Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 12 additions & 17 deletions docs/architecture/rfcs/monorepo-distribution-split-v0.md
Original file line number Diff line number Diff line change
Expand Up @@ -671,23 +671,18 @@ diaries in this RFC.

#### 3. Concrete retirement backlog and ratchet rules

1. **Pre-Turn pilot CLI subcommands and top-level modules.** Retire the `15`
superseded pre-Turn `codex-cli-*` pilot/proof subcommands replaced by
`loopx turn` (`turn_driver/codex_cli.py`), `start-goal`, and `codex_app`
heartbeat (`14` with zero non-pilot callers plus `codex-cli-session-probe`
referenced only as a hint in `loopx/project_prompt.py`), along with the `4`
top-level `loopx/codex_cli_*.py` modules (`3,767` LOC) and `6`
`loopx/cli_commands/starter_*.py` wrappers (`859` LOC). Keep live
`codex-cli-bootstrap-message` and `turn_driver/codex_cli.py`.
2. **Remaining top-level `chat_*` and `*_goal_mode` migrations with strict
downward ratchet.** Complete M1/M2 by migrating the remaining `20` top-level
`loopx/chat_*.py` modules (`20,562` LOC) into `loopx/chat/` and the `9`
`loopx/*_goal_mode/` directories (`59` files / `17,655` lines; `33` `.py` /
`7,062` LOC) into `loopx/hosts/`. Every PR deleting or moving top-level
modules must ratchet `max_top_level_modules` in
[`top_level_module_budget.json`](../../../tests/architecture/top_level_module_budget.json)
(currently `127`) strictly downward; keeping permanent top-level shim files
that freeze the count does not clear M1/M2.
1. **Pre-Turn pilot CLI retirement.** [#8884](https://github.com/loopx-project/loopx/pull/8884)
retired 15 superseded pilot commands and four flat `loopx/codex_cli_*.py`
modules, and lowered the top-level budget from 127 to 124. The live
`codex-cli-bootstrap-message` and `turn_driver/codex_cli.py` still have callers.
2. **Remaining top-level `chat_*` and `*_goal_mode` migrations.** M1 places
the internal action normalization and Goal lifecycle action modules under
`loopx/chat/`. This removes two flat names and lowers the budget from 124
to 122. M1 still needs to migrate the remaining Chat modules under I3.
M2 still moves host mode packages under `loopx/hosts/`. Each net removal
lowers [`top_level_module_budget.json`](../../../tests/architecture/top_level_module_budget.json)
by the same count in that PR. The Core wheel already excludes both Chat
paths, so this move reduces the source directory count.
3. **Unreferenced `examples/` scripts and source-text-pinning smokes.** Prune
unreferenced scripts under `examples/` (`200` unreferenced `.py` scripts /
`45,504` LOC; `209` scripts / `47,325` LOC across `.py/.sh/.mjs/.js/.ts` on
Expand Down
19 changes: 8 additions & 11 deletions docs/architecture/rfcs/monorepo-distribution-split-v0.zh-CN.md
Original file line number Diff line number Diff line change
Expand Up @@ -488,17 +488,14 @@ R0 可以先于目录移动交付,但不代表内核合格。R3/R4 可选,

