mirror of
https://github.com/jxxghp/MoviePilot.git
synced 2026-11-03 05:44:44 +08:00
Merge pull request #7050 from DDSRem-Dev/codex-comment-migration
docs: 统一全项目中文注释和结构化契约说明
This commit is contained in:
DDSRem
@@ -1,5 +1,7 @@
|
||||
#!/usr/bin/env python3
|
||||
"""运行订阅执行治理的确定性规模验收并输出 JSON 证据。"""
|
||||
"""
|
||||
运行订阅执行治理的确定性规模验收并输出 JSON 证据。
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
@@ -28,40 +30,89 @@ PROJECT_ROOT = Path(__file__).resolve().parents[2]
|
||||
if str(PROJECT_ROOT) not in sys.path:
|
||||
sys.path.insert(0, str(PROJECT_ROOT))
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.application.site.observation import report_site_search_outcome # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.application.subscription.candidates import CandidateIndex # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.application.subscription.contract import SubscriptionSnapshot # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.application.subscription.execution import SubscriptionExecutionAdmission # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.application.subscription.facts import FreshFactLease # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.application.subscription.sitebudget import ( # noqa: E402
|
||||
SubscriptionSiteBudget as SubscriptionSiteBudgetController,
|
||||
)
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.chain.download import batch as download_batch # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.chain.download.facade import DownloadChain # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.chain.search.facade import SearchChain # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.chain.subscribe import match as subscribe_match # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.chain.subscribe import policy as subscribe_policy # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.chain.subscribe.facade import SubscribeChain # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.db.adapters.subscriptionsearch import ( # noqa: E402
|
||||
TransactionalSubscriptionSearchRepository,
|
||||
)
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.db.base import Base # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.db.models.subscriptionsearch import ( # noqa: E402
|
||||
SubscriptionSearchTask,
|
||||
)
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.db.models.subscriptionsearch import ( # noqa: E402
|
||||
SubscriptionSiteBudget as SubscriptionSiteBudgetRecord,
|
||||
)
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.domain.context import Context, MediaInfo, TorrentInfo # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.domain.metainfo import MetaInfo # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.runtime.stop import ProcessStopState # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.schemas.mediaserver import NotExistMediaInfo # noqa: E402
|
||||
|
||||
# 直接执行脚本时需先把仓库根目录加入导入路径。
|
||||
from app.schemas.types import MediaSource, MediaType # noqa: E402
|
||||
|
||||
|
||||
