把 21 个内置 agent 的 system_prompt 从 specs.py 的 Python 常量外置为 prompts/<spec.name>.md,import 期由 load_prompt 确定性加载;建立 SPECS 名册 + SCHEMA_CATALOG(Pydantic 类型留 Python)+ 统一只读解析入口 SpecResolver。 纯重构、零功能/schema 变更,缓存断点前块字节级不变(不变量 #9)。 @llm packages/agents(步骤1-3) - spec_model.py:抽出 AgentSpec(frozen,字段不变) - prompt_loader.py:load_prompt = utf-8-sig 去BOM → LF 归一 → NFC → rstrip尾LF, 内存缓存 + fail-fast(PromptNotFoundError),import 期确定性 - schema_catalog.py:SCHEMA_CATALOG[name]→output type 唯一真相源(refiner=None) - prompts/*.md ×21:取常量「运行时值」程序化外迁(反斜杠折行已塌缩, 物理换行≡运行时换行);文件名按 spec.name 连字符(style.md/character-gen.md 等) - specs.py:删 21 常量 + AgentSpec 类;system_prompt=load_prompt(name)、 output_schema=SCHEMA_CATALOG[name];建 SPECS + REVIEW_RESERVED_NAMES; *_spec 兼容期保留且 SPECS[name] is *_spec(同一实例)。804→337 行 - __init__.py:显式 __all__ 重导出(避 F401) @backend packages/skills(步骤4-5) - SpecResolver:内置 SPECS(纯内存、零 DB)+ 用户 SkillRegistry 统一 get; 内置 name 永不触发 DB;output_schema_for 精确匹配 - skill_registry:保留命名空间守卫前移至入库校验,拒同名内置 → VALIDATION - toolbox_registry:GeneratorTool.spec 改走 SPECS[...],删 12 个 *_spec 直接 import @devops repo-root - .gitattributes:prompts/*.md text eol=lf(修正:须用完整嵌套路径才匹配) - packages/agents/pyproject:hatchling artifacts 纳入 prompts/*.md 随 wheel/sdist 分发 - ci.yml:新增 build wheel → 裸装 → import ww_agents.SPECS 冒烟 TDD 全程 mock 网关;门禁绿:ruff/format clean · mypy 209 files · pytest 744 passed (含金标准 sha256 回归 / md↔spec↔catalog 一一对应 / fail-fast / BOM+NFC / 内置守卫 / 同一实例 / 编排器无回归 / apps/api import-smoke / 打包冒烟)
168 lines
4.4 KiB
Python
168 lines
4.4 KiB
Python
"""T3.3 三审契约测试:foreshadow / pace spec + 输出 schema(C6 扩)。
|
||
|
||
契约测试——构造符合 schema 的 mock 响应,校验字段、tier、四审只读(writes==[])。
|
||
不联网、无 DB。
|
||
"""
|
||
|
||
from __future__ import annotations
|
||
|
||
import pytest
|
||
from pydantic import ValidationError
|
||
from ww_agents import (
|
||
AgentSpec,
|
||
ForeshadowReview,
|
||
ForeshadowSuggestion,
|
||
PaceIssue,
|
||
PaceReview,
|
||
foreshadow_spec,
|
||
pace_spec,
|
||
)
|
||
|
||
# ---- ForeshadowReview schema ----
|
||
|
||
|
||
def test_foreshadow_review_parses_mock_response() -> None:
|
||
# Arrange:模拟网关 instructor 校验后的结构化产出
|
||
mock = {
|
||
"planted": [
|
||
{
|
||
"code": "FS-MARK",
|
||
"title": "胸口的胎记",
|
||
"where": "第 2 段主角沐浴时被瞥见",
|
||
"note": "暗示皇族血统",
|
||
}
|
||
],
|
||
"resolved": [
|
||
{
|
||
"code": "FS-BIRTH",
|
||
"title": "身世之谜",
|
||
"where": "第 9 段老者道破来历",
|
||
"note": "回收第 1 章埋设",
|
||
}
|
||
],
|
||
}
|
||
|
||
# Act
|
||
review = ForeshadowReview.model_validate(mock)
|
||
|
||
# Assert
|
||
assert len(review.planted) == 1
|
||
assert review.planted[0].code == "FS-MARK"
|
||
assert review.resolved[0].code == "FS-BIRTH"
|
||
assert review.resolved[0].where == "第 9 段老者道破来历"
|
||
|
||
|
||
def test_foreshadow_review_defaults_to_empty_lists() -> None:
|
||
review = ForeshadowReview()
|
||
assert review.planted == []
|
||
assert review.resolved == []
|
||
|
||
|
||
def test_foreshadow_suggestion_only_title_required() -> None:
|
||