#### 3. 具名退役清单与棘轮规则

1. **退役 Turn 前置 pilot CLI 子命令与顶层模块。** 删除已被 `loopx turn`(`turn_driver/codex_cli.py`)、
`start-goal` 与 `codex_app` heartbeat 替代的 `15` 个 `codex-cli-*` pilot/证明子命令(其中
`14` 个在非 pilot 生产调用方中为零引用,另有 `codex-cli-session-probe` 仅作为提示出现在
`loopx/project_prompt.py`),以及 `4` 个顶层 `loopx/codex_cli_*.py`(`3,767` 行)和 `6` 个
`loopx/cli_commands/starter_*.py` 包装(`859` 行);保留现役 `codex-cli-bootstrap-message` 与
`turn_driver/codex_cli.py`。
2. **迁完剩余顶层 `chat_*` 与 `*_goal_mode` 并严格单向下调棘轮。** 完成 M1/M2:将剩余 `20` 个顶层
`loopx/chat_*.py`(`20,562` 行)迁入 `loopx/chat/`,将 `9` 个 `loopx/*_goal_mode/` 目录
(`59` 文件 / `17,655` 行;其中 `33` 个 `.py` / `7,062` 行)迁入 `loopx/hosts/`。任何删除或迁出
顶层模块的 PR 都必须在同一 PR 内同步调低 [`top_level_module_budget.json`](../../../tests/architecture/top_level_module_budget.json)
中的 `max_top_level_modules`(当前 `127`);靠保留永久顶层 shim 维持计数不变不构成 M1/M2 完成。
1. **退役 Turn 前置 pilot CLI。** [#8884](https://github.com/loopx-project/loopx/pull/8884)
已退役 15 个被替代的试验命令和四个顶层 `loopx/codex_cli_*.py` 模块,顶层预算从 127 降至 124。
`codex-cli-bootstrap-message` 与 `turn_driver/codex_cli.py` 仍有现役调用方。
2. **迁移剩余顶层 `chat_*` 和 `*_goal_mode`。** M1 将动作规范化和 Goal 生命周期动作两个内部模块
放入 `loopx/chat/`,移除两个顶层名字,预算由 124 降至 122。M1 剩余 Chat 模块仍按 I3 迁移,
M2 继续把宿主模式包迁入 `loopx/hosts/`。每次净减少都在同一 PR 中等量调低
[`top_level_module_budget.json`](../../../tests/architecture/top_level_module_budget.json)。
Core wheel 原已排除这两类 Chat 路径,这次减少的是源码顶层模块数量。
3. **清理无引用 `examples/` 与源码文本绑定式 smoke。** 清理 `examples/` 中无引用脚本(`main@f15ebdd16b`
实测 `200` 个无引用 `.py` / `45,504` 行;按 `.py/.sh/.mjs/.js/.ts` 计为 `209` 个 / `47,325` 行),
并依 [good-smokes](../../development/good-smokes.md) 退役 `14` 个绑定内部 Python 符号而非已交付
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -6,11 +6,11 @@
import re
from typing import Any, Mapping

from .agent_registry import registered_agent_ids_for_goal
from .control_plane.content_digest import BARE_SHA256_PATTERN
from .control_plane.runtime.time import now_utc, parse_timestamp, utc_isoformat
from .control_plane.todos.contract import require_supported_todo_resume_when
from .registry import registry_goals
from ..agent_registry import registered_agent_ids_for_goal
from ..control_plane.content_digest import BARE_SHA256_PATTERN
from ..control_plane.runtime.time import now_utc, parse_timestamp, utc_isoformat
from ..control_plane.todos.contract import require_supported_todo_resume_when
from ..registry import registry_goals


_AUTHORITY_PRINCIPAL = re.compile(r"^[a-z][a-z0-9._-]{0,30}:[A-Za-z0-9._:-]{1,200}$")
Expand All @@ -33,7 +33,7 @@ def _normalize(
) -> dict[str, Any]:
# Imported lazily to retain the compatibility helpers in chat_actions
# without introducing a module initialization cycle.
from .chat_actions import (
from ..chat_actions import (
ProtectedActionGate,
_digest,
_normalize_cadence,
Expand Down Expand Up @@ -174,15 +174,15 @@ def _normalize(
"agent_session",
"managed_turn",
}:
from .control_plane.collaboration.goal_instance_scope import (
from ..control_plane.collaboration.goal_instance_scope import (
collaboration_goal_scope,
)
from .control_plane.effect_runtime import (
from ..control_plane.effect_runtime import (
EffectRuntimeRejected,
effect_runtime_result,
)
from .control_plane.goals.activation import goal_is_stopped
from .thread_agent_binding import resolve_registry_thread_agent_binding
from ..control_plane.goals.activation import goal_is_stopped
from ..thread_agent_binding import resolve_registry_thread_agent_binding