@dataclass(frozen=True, slots=True)
|
||||
class ScaleCase:
|
||||
"""声明一档订阅、站点、候选和唯一媒体规模。"""
|
||||
"""
|
||||
声明一档订阅、站点、候选和唯一媒体规模。
|
||||
|
||||
:param name: 验收报告中的规模档位标识
|
||||
:param subscription_count: 固定输入中的订阅总数
|
||||
:param site_count: 固定输入中的站点总数
|
||||
:param candidate_count: 索引对照场景中的候选总数
|
||||
:param media_count: 候选所覆盖的唯一媒体身份数量
|
||||
"""
|
||||
|
||||
name: str
|
||||
subscription_count: int
|
||||
@@ -71,7 +122,9 @@ class ScaleCase:
|
||||
|
||||
|
||||
def _percentile(values: list[float], percentile: int) -> float:
|
||||
"""用 nearest-rank 计算稳定百分位，空集合返回零。"""
|
||||
"""
|
||||
用 最近秩法 计算稳定百分位，空集合返回零。
|
||||
"""
|
||||
if not values:
|
||||
return 0.0
|
||||
ordered = sorted(values)
|
||||
@@ -80,7 +133,9 @@ def _percentile(values: list[float], percentile: int) -> float:
|
||||
|
||||
|
||||
def _git_revision() -> str:
|
||||
"""返回当前工作树 HEAD，脚本失败时保留 unknown。"""
|
||||
"""
|
||||
返回当前工作树 HEAD，脚本失败时保留 unknown。
|
||||
"""
|
||||
try:
|
||||
return subprocess.check_output(
|
||||
["git", "rev-parse", "HEAD"],
|
||||
@@ -92,7 +147,9 @@ def _git_revision() -> str:
|
||||
|
||||
|
||||
def _context(*, candidate_index: int, media_id: str, site_id: int) -> Context:
|
||||
"""构造同时携带 canonical 媒体身份和稳定资源身份的候选。"""
|
||||
"""
|
||||
构造同时携带 规范媒体身份和稳定资源身份的候选。
|
||||
"""
|
||||
title = f"Governance Show {media_id} S01E01"
|
||||
meta = MetaInfo(title)
|
||||
meta.type = MediaType.TV
|
||||
@@ -125,7 +182,9 @@ def _context(*, candidate_index: int, media_id: str, site_id: int) -> Context:
|
||||
|
||||
|
||||
def _subscription(subscription_id: int, media_id: str, site_count: int) -> SubscriptionSnapshot:
|
||||
"""构造覆盖全部受控站点的同季电视剧订阅。"""
|
||||
"""
|
||||
构造覆盖全部受控站点的同季电视剧订阅。
|
||||
"""
|
||||
return SubscriptionSnapshot(
|
||||
id=subscription_id,
|
||||
name=f"Governance Show {media_id}",
|
||||
@@ -144,7 +203,9 @@ def _subscription(subscription_id: int, media_id: str, site_count: int) -> Subsc
|
||||
|
||||
|
||||
def _candidate_key(context: Context) -> str:
|
||||
"""返回候选稳定资源键，用于集合与顺序对照。"""
|
||||
"""
|
||||
返回候选稳定资源键，用于集合与顺序对照。
|
||||
"""
|
||||
return str(context.torrent_info.enclosure if context.torrent_info else "")
|
||||
|
||||
|
||||
@@ -152,7 +213,9 @@ def _baseline_route(
|
||||
candidates: dict[str, list[Context]],
|
||||
subscribe: SubscriptionSnapshot,
|
||||
) -> list[str]:
|
||||
"""用完整扫描实现受控显式身份样本的 canonical 参考集合。"""
|
||||
"""
|
||||
用完整扫描实现受控显式身份样本的 规范参考集合。
|
||||
"""
|
||||
target = CandidateIndex.media_identity(subscribe)
|
||||
routed: list[str] = []
|
||||
for contexts in candidates.values():
|
||||
@@ -170,7 +233,9 @@ def _baseline_route(
|
||||
def _build_match_inputs(
|
||||
case: ScaleCase,
|
||||
) -> tuple[dict[str, list[Context]], list[SubscriptionSnapshot]]:
|
||||
"""为索引基线和实际 Match 构造彼此独立的固定输入快照。"""
|
||||
"""
|
||||
为索引基线和实际 Match 构造彼此独立的固定输入快照。
|
||||
"""
|
||||
candidates = {
|
||||
f"site-{site_id}.example": []
|
||||
for site_id in range(1, case.site_count + 1)
|
||||
@@ -197,13 +262,17 @@ def _build_match_inputs(
|
||||
|
||||
|
||||
def _collection_digest(values: list[str]) -> str:
