Files
cleveragents-core/benchmarks/phase_reversion_bench.py
freemo 31472b5413
CI / benchmark-publish (pull_request) Has been skipped
CI / build (pull_request) Successful in 21s
CI / helm (pull_request) Successful in 22s
CI / lint (pull_request) Successful in 3m20s
CI / quality (pull_request) Successful in 3m43s
CI / typecheck (pull_request) Successful in 3m58s
CI / security (pull_request) Successful in 4m8s
CI / integration_tests (pull_request) Successful in 9m33s
CI / unit_tests (pull_request) Successful in 10m12s
CI / docker (pull_request) Successful in 1m28s
CI / coverage (pull_request) Successful in 12m18s
CI / e2e_tests (pull_request) Successful in 19m51s
CI / status-check (pull_request) Successful in 1s
CI / build (push) Successful in 15s
CI / helm (push) Successful in 22s
CI / lint (push) Successful in 3m18s
CI / quality (push) Successful in 3m41s
CI / typecheck (push) Successful in 3m57s
CI / benchmark-regression (push) Has been skipped
CI / security (push) Successful in 4m10s
CI / unit_tests (push) Failing after 6m58s
CI / docker (push) Has been skipped
CI / integration_tests (push) Successful in 9m15s
CI / coverage (push) Successful in 12m26s
CI / e2e_tests (push) Successful in 23m11s
CI / status-check (push) Failing after 1s
CI / benchmark-publish (push) Successful in 28m31s
CI / benchmark-regression (pull_request) Successful in 54m53s
test(coverage): add Behave scenarios for 39 under-covered modules
Add Behave feature/step pairs that exercise uncovered branches across handlers, LSP, CLI, and service layers to reach the coverage gate.

ISSUES CLOSED: #1232
2026-03-31 21:47:12 +00:00

148 lines
4.7 KiB
Python

"""ASV benchmarks for phase reversion state machine.
Measures the performance of:
- Plan.can_revert_to() validation
- PlanLifecycleService.revert_plan() overhead
- PlanLifecycleService.try_auto_revert_from_apply() overhead
- Decision recording during reversion
- Reversion loop guard check overhead
"""
from __future__ import annotations
import sys
from pathlib import Path
try:
from cleveragents.application.services.plan_lifecycle_service import (
PlanLifecycleService,
)
from cleveragents.config.settings import Settings
from cleveragents.domain.models.core.plan import (
AutomationProfileProvenance,
AutomationProfileRef,
PlanPhase,
)
except ModuleNotFoundError:
sys.path.insert(0, str(Path(__file__).resolve().parents[1] / "src"))
from cleveragents.application.services.plan_lifecycle_service import (
PlanLifecycleService,
)
from cleveragents.config.settings import Settings
from cleveragents.domain.models.core.plan import (
AutomationProfileProvenance,
AutomationProfileRef,
PlanPhase,
)
def _make_service() -> PlanLifecycleService:
return PlanLifecycleService(settings=Settings())
def _create_plan_in_execute(
service: PlanLifecycleService,
action_name: str = "local/bench-action",
profile: str = "manual",
) -> str:
service.create_action(
name=action_name,
description="Benchmark action",
definition_of_done="Tests pass",
strategy_actor="local/strategist",
execution_actor="local/executor",
)
plan = service.use_action(action_name)
plan.automation_profile = AutomationProfileRef(
profile_name=profile,
provenance=AutomationProfileProvenance.PLAN,
)
plan_id = plan.identity.plan_id
service.start_strategize(plan_id)
service.complete_strategize(plan_id)
service.execute_plan(plan_id)
return plan_id
class CanRevertToSuite:
"""Benchmark Plan.can_revert_to() validation."""
def setup(self) -> None:
service = _make_service()
plan_id = _create_plan_in_execute(service, "local/bench-can-revert")
self.plan = service.get_plan(plan_id)
def time_can_revert_to_valid(self) -> None:
self.plan.can_revert_to(PlanPhase.STRATEGIZE)
def time_can_revert_to_invalid(self) -> None:
self.plan.can_revert_to(PlanPhase.APPLY)
class RevertPlanSuite:
"""Benchmark PlanLifecycleService.revert_plan() overhead."""
def setup(self) -> None:
self.service = _make_service()
self.counter = 0
def time_revert_plan(self) -> None:
self.counter += 1
name = f"local/bench-revert-{self.counter}"
plan_id = _create_plan_in_execute(self.service, name)
self.service.revert_plan(plan_id, PlanPhase.STRATEGIZE, reason="benchmark")
class AutoRevertSuite:
"""Benchmark auto-reversion from constrained apply."""
def setup(self) -> None:
self.service = _make_service()
self.counter = 0
def time_auto_revert_from_apply(self) -> None:
self.counter += 1
name = f"local/bench-auto-{self.counter}"
self.service.create_action(
name=name,
description="Benchmark action",
definition_of_done="Tests pass",
strategy_actor="local/strategist",
execution_actor="local/executor",
)
plan = self.service.use_action(name)
plan.automation_profile = AutomationProfileRef(
profile_name="ci",
provenance=AutomationProfileProvenance.PLAN,
)
plan_id = plan.identity.plan_id
self.service.start_strategize(plan_id)
# complete_strategize auto-progresses to EXECUTE with "ci" profile
self.service.complete_strategize(plan_id)
plan = self.service.get_plan(plan_id)
if plan.phase != PlanPhase.EXECUTE:
self.service.execute_plan(plan_id)
self.service.start_execute(plan_id)
# complete_execute auto-progresses to APPLY with "ci" profile
self.service.complete_execute(plan_id)
self.service.start_apply(plan_id)
self.service.constrain_apply(plan_id, "benchmark constraint")
self.service.try_auto_revert_from_apply(plan_id, "benchmark")
class LoopGuardSuite:
"""Benchmark loop guard check overhead."""
def setup(self) -> None:
self.service = _make_service()
plan_id = _create_plan_in_execute(self.service, "local/bench-guard")
self.plan = self.service.get_plan(plan_id)
def time_loop_guard_under_limit(self) -> None:
self.plan.reversion_count = 2
self.plan.can_revert_to(PlanPhase.STRATEGIZE)
def time_loop_guard_at_limit(self) -> None:
self.plan.reversion_count = 3
self.plan.can_revert_to(PlanPhase.STRATEGIZE)