sync identified tasks of existing workplans on projection reset

Hub tasks now carry record_id, so reset_repository_projection can create
missing derived tasks, update title/status, and cancel open tasks the
forge no longer derives. Rows with no record_id are left alone.

Assistant: grok
Assistant-Session: 01a04996-76e8-7f53-b971-1885cfbed436

api/services/forge_projection.py

@@ -577,6 +577,9 @@ class ResetOutcome:
     refused: list[dict[str, Any]] = field(default_factory=list)
     # Identifiers freed from rows retired before retirement released them.
     released: list[str] = field(default_factory=list)
+    created_tasks: list[str] = field(default_factory=list)
+    updated_tasks: list[str] = field(default_factory=list)
+    cancelled_tasks: list[str] = field(default_factory=list)
     notes: list[str] = field(default_factory=list)
 
     def to_dict(self) -> dict[str, Any]:
@@ -591,11 +594,17 @@ class ResetOutcome:
                 "retired": len(self.retired),
                 "refused": len(self.refused),
                 "released": len(self.released),
+                "created_tasks": len(self.created_tasks),
+                "updated_tasks": len(self.updated_tasks),
+                "cancelled_tasks": len(self.cancelled_tasks),
             },
             "created": self.created,
             "updated": self.updated,
             "retired": self.retired,
             "released": self.released,
+            "created_tasks": self.created_tasks,
+            "updated_tasks": self.updated_tasks,
+            "cancelled_tasks": self.cancelled_tasks,
             "refused": self.refused,
             "notes": self.notes,
         }
@@ -626,6 +635,89 @@ def _tombstone_slug(slug: str, when: datetime) -> str:
 RETIRE_REASON = "no longer derived from the forge"
 
 
+def _coerce_task_status(raw: str | None) -> Any:
+    from api.models.task import TaskStatus
+
+    if not raw:
+        return None
+    try:
+        return TaskStatus(str(raw).strip().lower())
+    except ValueError:
+        return None
+
+
+def _task_status_value(status: Any) -> str:
+    return str(getattr(status, "value", status) or "")
+
+
+def _sync_existing_workplan_tasks(
+    session: Any,
+    row: Any,
+    derived_wp: DerivedWorkplan,
+    hub_tasks: list[Any],
+    outcome: ResetOutcome,
+) -> None:
+    """Match identified tasks of an existing workplan by record_id.
+
+    Rows with no ``record_id`` predate the backfill and cannot be qualified
+    from the file alone (CUST-WP-0068-T09). They are not cancelled or
+    overwritten.
+    """
+    from api.models.task import Task, TaskPriority, TaskStatus
+
+    want = {
+        t.record_id.strip().lower(): t
+        for t in derived_wp.tasks
+        if t.record_id and t.record_id.strip()
+    }
+    matched: dict[str, Any] = {}
+    stale: list[Any] = []
+    for ht in hub_tasks:
+        rid = (ht.record_id or "").strip().lower()
+        if not rid:
+            continue
+        if rid in want:
+            matched[rid] = ht
+        else:
+            stale.append(ht)
+
+    for key, dt in want.items():
+        ht = matched.get(key)
+        if ht is None:
+            kwargs: dict[str, Any] = {
+                "id": uuid.UUID(dt.uuid),
+                "workplan_id": row.id,
+                "record_id": dt.record_id,
+                "title": (dt.title or dt.record_id),
+            }
+            st = _coerce_task_status(dt.status)
+            if st is not None:
+                kwargs["status"] = st
+            if dt.priority:
+                try:
+                    kwargs["priority"] = TaskPriority(dt.priority.strip().lower())
+                except ValueError:
+                    pass
+            session.add(Task(**kwargs))
+            outcome.created_tasks.append(dt.record_id)
+            continue
+        changed = False
+        if dt.title and dt.title.strip() and ht.title != dt.title.strip():
+            ht.title = dt.title.strip()
+            changed = True
+        st = _coerce_task_status(dt.status)
+        if st is not None and ht.status != st:
+            ht.status = st
+            changed = True
+        if changed:
+            outcome.updated_tasks.append(dt.record_id)
+
+    for ht in stale:
+        if _task_status_value(ht.status) in {"wait", "todo", "progress"}:
+            ht.status = TaskStatus.cancel
+            outcome.cancelled_tasks.append(ht.record_id or str(ht.id))
+
+
 async def reset_repository_projection(
     session: Any,
     repo_slug: str,
@@ -646,10 +738,10 @@ async def reset_repository_projection(
     file was removed deliberately — or that someone pointed this at the wrong
     branch. The caller must say which.
 