|
||||
"""返回保留重复次数的稳定摘要，便于审计完整集合而不膨胀输出。"""
|
||||
"""
|
||||
返回保留重复次数的稳定摘要，便于审计完整集合而不膨胀输出。
|
||||
"""
|
||||
payload = "\n".join(sorted(values)).encode("utf-8")
|
||||
return hashlib.sha256(payload).hexdigest()
|
||||
|
||||
|
||||
def _compare_collections(expected: set[str], actual: list[str]) -> dict[str, Any]:
|
||||
"""比较完整多重集合并保留计数、摘要、重复项和有限差异样本。"""
|
||||
"""
|
||||
比较完整多重集合并保留计数、摘要、重复项和有限差异样本。
|
||||
"""
|
||||
expected_values = sorted(expected)
|
||||
actual_values = sorted(actual)
|
||||
expected_counts = Counter(expected_values)
|
||||
@@ -224,46 +293,71 @@ def _compare_collections(expected: set[str], actual: list[str]) -> dict[str, Any
|
||||
|
||||
|
||||
class _ScaleSubscriptionRepository:
|
||||
"""为实际 Match 提供固定订阅列表和准入后的最新快照。"""
|
||||
"""
|
||||
为实际 Match 提供固定订阅列表和准入后的最新快照。
|
||||
"""
|
||||
|
||||
def __init__(self, subscribes: list[SubscriptionSnapshot]) -> None:
|
||||
"""
|
||||
保存订阅快照，并建立按标识复读的索引。
|
||||
"""
|
||||
self._subscribes = list(subscribes)
|
||||
self._by_id = {subscribe.id: subscribe for subscribe in subscribes}
|
||||
|
||||
def list(self, _state: str) -> list[SubscriptionSnapshot]:
|
||||
"""返回固定顺序的完整订阅快照。"""
|
||||
"""
|
||||
返回固定顺序的完整订阅快照。
|
||||
"""
|
||||
return list(self._subscribes)
|
||||
|
||||
def get(self, subscription_id: int) -> Optional[SubscriptionSnapshot]:
|
||||
"""返回准入和提交前复读使用的当前订阅快照。"""
|
||||
"""
|
||||
返回准入和提交前复读使用的当前订阅快照。
|
||||
"""
|
||||
return self._by_id.get(subscription_id)
|
||||
|
||||
|
||||
class _ScaleSiteRepository:
|
||||
"""把受控站点 ID 映射为候选分组域名。"""
|
||||
"""
|
||||
把受控站点 ID 映射为候选分组域名。
|
||||
"""
|
||||
|
||||
@staticmethod
|
||||
def get_domains_by_ids(site_ids: list[int]) -> list[str]:
|
||||
"""返回固定快照内与站点 ID 一一对应的域名。"""
|
||||
"""
|
||||
返回固定快照内与站点 ID 一一对应的域名。
|
||||
"""
|
||||
return [f"site-{site_id}.example" for site_id in site_ids]
|
||||
|
||||
|
||||
class _ScaleTorrentHelper:
|
||||
"""隔离实际 Match 的种子属性过滤外部依赖。"""
|
||||
"""
|
||||
隔离实际 Match 的种子属性过滤外部依赖。
|
||||
"""
|
||||
|
||||
@staticmethod
|
||||
def filter_torrent(**_kwargs: Any) -> bool:
|
||||
"""固定样本中的候选均通过订阅属性过滤。"""
|
||||
"""
|
||||
固定样本中的候选均通过订阅属性过滤。
|
||||
"""
|
||||
return True
|
||||
|
||||
|
||||
# 线程同步最多等待 5 秒，避免验收异常时无限挂起。
|
||||
_SITE_PRESSURE_SYNC_TIMEOUT = 5.0
|
||||
|
||||
|
||||
class _ScaleSiteRequestBoundary:
|
||||
"""记录确定性站点请求替身的在途数量和请求结果。"""
|
||||
"""
|
||||
记录确定性站点请求替身的在途数量和请求结果。
|
||||
"""
|
||||
|
||||
def __init__(self, *, site_id: int) -> None:
|
||||
"""
|
||||
初始化站点请求计数、线程同步信号和默认成功结果。
|
||||
|
||||
:param site_id: 本次压力场景中唯一的受控站点标识
|
||||
"""
|
||||
self.site_id = site_id
|
||||
self._lock = threading.Lock()
|
||||
self._active = 0
|
||||
@@ -279,48 +373,64 @@ class _ScaleSiteRequestBoundary:
|
||||
|
||||
@property
|
||||
def active(self) -> int:
|
||||
"""返回请求边界当前真实在途数量。"""
|
||||
"""
|
||||
返回请求边界当前真实在途数量。
|
||||
"""
|
||||
with self._lock:
|
||||
return self._active
|
||||
|
||||
@property
|
||||
def peak(self) -> int:
|
||||
"""返回请求边界观测到的最大在途数量。"""
|
||||
"""
|
||||
返回请求边界观测到的最大在途数量。
|
||||
"""
|
||||