sug = ForeshadowSuggestion.model_validate({"title": "新埋的悬念"})
|
||
assert sug.title == "新埋的悬念"
|
||
assert sug.code is None
|
||
assert sug.where is None
|
||
assert sug.note is None
|
||
|
||
|
||
def test_foreshadow_suggestion_requires_title() -> None:
|
||
with pytest.raises(ValidationError):
|
||
ForeshadowSuggestion.model_validate({"code": "FS-X"})
|
||
|
||
|
||
# ---- PaceReview schema ----
|
||
|
||
|
||
def test_pace_review_parses_mock_response() -> None:
|
||
mock = {
|
||
"water": [{"where": "第 4–6 段反复描写天气", "reason": "信息密度低、与主线无关"}],
|
||
"hook": True,
|
||
"beat_map": [1, 3, 2, 0, 4, 5],
|
||
}
|
||
|
||
review = PaceReview.model_validate(mock)
|
||
|
||
assert len(review.water) == 1
|
||
assert review.water[0].reason == "信息密度低、与主线无关"
|
||
assert review.hook is True
|
||
assert review.beat_map == [1, 3, 2, 0, 4, 5]
|
||
|
||
|
||
def test_pace_review_defaults() -> None:
|
||
review = PaceReview()
|
||
assert review.water == []
|
||
assert review.hook is False
|
||
assert review.beat_map == []
|
||
|
||
|
||
def test_pace_issue_requires_where_and_reason() -> None:
|
||
with pytest.raises(ValidationError):
|
||
PaceIssue.model_validate({"where": "x"})
|
||
|
||
|
||
# ---- foreshadow_spec 声明 ----
|
||
|
||
|
||
def test_foreshadow_spec_is_analyst_tier() -> None:
|
||
assert foreshadow_spec.tier == "analyst"
|
||
assert foreshadow_spec.name == "foreshadow"
|
||
|
||
|
||
def test_foreshadow_spec_is_read_only() -> None:
|
||
# 不变量 #3:四审只读
|
||
assert foreshadow_spec.writes == ()
|
||
|
||
|
||
def test_foreshadow_spec_reads_foreshadow() -> None:
|
||
assert foreshadow_spec.reads == ("foreshadow",)
|
||
|
||
|
||
def test_foreshadow_spec_output_schema() -> None:
|
||
assert foreshadow_spec.output_schema is ForeshadowReview
|
||
|
||
|
||
def test_foreshadow_spec_has_nonempty_system_prompt() -> None:
|
||
assert foreshadow_spec.system_prompt.strip()
|
||
|
||
|
||
# ---- pace_spec 声明 ----
|
||
|
||
|
||
def test_pace_spec_is_light_tier() -> None:
|
||
# 不变量 #2:节奏审用轻量档,只声明 tier
|
||
assert pace_spec.tier == "light"
|
||
assert pace_spec.name == "pace"
|
||
|
||
|
||
def test_pace_spec_is_read_only() -> None:
|
||
assert pace_spec.writes == ()
|
||
|
||
|
||
def test_pace_spec_reads_rules() -> None:
|
||
# genre 模板 DSL 经 rules(genre 级)注入
|
||
assert pace_spec.reads == ("rules",)
|
||
|
||
|
||
def test_pace_spec_output_schema() -> None:
|
||
assert pace_spec.output_schema is PaceReview
|
||
|
||
|
||
def test_pace_spec_has_nonempty_system_prompt() -> None:
|
||
assert pace_spec.system_prompt.strip()
|
||
|
||
|
||
# ---- 共性:immutable + 是 AgentSpec ----
|
||
|
||
|
||
def test_review_specs_are_agent_specs() -> None:
|
||
assert isinstance(foreshadow_spec, AgentSpec)
|
||
assert isinstance(pace_spec, AgentSpec)
|
||
|
||
|
||
def test_review_specs_are_immutable() -> None:
|
||
with pytest.raises(ValidationError):
|
||
foreshadow_spec.tier = "writer"
|
||
with pytest.raises(ValidationError):
|
||
pace_spec.tier = "writer"
|