try:
executor = dict(
Expand Down Expand Up @@ -227,7 +227,7 @@ def _normalize(
raise ValueError("agent operation Goal is stopped")
origin_goal_ref = scope.current_goal_ref
if executor["kind"] == "managed_turn":
from .control_plane.collaboration.operation_handoff import (
from ..control_plane.collaboration.operation_handoff import (
managed_operation_binding_current,
)

Expand Down Expand Up @@ -260,8 +260,8 @@ def _normalize(
raise ValueError("source route selection requires a managed executor")
source_route = None
if managed_source:
from .control_plane.effect_runtime import effect_runtime_result
from .thread_agent_binding import (
from ..control_plane.effect_runtime import effect_runtime_result
from ..thread_agent_binding import (
collect_accepted_bindings,
resolve_thread_agent_binding,
)
Expand Down Expand Up @@ -583,7 +583,7 @@ def _normalize(
if values.get(field):
result[field] = _text(values[field], field=field, limit=1000)
if "permission" in values or "start_execution" in values:
from .control_plane.effect_runtime import effect_runtime_result
from ..control_plane.effect_runtime import effect_runtime_result
result.update(effect_runtime_result("goal.creation.execution.normalize", values))
if result.get("open_session") is False and result.get("start_execution"):
# start_execution queues the original task as the first Turn of the Session the
Expand Down Expand Up @@ -721,10 +721,10 @@ def _normalize(
raise ValueError("heartbeat edit requires a configuration change")
return result
if action_kind == "team.plan":
from .control_plane.todos.contract import (
from ..control_plane.todos.contract import (
TODO_ACTION_KIND_ADVANCEMENT_VALUES,
)
from .control_plane.work_items.governed_transition_proposal import (
from ..control_plane.work_items.governed_transition_proposal import (
validate_steward_team_plan_preview,
)

Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,9 @@

from typing import Any

from .control_plane.goals.activation import GoalActivationState, goal_activation_state
from .control_plane.goals.activation_service import set_goal_activation_state
from .control_plane.goals.deletion_service import delete_stopped_goal
from ..control_plane.goals.activation import GoalActivationState, goal_activation_state
from ..control_plane.goals.activation_service import set_goal_activation_state
from ..control_plane.goals.deletion_service import delete_stopped_goal


GOAL_LIFECYCLE_SOURCE_BASIS_SCHEMA_VERSION = "loopx_goal_lifecycle_source_basis_v1"
Expand Down Expand Up @@ -75,7 +75,7 @@ def _goal_lifecycle_preview(
def _normalize_goal_lifecycle(
self, parameters: dict[str, Any]
) -> dict[str, Any]:
from .chat_actions import _opaque, _text
from ..chat_actions import _opaque, _text

values = self._allowed_parameters(
parameters,
Expand All @@ -100,7 +100,7 @@ def _apply_goal_delete(
goal_id: str,
current_fingerprint: str,
) -> dict[str, Any]:
from .chat_actions import _digest
from ..chat_actions import _digest

expected_fingerprint = str(proposal.get("expected_state_fingerprint") or "")
if current_fingerprint != expected_fingerprint:
Expand Down Expand Up @@ -175,7 +175,7 @@ def _goal_delete_failed(
def _apply_goal_lifecycle(
self, proposal_id: str, proposal: dict[str, Any], parameters: dict[str, Any]
) -> dict[str, Any]:
from .chat_actions import _digest
from ..chat_actions import _digest