with self._lock:
|
||||
return self._peak
|
||||
|
||||
@property
|
||||
def calls(self) -> int:
|
||||
"""返回请求边界实际进入次数。"""
|
||||
"""
|
||||
返回请求边界实际进入次数。
|
||||
"""
|
||||
with self._lock:
|
||||
return self._calls
|
||||
|
||||
@property
|
||||
def outcomes(self) -> tuple[str, ...]:
|
||||
"""返回请求边界按进入顺序记录的结果分类。"""
|
||||
"""
|
||||
返回请求边界按进入顺序记录的结果分类。
|
||||
"""
|
||||
with self._lock:
|
||||
return tuple(self._outcomes)
|
||||
|
||||
@property
|
||||
def active_at_rejection(self) -> int:
|
||||
"""返回预算拒绝发生时请求边界的在途数量。"""
|
||||
"""
|
||||
返回预算拒绝发生时请求边界的在途数量。
|
||||
"""
|
||||
with self._lock:
|
||||
return self._active_at_rejection
|
||||
|
||||
def set_outcome(self, outcome: str, error: Optional[str] = None) -> None:
|
||||
"""设置后续确定性请求要发布的站点观察结果。"""
|
||||
"""
|
||||
设置后续确定性请求要发布的站点观察结果。
|
||||
"""
|
||||
with self._lock:
|
||||
self._outcome = outcome
|
||||
self._error = error
|
||||
|
||||
def record_budget_rejection(self) -> None:
|
||||
"""记录 wrapper 拒绝时请求边界仍保持在途的事实。"""
|
||||
"""
|
||||
记录 请求包装层拒绝时请求边界仍保持在途的事实。
|
||||
"""
|
||||
with self._lock:
|
||||
self._active_at_rejection = max(self._active_at_rejection, self._active)
|
||||
self.round_observed.set()
|
||||
|
||||
def request(self, *, owner: str) -> list[str]:
|
||||
"""在真实请求边界更新 active/peak 并返回固定站点结果。"""
|
||||
"""
|
||||
在真实请求边界更新 active/peak 并返回固定站点结果。
|
||||
"""
|
||||
with self._lock:
|
||||
call_index = self._calls
|
||||
self._calls += 1
|
||||
@@ -353,7 +463,9 @@ def _site_budget_record(
|
||||
engine: Any,
|
||||
site_id: int,
|
||||
) -> Optional[SubscriptionSiteBudgetRecord]:
|
||||
"""读取站点预算最终记录，核对 observation 是否实际持久化。"""
|
||||
"""
|
||||
读取站点预算最终记录，核对 观察结果是否实际持久化。
|
||||
"""
|
||||
with Session(engine) as session:
|
||||
return session.execute(
|
||||
select(SubscriptionSiteBudgetRecord).where(
|
||||
@@ -367,7 +479,9 @@ def _run_site_pressure_case(
|
||||
search: TransactionalSubscriptionSearchRepository,
|
||||
engine: Any,
|
||||
) -> dict[str, Any]:
|
||||
"""让两个并发 SearchChain owner 通过 wrapper 验证每个站点的容量和冷却。"""
|
||||
"""
|
||||
让两个并发 SearchChain 执行方通过请求包装层 验证每个站点的容量和冷却。
|
||||
"""
|
||||
site_observations: list[dict[str, Any]] = []
|
||||
duplicate_site_claims_blocked = 0
|
||||
error_cooldown_claims_blocked = 0
|
||||
@@ -399,7 +513,9 @@ def _run_site_pressure_case(
|
||||
original=original_record_deferred,
|
||||
owner_name=owner,
|
||||
) -> None:
|
||||
"""保留 wrapper 延后事实并同步压力轮次观测。"""
|
||||
"""
|
||||
保留 请求包装层的延后事实并同步压力轮次观测。
|
||||
"""
|
||||
original(deferral)
|
||||
owner_deferrals[owner_name].append(deferral)
|
||||
boundary.record_budget_rejection()
|
||||
@@ -408,7 +524,9 @@ def _run_site_pressure_case(
|
||||
chain.record_subscription_site_budget_deferred = record_deferred
|
||||
|
||||
def search_site_torrents(*, _owner=owner, **_kwargs: Any) -> list[str]:
|
||||
"""将 SearchChain 的真实站点请求委托给固定边界。"""
|
||||
"""
|
||||
将 SearchChain 的真实站点请求委托给固定边界。
|
||||
"""
|
||||
return boundary.request(owner=_owner)
|
||||
|
||||
chain.search_site_torrents = search_site_torrents
|
||||
@@ -420,7 +538,9 @@ def _run_site_pressure_case(
