
+__all__ = ["write_json", "WorkspaceRepository", "ConversationRepository", "TaskRepository"]

infrastructure/persistence/json/task_repository_impl.py

@@ -0,0 +1,63 @@
+"""TaskRepository - an object-shaped, atomic-write-backed facade over
+``core/tasks.py`` (R07-T01).
+
+``core/tasks.py``'s module-level functions (``list_tasks``, ``load_task``,
+``save_task``, ``delete_task``, ``new_task``, ``duplicate_task``) are still
+what every existing call site (``core/task_scheduler.py``,
+``core/task_executors.py``, ``ui/schedule_task_tab.py``) uses, and stay that
+way - ``save_task`` now writes through :func:`atomic_write.write_json`
+itself (R07-T01, same class of durability fix already applied to
+``core/projects.py``/``core/history.py`` at R06-T02), so the fix applies
+whether or not a caller ever touches this class.
+
+This repository exists for the application layer
+(``application/scheduling``, R07-T04) to depend on an interface instead of
+reaching into ``core/`` directly. It is a thin pass-through today, not a
+re-implementation: same on-disk format, same directory, same functions
+underneath.
+"""
+from __future__ import annotations
+
+from pathlib import Path
+from typing import Any, Dict, List, Optional
+
+from cowork_local.core.tasks import (
+    TASKS_DIR,
+    delete_task,
+    duplicate_task,
+    list_tasks,
+    load_task,
+    new_task,
+    save_task,
+)
+
+
+class TaskRepository:
+    """CRUD over task dicts (see ``core/tasks.py::DEFAULT_TASK`` for shape),
+    scoped to one ``directory`` (defaults to the app's real ``TASKS_DIR``;
+    tests pass a ``tmp_path`` so nothing touches the user's real config
+    folder)."""
+
+    def __init__(self, directory: Optional[Path] = None) -> None:
+        self._directory = directory or TASKS_DIR
+
+    def list(self) -> List[Dict[str, Any]]:
+        return list_tasks(self._directory)
+
+    def get(self, task_id: str) -> Optional[Dict[str, Any]]:
+        return load_task(task_id, self._directory)
+
+    def save(self, task: Dict[str, Any]) -> Path:
+        return save_task(task, self._directory)
+
+    def create(self, title: str = "", **overrides: Any) -> Dict[str, Any]:
+        return new_task(title, **overrides)
+
+    def duplicate(self, task: Dict[str, Any]) -> Dict[str, Any]:
+        return duplicate_task(task)
+
+    def delete(self, task_id: str) -> None:
+        delete_task(task_id, self._directory)
+
+
+__all__ = ["TaskRepository"]

infrastructure/qt/__init__.py

@@ -0,0 +1,18 @@
+"""Qt-backed adapters for pure interfaces used elsewhere in the app (EPIC R07).
+
+Note: the original plan (``docs/refactor/Feature_Architecture_Proposal.md``)
+placed this adapter at a new top-level ``platform/qt/`` package. That name
+was dropped after it was shown to actually shadow the stdlib ``platform``
+module (used by ``core/windows_sandbox_vm.py``/``core/appcontainer_sandbox.
+py``) whenever the repo root ends up on ``sys.path`` directly - e.g. running
+``python -c "..."`` (or any script) with the repo root as the working
+directory, which resolves a bare ``import platform`` to this package instead
+of the standard library one. ``infrastructure/`` already exists as a layer
+for exactly this kind of toolkit-specific implementation
+(``infrastructure/filesystem/``, ``infrastructure/mcp/``, ...), so the
+adapter lives here instead - same content, safer location.
+"""
+
+from .qt_scheduler_clock import QtSchedulerClock
+
+__all__ = ["QtSchedulerClock"]

infrastructure/qt/qt_scheduler_clock.py

@@ -0,0 +1,70 @@
+"""QtSchedulerClock - the ``QTimer``-backed periodic ticker `TaskScheduler`
+needs, pulled out from ``core/task_scheduler.py`` into its own adapter
+(R07-T03).
+
+``core/task_scheduler.py::TaskScheduler`` is the only file in the scheduling
+stack that imports Qt at all (confirmed by grep — ``core/tasks.py`` and
+``core/task_executors.py`` are Qt-free). Everything it needs Qt FOR is small
+and mechanical: an interval timer that calls back into ``tick()`` every
+``TICK_MS``, plus, during ``stop()``, a way to pump the event loop so a
+worker thread's queued ``finished_ok``/``failed`` signal still gets delivered
+while draining running tasks (see the long comment on ``TaskScheduler.stop()``
+for why that pump matters).
+
+Wrapping exactly that surface — ``start(interval_ms, callback)``, ``stop()``,
+``pump()`` — behind :class:`QtSchedulerClock` lets ``TaskScheduler`` take a
+clock as a constructor parameter instead of constructing a ``QTimer``
+itself. Production wiring is unchanged (``TaskScheduler`` defaults to a real
+``QtSchedulerClock`` when no clock is passed); tests can inject
+``tests/fakes/fake_clock.py::FakeClock`` to control ticks by hand with no Qt
+event loop running at all.
+
+See ``infrastructure/qt/__init__.py`` for why this lives under
+``infrastructure/qt/`` and not the ``platform/qt/`` path the original plan
+named.
+"""
+from __future__ import annotations
+
+from typing import Callable, Optional
+
+from PySide6.QtCore import QCoreApplication, QObject, QTimer
+
+
+class QtSchedulerClock:
+    """Owns one ``QTimer``. Not itself a ``QObject`` subclass — it OWNS a
+    ``QObject``-parented timer instead of inheriting from one, so callers
+    (like ``FakeClock`` in tests) can satisfy the same duck-typed interface
+    without any Qt base class at all."""
+
+    def __init__(self, parent: Optional[QObject] = None) -> None:
+        # Parented so the timer is torn down with its owner instead of
+        # outliving it — the same lifetime QTimer(self) gave it inside
+        # TaskScheduler before this extraction.
+        self._timer = QTimer(parent)
+        self._timer.timeout.connect(self._on_timeout)
+        self._callback: Optional[Callable[[], None]] = None
+
+    def _on_timeout(self) -> None:
+        if self._callback is not None:
+            self._callback()
+
+    def start(self, interval_ms: int, callback: Callable[[], None]) -> None:
+        """Arm and start the timer. Calling this again while already
+        running re-arms it with the new interval/callback (matches
+        ``QTimer.start()``'s own restart-on-repeat-call behaviour)."""
+        self._callback = callback
+        self._timer.setInterval(interval_ms)
+        self._timer.start()
+
+    def stop(self) -> None:
+        self._timer.stop()
+
+    def pump(self) -> None:
+        """Process one batch of pending Qt events — used by
+        ``TaskScheduler.stop()``'s bounded drain loop so a worker thread's
+        queued completion signal can still be delivered while we wait for it
+        to exit."""
+        QCoreApplication.processEvents()
+
+
+__all__ = ["QtSchedulerClock"]