-    Scope: workplans, and the tasks of workplans being created. Tasks of
+    Scope: workplans, and their tasks. Tasks of existing workplans are matched
-    *existing* workplans are left alone, because hub task rows carry no
+    by ``record_id`` (STATE-WP-0083-T06); rows with no ``record_id`` are left
-    canonical identifier and can only be matched by title — renaming a heading
+    alone. A derived task the hub lacks is created; an identified hub task the
-    would otherwise destroy and recreate its record (`T06`).
+    forge no longer derives is cancelled if it is still open.
     """
     from datetime import datetime, timezone
 
@@ -932,6 +1024,17 @@ async def reset_repository_projection(
 
     now = datetime.now(tz=timezone.utc)
 
+    existing_ids = [matched[k].id for k, _w in want.items() if k in matched]
+    tasks_by_wp: dict[Any, list[Any]] = {}
+    if existing_ids:
+        loaded = list(
+            (
+                await session.execute(select(Task).where(Task.workplan_id.in_(existing_ids)))
+            ).scalars()
+        )
+        for task_row in loaded:
+            tasks_by_wp.setdefault(task_row.workplan_id, []).append(task_row)
+
     for key, w in want.items():
         row = matched.get(key)
         if row is None:
@@ -995,6 +1098,9 @@ async def reset_repository_projection(
             changed = True
         if changed:
             outcome.updated.append(w.record_id)
+        _sync_existing_workplan_tasks(
+            session, row, w, tasks_by_wp.get(row.id, []), outcome
+        )
 
     for r in stale:
         outcome.retired.append(r.slug or str(r.id))
@@ -1002,12 +1108,16 @@ async def reset_repository_projection(
         r.projection_retired_reason = RETIRE_REASON
         r.slug = _tombstone_slug(r.slug or str(r.id), now)
 
-    if outcome.created or outcome.updated or outcome.retired or outcome.released:
+    if (
+        outcome.created
+        or outcome.updated
+        or outcome.retired
+        or outcome.released
+        or outcome.created_tasks
+        or outcome.updated_tasks
+        or outcome.cancelled_tasks
+    ):
         outcome.status = "applied"
-    outcome.notes.append(
-        "Tasks of existing workplans were not touched; hub tasks carry no "
-        "canonical identifier (STATE-WP-0083-T06)."
-    )
     return outcome
 
 

tests/test_forge_projection.py

@@ -218,11 +218,12 @@ class _Row:
 class _FakeSession:
     """Stands in for AsyncSession: enough to prove intent without a database."""
 
-    def __init__(self, repo, rows, foreign=None, slug_clash=None):
+    def __init__(self, repo, rows, foreign=None, slug_clash=None, task_rows=None):
         self._repo = repo
         self.rows = list(rows)
         self._foreign = list(foreign or [])
         self._slug_clash = list(slug_clash or [])
+        self._task_rows = task_rows
         self.added = []
         self.deleted = []
         self.committed = False
@@ -231,16 +232,24 @@ class _FakeSession:
     async def execute(self, *_a, **_k):
         self._calls += 1
         repo, rows = self._repo, self.rows
-        # 1st call resolves the repo, 2nd loads its workplans, 3rd is the
-        # foreign-identifier lookup.
         # 1 resolves the repo, 2 loads its workplans, 3 is the identifier
-        # lookup, 4 the slug lookup.
+        # lookup, 4 the slug lookup. On the update-only path (no creates)
+        # the 3rd call is the existing-task load instead.
+        update_only = (
+            self._task_rows is not None
+            and not self._foreign
+            and not self._slug_clash
+        )
         if self._calls == 2:
             payload = rows
+        elif update_only and self._calls == 3:
+            payload = self._task_rows
         elif self._calls == 3:
             payload = self._foreign
         elif self._calls == 4:
             payload = self._slug_clash
+        elif self._calls == 5:
+            payload = self._task_rows or []
         else:
             payload = []
 
