[NOTICKET][AI] Planner: drop parallelizable_with from Task

Remove the unused parallelizable_with field and its validator check (5).
TaskRunner derives parallelism from depends_on alone, so the field was
redundant. Strips it from schema, validator, all few-shots (A-D), and the
planner prompt.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

src/agents/planner/examples.py

@@ -1,7 +1,7 @@
 """Few-shot examples for the planner prompt.
 
 Two illustrative (question -> TaskList) pairs that teach the OUTPUT SHAPE:
-stages, dependency edges, parallelism, ordered tool-call chains, inline QueryIR,
+stages, dependency edges, ordered tool-call chains, inline QueryIR,
 "${t<id>}" placeholders, and the assumed data-flow convention — `query_structured`
 pulls rows, then a composite `analyze_*` tool consumes them via a `data` placeholder
 referencing the upstream result's column aliases (Pattern A; the tool team may
@@ -40,7 +40,6 @@ _EXAMPLE_A = TaskList(
             expected_output="source_shape",
             success_criteria="Produced the orders table schema; the 3 needed columns are present.",
             depends_on=[],
-            parallelizable_with=[],
             estimated_cost="low",
         ),
         Task(
@@ -74,7 +73,6 @@ _EXAMPLE_A = TaskList(
             expected_output="quarter_rows",
             success_criteria="Produced last quarter's order rows with category and revenue.",
             depends_on=["t1"],
-            parallelizable_with=[],
             estimated_cost="medium",
         ),
         Task(
@@ -95,7 +93,6 @@ _EXAMPLE_A = TaskList(
             expected_output="category_contribution",
             success_criteria="Produced each category's revenue share, ranked high to low.",
             depends_on=["t2"],
-            parallelizable_with=[],
             estimated_cost="low",
         ),
     ],
@@ -120,7 +117,6 @@ _EXAMPLE_B = TaskList(
             expected_output="source_shape",
             success_criteria="Produced the orders table schema; the needed columns are present.",
             depends_on=[],
-            parallelizable_with=[],
             estimated_cost="low",
         ),
         Task(
@@ -159,7 +155,6 @@ _EXAMPLE_B = TaskList(
             expected_output="ytd_rows",
             success_criteria="Produced this year's order-level rows with date, revenue, region.",
             depends_on=["t1"],
-            parallelizable_with=[],
             estimated_cost="medium",
         ),
         Task(
@@ -184,7 +179,6 @@ _EXAMPLE_B = TaskList(
                 "flag months above/below the typical level."
             ),
             depends_on=["t2"],
-            parallelizable_with=[],
             estimated_cost="low",
         ),
     ],
@@ -215,7 +209,6 @@ _EXAMPLE_C = TaskList(
             expected_output="source_shape",
             success_criteria="Produced the orders table schema; date and revenue columns present.",
             depends_on=[],
-            parallelizable_with=["t4"],
             estimated_cost="low",
         ),
         Task(
@@ -253,7 +246,6 @@ _EXAMPLE_C = TaskList(
             expected_output="q1_rows",
             success_criteria="Produced Q1 order rows with date and revenue.",
             depends_on=["t1"],
-            parallelizable_with=[],
             estimated_cost="medium",
         ),
         Task(
@@ -275,7 +267,6 @@ _EXAMPLE_C = TaskList(
             expected_output="q1_trend",
             success_criteria="Produced a per-month revenue series showing where revenue fell.",
             depends_on=["t2"],
-            parallelizable_with=[],
             estimated_cost="low",
         ),
         Task(
@@ -294,7 +285,6 @@ _EXAMPLE_C = TaskList(
             expected_output="q1_context_chunks",
             success_criteria="Produced relevant document chunks about Q1 operations.",
             depends_on=[],
-            parallelizable_with=["t1"],
             estimated_cost="low",
         ),
     ],
@@ -324,7 +314,6 @@ _EXAMPLE_D = TaskList(
             expected_output="source_shape",
             success_criteria="Produced the orders table schema; region and revenue present.",
             depends_on=[],
-            parallelizable_with=[],
             estimated_cost="low",
         ),
         Task(
@@ -350,7 +339,6 @@ _EXAMPLE_D = TaskList(
             expected_output="region_rows",
             success_criteria="Produced order rows with region and revenue.",
             depends_on=["t1"],
-            parallelizable_with=[],
             estimated_cost="medium",
         ),
         Task(
@@ -370,7 +358,6 @@ _EXAMPLE_D = TaskList(
             expected_output="region_aggregates",
             success_criteria="Produced one row per region with mean and total revenue.",
             depends_on=["t2"],
-            parallelizable_with=[],
             estimated_cost="low",
         ),
     ],

src/agents/planner/schemas.py

@@ -42,7 +42,6 @@ class Task(BaseModel):
     expected_output: str  # named result this task produces
     success_criteria: str  # REPORTING signal, not a control trigger
     depends_on: list[str] = Field(default_factory=list)  # task ids
-    parallelizable_with: list[str] = Field(default_factory=list)
     estimated_cost: Literal["low", "medium", "high"] = "low"
 
 

src/agents/planner/validator.py

@@ -113,9 +113,6 @@ class PlannerValidator:
         # Check 4 — DAG: edges resolve, placeholders resolve, no cycles.
         self._validate_dag(tasks_by_id, id_set)
 
-        # Check 5 — parallelizable_with is consistent with the dependency graph.
-        self._validate_parallelism(tasks_by_id, id_set)
-
     def _validate_inline_ir(self, task_id: str, args: dict, catalog: Catalog) -> None:
         raw_ir = args.get("ir")
         if not isinstance(raw_ir, dict):
@@ -170,26 +167,6 @@ class PlannerValidator:
                         f"not a (transitive) dependency — add it to depends_on"
                     )
 
-    @staticmethod
-    def _validate_parallelism(tasks_by_id: dict, id_set: set[str]) -> None:
-        ancestors = _all_ancestors(tasks_by_id)
-        for task in tasks_by_id.values():
-            for other in task.parallelizable_with:
-                if other not in id_set:
-                    raise PlannerValidationError(
-                        f"task {task.id}: parallelizable_with references unknown task "
-                        f"{other!r}"
-                    )
-                if other == task.id:
-                    raise PlannerValidationError(
-                        f"task {task.id}: parallelizable_with includes itself"
-                    )
-                if other in ancestors[task.id] or task.id in ancestors[other]:
-                    raise PlannerValidationError(
-                        f"task {task.id}: parallelizable_with {other!r} conflicts with a "
-                        f"(transitive) depends_on relationship between them"
-                    )
-
 
 def _is_placeholder(value: str) -> bool:
     return bool(_PLACEHOLDER_RE.fullmatch(value.strip()))

src/config/prompts/planner.md

@@ -41,8 +41,6 @@ only a `TaskList` object that conforms to the provided schema.
   (referencing the upstream result's column aliases).
 - **Mixing structured + unstructured.** If qualitative context helps, add a
   `retrieve_documents` task against an unstructured source listed in the catalog.
-- **Parallelism.** List sibling tasks that have no data dependency on each other
-  in `parallelizable_with` (must be mutually consistent with `depends_on`).
 - **CRISP-DM stages.** Tag each task with the stage it serves:
   `data_understanding`, `data_preparation`, or `evaluation`. (Never `modeling`.)
 - **success_criteria is a reporting signal**, not a control trigger. State, in