langgoap.ExecutionRecord#
- class ExecutionRecord(goal_hash, goal_conditions, plan_actions, expected_cost, actual_cost, outcome, replan_count, timestamp, metadata=<factory>)[source]#
One planner → executor → observer loop outcome.
- Parameters:
goal_hash (str)
goal_conditions (Mapping[str, Any])
plan_actions (tuple[str, ...])
expected_cost (float)
actual_cost (float)
outcome (str)
replan_count (int)
timestamp (datetime)
metadata (Mapping[str, Any])
- goal_hash#
Stable identifier for the goal (user-chosen or derived from
GoalSpec).- Type:
str
- goal_conditions#
The goal’s target condition dict, stored for later diagnostics.
- Type:
Mapping[str, Any]
- plan_actions#
Tuple of action names in execution order.
- Type:
tuple[str, …]
- expected_cost#
The plan’s
total_costat planning time.- Type:
float
- actual_cost#
Realised cost after execution. Equal to
expected_costfor purely deterministic runs; may differ when dynamic costs change mid-execution.- Type:
float
- outcome#
One of
"success","failed","partial". Open string so new tracer implementations can introduce extra states without churning the schema.- Type:
str
- replan_count#
Number of replans that happened during execution.
- Type:
int
- timestamp#
UTC timestamp at record creation.
- Type:
datetime.datetime
- metadata#
Optional free-form dict for tracer-specific extras (token counts, provider, model name, etc.). Never required by the indexing code.
- Type:
Mapping[str, Any]
- __init__(goal_hash, goal_conditions, plan_actions, expected_cost, actual_cost, outcome, replan_count, timestamp, metadata=<factory>)#
- Parameters:
goal_hash (str)
goal_conditions (Mapping[str, Any])
plan_actions (tuple[str, ...])
expected_cost (float)
actual_cost (float)
outcome (str)
replan_count (int)
timestamp (datetime)
metadata (Mapping[str, Any])
- Return type:
None
Methods
__init__(goal_hash, goal_conditions, ...[, ...])from_dict(data)Deserialise from the format produced by
to_dict().to_dict()Serialise to a plain
dictsuitable forBaseStore.put.Attributes