goal_id = str(parameters["goal_id"])
operation = str(parameters["operation"])
Expand Down
4 changes: 2 additions & 2 deletions loopx/chat_actions.py
Original file line number Diff line number Diff line change
Expand Up @@ -12,13 +12,13 @@
from .control_plane.projects.identity import GoalCreationConflictError
from .control_plane.coordination.new_goal_storage_transport import initialize_goal_storage_target
from .chat import apply_todo_review_preview, build_todo_review_preview
from .chat_action_normalization import ChatActionNormalizationMixin
from .chat.action_normalization import ChatActionNormalizationMixin
from .chat_action_store import (
ACTION_KINDS as SUPPORTED_ACTION_KINDS,
ActionConflictError,
ChatActionStore,
)
from .chat_goal_lifecycle_actions import ChatGoalLifecycleActionMixin
from .chat.goal_lifecycle_actions import ChatGoalLifecycleActionMixin
from .chat.monitor_actions import ChatMonitorActionMixin
from .chat_store import ChatSessionStore, TERMINAL_TURN_STATES
from .chat.todo_actions import ChatTodoActionMixin
Expand Down
2 changes: 1 addition & 1 deletion tests/architecture/test_content_digest_single_owner.py
Original file line number Diff line number Diff line change
Expand Up @@ -181,7 +181,7 @@
"loopx.capabilities.periodic_report.review",
"loopx.capabilities.progress_review.receipt",
"loopx.chat.result_files_api",
"loopx.chat_action_normalization",
"loopx.chat.action_normalization",
"loopx.capabilities.manager_context.chat_history",
"loopx.cli_commands.quota_context_reader",
"loopx.cli_commands.update_artifact",
Expand Down
4 changes: 2 additions & 2 deletions tests/architecture/top_level_module_budget.json
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
{
"schema": "loopx_top_level_module_budget_v0",
"baseline_commit": "0514512ee7fbdf5548c88d07763a41dbfc5df931",
"max_top_level_modules": 124,
"baseline_commit": "41ea75e89892ed88d0ae543e97e31aad0a181c47",
"max_top_level_modules": 122,
"allowlist": ["__init__.py", "entrypoint.py", "cli.py"]
}
7 changes: 4 additions & 3 deletions tests/control_plane/test_compact_identifier_shape_owner.py
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,8 @@

import pytest

from loopx import chat_action_normalization, chat_actions, chat_action_store, domain_state
from loopx import chat_actions, chat_action_store, domain_state
from loopx.chat import action_normalization
from loopx.capabilities.benchmark_toolkit import integrity
from loopx.capabilities.context_providers import service_ownership
from loopx.capabilities.decision_context import packets, profile, sources
Expand Down Expand Up @@ -254,10 +255,10 @@ def test_a_principal_reference_is_a_different_question_and_stays_separate() -> N
# ``^[a-z][a-z0-9._-]{0,30}:<opaque>$`` embeds the same body behind a required
# family prefix. Merging them would either widen the opaque id to accept a
# prefix it never allowed or narrow a principal to a bare id.
principal = chat_action_normalization._AUTHORITY_PRINCIPAL
principal = action_normalization._AUTHORITY_PRINCIPAL
assert principal.fullmatch("codex:thread-1")
assert not principal.fullmatch("thread-1")
assert PUBLIC_OPAQUE_ID.fullmatch("thread-1")
assert OPAQUE_SHAPE not in (
REPOSITORY_ROOT / "loopx/chat_action_normalization.py"
REPOSITORY_ROOT / "loopx/chat/action_normalization.py"
).read_text(encoding="utf-8")
18 changes: 9 additions & 9 deletions tests/control_plane/test_goal_activation.py
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@

import pytest

from loopx import chat_goal_lifecycle_actions
from loopx.chat import goal_lifecycle_actions
from loopx.chat_action_store import ChatActionStore
from loopx.chat_actions import ChatActionService
from loopx.cli_commands import registry_admin
Expand Down Expand Up @@ -879,7 +879,7 @@ def test_owner_confirmed_lifecycle_action_rechecks_source_inside_write_lock(
"idempotency_key": "stop-goal-one-before-lock-race",
}
)
original = chat_goal_lifecycle_actions.set_goal_activation_state
original = goal_lifecycle_actions.set_goal_activation_state
source_changed = False