@@ -1030,3 +1039,81 @@ class TestTitleSyncBehaviour:
         session = _FakeSession(repo=_Repo(), rows=[row])
         await fp.reset_repository_projection(session, "demo", derived=d)
         assert row.title == "Real Title"
+
+
+class _TaskRow:
+    def __init__(self, record_id, status="todo", title="t", workplan_id=None):
+        import uuid as _u
+        self.id = _u.uuid4()
+        self.record_id = record_id
+        self.status = status
+        self.title = title
+        self.priority = "medium"
+        self.workplan_id = workplan_id
+
+
+class TestExistingWorkplanTasks:
+    """Tasks of an existing workplan are matched by record_id (CUST-WP-0068-T09)."""
+
+    def _derived(self, *tasks):
+        dts = [
+            fp.DerivedTask(
+                record_id=tid,
+                uuid=fp.derived_record_uuid(tid),
+                title=title,
+                status=st,
+                priority="medium",
+            )
+            for tid, title, st in tasks
+        ]
+        return fp.DerivedProjection(
+            repo_slug="demo",
+            commit="c0ffee",
+            workplans=[
+                fp.DerivedWorkplan(
+                    record_id="DEMO-WP-0001",
+                    uuid=fp.derived_record_uuid("DEMO-WP-0001"),
+                    title="DEMO-WP-0001",
+                    status="active",
+                    relative_path="workplans/a.md",
+                    archived=False,
+                    tasks=dts,
+                )
+            ],
+        )
+
+    @pytest.mark.asyncio
+    async def test_creates_a_missing_identified_task(self):
+        row = _Row(slug="demo-wp-0001", status="active", path="workplans/a.md")
+        session = _FakeSession(repo=_Repo(), rows=[row], task_rows=[])
+        out = await fp.reset_repository_projection(
+            session, "demo",
+            derived=self._derived(("DEMO-WP-0001-T01", "Do it", "todo")),
+        )
+        assert out.created_tasks == ["DEMO-WP-0001-T01"]
+        assert session.added
+        assert session.added[0].record_id == "DEMO-WP-0001-T01"
+
+    @pytest.mark.asyncio
+    async def test_cancels_an_open_task_the_file_no_longer_derives(self):
+        row = _Row(slug="demo-wp-0001", status="active", path="workplans/a.md")
+        stale = _TaskRow("DEMO-WP-0001-T09", status="todo", workplan_id=row.id)
+        session = _FakeSession(repo=_Repo(), rows=[row], task_rows=[stale])
+        out = await fp.reset_repository_projection(
+            session, "demo",
+            derived=self._derived(("DEMO-WP-0001-T01", "Do it", "todo")),
+        )
+        assert stale.status.value == "cancel" or stale.status == "cancel" or str(stale.status).endswith("cancel")
+        assert "DEMO-WP-0001-T09" in out.cancelled_tasks
+        assert session.deleted == []
+
+    @pytest.mark.asyncio
+    async def test_leaves_unidentified_tasks_alone(self):
+        row = _Row(slug="demo-wp-0001", status="active", path="workplans/a.md")
+        orphan = _TaskRow(None, status="todo", title="legacy", workplan_id=row.id)
+        session = _FakeSession(repo=_Repo(), rows=[row], task_rows=[orphan])
+        await fp.reset_repository_projection(
+            session, "demo",
+            derived=self._derived(("DEMO-WP-0001-T01", "Do it", "todo")),
+        )
+        assert orphan.status == "todo"