|
||||
result_lock = threading.Lock()
|
||||
|
||||
def run_owner(owner: str) -> None:
|
||||
"""同步启动一个 owner，并保存 wrapper 的结果或异常。"""
|
||||
"""
|
||||
同步启动一个执行方，并保存请求包装层 的结果或异常。
|
||||
"""
|
||||
try:
|
||||
start_barrier.wait(timeout=_SITE_PRESSURE_SYNC_TIMEOUT)
|
||||
# 租约取得不代表请求已经开始；竞争方必须在真实请求在途时验证拒绝。
|
||||
@@ -428,6 +548,7 @@ def _run_site_pressure_case(
|
||||
raise TimeoutError("首个站点请求未进入受控请求边界")
|
||||
with result_lock:
|
||||
owner_invocations[owner] += 1
|
||||
# 验收需经过真实站点预算包装层，直接调用内部入口以隔离外部请求。
|
||||
result = chains[owner]._search_site_torrents_with_budget( # pylint: disable=protected-access
|
||||
site={"id": site_id, "name": f"Site {site_id}"},
|
||||
keyword="governance-pressure",
|
||||
@@ -436,6 +557,7 @@ def _run_site_pressure_case(
|
||||
)
|
||||
with result_lock:
|
||||
owner_results[owner] = result
|
||||
# 汇总各线程的全部失败，让验收报告统一判定结果。
|
||||
except Exception as error: # noqa: BLE001
|
||||
with result_lock:
|
||||
owner_errors[owner] = str(error)
|
||||
@@ -505,6 +627,7 @@ def _run_site_pressure_case(
|
||||
if site_id != case.site_count:
|
||||
boundary.set_outcome("success")
|
||||
calls_before_reuse = boundary.calls
|
||||
# 验收需经过真实站点预算包装层，直接调用内部入口以隔离外部请求。
|
||||
reused = winner._search_site_torrents_with_budget( # pylint: disable=protected-access
|
||||
site={"id": site_id, "name": f"Site {site_id}"},
|
||||
keyword="governance-pressure-reuse",
|
||||
@@ -526,6 +649,7 @@ def _run_site_pressure_case(
|
||||
if site_id == case.site_count:
|
||||
boundary.set_outcome("rate_limited", error="HTTP 429")
|
||||
calls_before_error = boundary.calls
|
||||
# 验收需经过真实站点预算包装层，直接调用内部入口以隔离外部请求。
|
||||
error_result = winner._search_site_torrents_with_budget( # pylint: disable=protected-access
|
||||
site={"id": site_id, "name": f"Site {site_id}"},
|
||||
keyword="governance-pressure-error",
|
||||
@@ -543,6 +667,7 @@ def _run_site_pressure_case(
|
||||
winner.consume_subscription_site_budget_failures()
|
||||
|
||||
calls_before_cooldown = boundary.calls
|
||||
# 验收需经过真实站点预算包装层，直接调用内部入口以隔离外部请求。
|
||||
cooldown_result = loser._search_site_torrents_with_budget( # pylint: disable=protected-access
|
||||
site={"id": site_id, "name": f"Site {site_id}"},
|
||||
keyword="governance-pressure-cooldown",
|
||||
@@ -637,9 +762,17 @@ def _run_site_pressure_case(
|
||||
|
||||
|
||||
class _ScaleDownloadBoundary(DownloadChain):
|
||||
"""运行 canonical 批量选择，并在实际下载器边界返回确定性成功。"""
|
||||
"""
|
||||
运行 规范批量选择，并在实际下载器边界返回确定性成功。
|
||||
"""
|
||||
|
||||
def __init__(self, matched: list[str], downloaded: list[str]) -> None:
|
||||
"""
|
||||
记录匹配与下载输出容器，供批量选择后核对完整集合。
|
||||
|
||||
:param matched: 接收实际进入批量下载的候选键，调用期间原地追加
|
||||
:param downloaded: 接收选择后模拟成功的候选键，调用期间原地追加
|
||||
"""
|
||||
self._matched = matched
|
||||
self._downloaded = downloaded
|
||||
self._subscription_id = 0
|
||||
@@ -650,7 +783,9 @@ class _ScaleDownloadBoundary(DownloadChain):
|
||||
contexts: list[Context],
|
||||
**_kwargs: Any,
|
||||
) -> tuple[list[Context], dict[str, Any]]:
|
||||
"""固定快照保持候选顺序且不读取历史失败冷却。"""