def change_source_before_execute(**kwargs: object) -> dict[str, object]:
Expand All @@ -894,7 +894,7 @@ def change_source_before_execute(**kwargs: object) -> dict[str, object]:
return original(**kwargs)

monkeypatch.setattr(
chat_goal_lifecycle_actions,
goal_lifecycle_actions,
"set_goal_activation_state",
change_source_before_execute,
)
Expand Down Expand Up @@ -939,7 +939,7 @@ def test_owner_confirmed_lifecycle_action_rechecks_route_before_execute(
"idempotency_key": "stop-goal-one-before-execute-route-race",
}
)
original = chat_goal_lifecycle_actions.set_goal_activation_state
original = goal_lifecycle_actions.set_goal_activation_state
switched_global_bytes: bytes | None = None

def change_route_before_execute(**kwargs: object) -> dict[str, object]:
Expand All @@ -952,7 +952,7 @@ def change_route_before_execute(**kwargs: object) -> dict[str, object]:
return original(**kwargs)

monkeypatch.setattr(
chat_goal_lifecycle_actions,
goal_lifecycle_actions,
"set_goal_activation_state",
change_route_before_execute,
)
Expand Down Expand Up @@ -1656,7 +1656,7 @@ def test_owner_confirmed_typed_action_recovers_concurrent_same_operation_delete(
action_store=tmp_path / "actions",
idempotency_key="delete-goal-one-concurrently",
)
original_delete = chat_goal_lifecycle_actions.delete_stopped_goal
original_delete = goal_lifecycle_actions.delete_stopped_goal
delete_calls = 0

def delete_between_preflight_and_execute(**kwargs: object) -> dict[str, object]:
Expand All @@ -1675,7 +1675,7 @@ def delete_between_preflight_and_execute(**kwargs: object) -> dict[str, object]:
return current

monkeypatch.setattr(
chat_goal_lifecycle_actions,
goal_lifecycle_actions,
"delete_stopped_goal",
delete_between_preflight_and_execute,
)
Expand Down Expand Up @@ -1746,7 +1746,7 @@ def test_owner_confirmed_typed_action_records_delete_failures(
action_store=tmp_path / "actions",
idempotency_key=f"delete-goal-one-lock-failure-{failure_call}",
)
original_delete = chat_goal_lifecycle_actions.delete_stopped_goal
original_delete = goal_lifecycle_actions.delete_stopped_goal
delete_calls = 0

def fail_delete(**kwargs: object) -> dict[str, object]:
Expand All @@ -1757,7 +1757,7 @@ def fail_delete(**kwargs: object) -> dict[str, object]:
return original_delete(**kwargs)

monkeypatch.setattr(
chat_goal_lifecycle_actions,
goal_lifecycle_actions,
"delete_stopped_goal",
fail_delete,
)
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

import pytest

import loopx.chat_action_normalization as chat_action_normalization
from loopx.chat import action_normalization
from loopx.chat_action_store import ChatActionStore
from loopx.chat_actions import ChatActionService
from loopx.extensions.lark.goal_channel_operation import (
Expand Down Expand Up @@ -113,7 +113,7 @@ def test_builder_feeds_one_canonical_core_dashboard_and_lark_projection(
) -> None:
module = _module()
now = datetime(2026, 9, 14, 1, 0, tzinfo=UTC)
monkeypatch.setattr(chat_action_normalization, "now_utc", lambda: now)
monkeypatch.setattr(action_normalization, "now_utc", lambda: now)
packet = module.build_finance_transaction_approval_packet(_input(now=now), now=now)

assert packet["boundary"] == {
Expand Down
Loading