|
||||
"""
|
||||
固定快照保持候选顺序且不读取历史失败冷却。
|
||||
"""
|
||||
return contexts, {}
|
||||
|
||||
def batch_download(
|
||||
@@ -662,7 +797,9 @@ class _ScaleDownloadBoundary(DownloadChain):
|
||||
governance: Any,
|
||||
**_kwargs: Any,
|
||||
) -> tuple[list[Context], dict[Any, dict[int, NotExistMediaInfo]]]:
|
||||
"""记录 Match 提交集合后运行真实批量候选选择。"""
|
||||
"""
|
||||
记录 Match 提交集合后运行真实批量候选选择。
|
||||
"""
|
||||
source_data = json.loads(source.removeprefix("Subscribe|"))
|
||||
self._subscription_id = int(source_data["id"])
|
||||
for context in contexts:
|
||||
@@ -682,7 +819,9 @@ class _ScaleDownloadBoundary(DownloadChain):
|
||||
governance: Any,
|
||||
**_kwargs: Any,
|
||||
) -> str:
|
||||
"""模拟下载器明确成功，并记录 canonical 选择出的资源。"""
|
||||
"""
|
||||
模拟下载器明确成功，并记录 规范选择出的资源。
|
||||
"""
|
||||
if governance.cancelled and governance.cancelled():
|
||||
raise AssertionError("固定快照执行不应在下载前取消")
|
||||
if governance.mark_started:
|
||||
@@ -692,7 +831,9 @@ class _ScaleDownloadBoundary(DownloadChain):
|
||||
|
||||
|
||||
def _run_match_case(case: ScaleCase) -> dict[str, Any]:
|
||||
"""比较完整扫描与无损索引，并记录单轮事实租约和本地等待分位。"""
|
||||
"""
|
||||
比较完整扫描与无损索引，并记录单轮事实租约和本地等待分位。
|
||||
"""
|
||||
candidates, subscribes = _build_match_inputs(case)
|
||||
|
||||
index = CandidateIndex(candidates)
|
||||
@@ -751,7 +892,9 @@ def _run_match_case(case: ScaleCase) -> dict[str, Any]:
|
||||
|
||||
|
||||
def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
"""运行实际 SubscribeChain Match 并对照四类完整业务集合。"""
|
||||
"""
|
||||
运行实际 SubscribeChain Match 并对照四类完整业务集合。
|
||||
"""
|
||||
candidates, subscribes = _build_match_inputs(case)
|
||||
expected_matched: set[str] = set()
|
||||
expected_downloaded: set[str] = set()
|
||||
@@ -784,36 +927,52 @@ def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
started = time.perf_counter()
|
||||
|
||||
class _ScaleMatchLogger:
|
||||
"""只收集 Match 模块 INFO，用于验证日志量不随候选乘积增长。"""
|
||||
"""
|
||||
只收集 Match 模块 INFO，用于验证日志量不随候选乘积增长。
|
||||
"""
|
||||
|
||||
@staticmethod
|
||||
def debug(*_args: Any, **_kwargs: Any) -> None:
|
||||
"""忽略受控样本的逐项诊断日志。"""
|
||||
"""
|
||||
忽略受控样本的逐项诊断日志。
|
||||
"""
|
||||
|
||||
@staticmethod
|
||||
def scoped_info(*_args: Any, **_kwargs: Any) -> None:
|
||||
"""独立订阅日志不计入主日志的批次 INFO 预算。"""
|
||||
"""
|
||||
独立订阅日志不计入主日志的批次 INFO 预算。
|
||||
"""
|
||||
|
||||
@staticmethod
|
||||
def warning(*_args: Any, **_kwargs: Any) -> None:
|
||||
"""忽略受控样本中预期不会出现的警告。"""
|
||||
"""
|
||||
忽略受控样本中预期不会出现的警告。
|
||||
"""
|
||||
|
||||
warn = warning
|
||||
|
||||
@staticmethod
|
||||
def error(*_args: Any, **_kwargs: Any) -> None:
|
||||
"""错误仍由集合和终态门禁暴露，不污染 INFO 计数。"""
|
||||
"""
|
||||
错误仍由集合和终态门禁暴露，不污染 INFO 计数。
|
||||
"""
|
||||
|
||||
@staticmethod
|
||||
def info(message: Any, *_args: Any, **_kwargs: Any) -> None:
|
||||
"""保存开始和结束摘要。"""
|
||||
"""
|
||||
保存开始和结束摘要。
|
||||
"""
|
||||
info_logs.append(str(message))
|
||||
|
||||
class _ScaleMediaChain:
|
||||
"""用固定新鲜媒体事实替代目标规模下的真实 TMDB 请求。"""
|
||||
"""
|
||||
用固定新鲜媒体事实替代目标规模下的真实 TMDB 请求。
|
||||
"""
|
||||
|
||||
def recognize_media(self, **kwargs: Any) -> MediaInfo:
|
||||
"""按订阅身份返回两集电视剧的新鲜事实。"""
|
||||
"""
|
||||
按订阅身份返回两集电视剧的新鲜事实。
|
||||
"""
|
||||
nonlocal recognition_calls
|
||||
recognition_calls += 1
|
||||
media_id = str(kwargs["media_id"])
|
||||
@@ -829,12 +988,16 @@ def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
|
||||
@staticmethod
|
||||
def recognize_by_meta(*_args: Any, **_kwargs: Any) -> MediaInfo:
|
||||
"""明确身份候选不应进入二次识别。"""
|
||||
"""
|
||||
明确身份候选不应进入二次识别。
|
||||
"""
|
||||
raise AssertionError("明确身份候选不应重新识别")
|
||||
|
||||
@staticmethod
|
||||
def supplement_media_info(mediainfo: MediaInfo) -> MediaInfo:
|
||||
"""固定样本无需访问其他元数据来源。"""
|
||||
"""
|
||||
固定样本无需访问其他元数据来源。
|
||||
"""
|
||||
return mediainfo
|
||||
|
||||
chain = object.__new__(SubscribeChain)
|
||||
@@ -858,7 +1021,9 @@ def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
mediakey: Any,
|
||||
**_kwargs: Any,
|
||||
) -> tuple[bool, dict[Any, dict[int, NotExistMediaInfo]]]:
|
||||
"""返回每条订阅固定的初始缺集快照。"""
|
||||
"""
|
||||
返回每条订阅固定的初始缺集快照。
|
||||
"""
|
||||
return False, {
|
||||
mediakey: {
|
||||
1: NotExistMediaInfo(
|
||||
@@ -874,7 +1039,9 @@ def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
subscribe: SubscriptionSnapshot,
|
||||
**_kwargs: Any,
|
||||
) -> dict[str, Any]:
|
||||
"""让完成判断消费下载边界结果，不引入持久化测试替身。"""
|
||||
"""
|
||||
让完成判断消费下载边界结果，不引入持久化测试替身。
|
||||
"""
|
||||
return {
|
||||
"episodes": [1],
|
||||
"fields": [],
|
||||
@@ -888,7 +1055,9 @@ def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
subscribe: SubscriptionSnapshot,
|
||||
**_kwargs: Any,
|
||||
) -> None:
|
||||
"""记录实际完成判断收到的剩余缺集集合。"""
|
||||
"""
|
||||
记录实际完成判断收到的剩余缺集集合。
|
||||
"""
|
||||
for seasons in no_exists.values():
|
||||
for missing in seasons.values():
|
||||
actual_missing.extend(
|
||||
@@ -905,7 +1074,9 @@ def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
original_finish = chain.finish_subscribe_or_not
|
||||
|
||||
def finish_and_record(**kwargs: Any) -> None:
|
||||
"""运行 canonical 完成判断并记录每条订阅收口时间。"""
|
||||
"""
|
||||
运行 规范完成判断并记录每条订阅收口时间。
|
||||
"""
|
||||
original_finish(**kwargs)
|
||||
settlement_ms.append((time.perf_counter() - started) * 1000)
|
||||
|
||||
@@ -913,7 +1084,9 @@ def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
download_boundary = _ScaleDownloadBoundary(actual_matched, actual_downloaded)
|
||||
|
||||
def record_progress(**kwargs: Any) -> None:
|
||||
"""保存 Match 对外发布的真实批次聚合状态。"""
|
||||
"""
|
||||
保存 Match 对外发布的真实批次聚合状态。
|
||||
"""
|
||||
if kwargs.get("data"):
|
||||
progress_snapshots.append(dict(kwargs["data"]))
|
||||
|
||||
@@ -978,7 +1151,9 @@ def _run_match_execution_case(case: ScaleCase) -> dict[str, Any]:
|
||||
|
||||
|
||||
def _run_durable_governance(case: ScaleCase, workdir: Path) -> dict[str, Any]:
|
||||
"""验证持久队列、站点预算、恢复、取消与订阅级准入。"""
|
||||
"""
|
||||
验证持久队列、站点预算、恢复、取消与订阅级准入。
|
||||
"""
|
||||
engine = create_engine(
|
||||
f"sqlite:///{workdir / (case.name + '.db')}",
|
||||
connect_args={"check_same_thread": False},
|
||||
@@ -1150,7 +1325,14 @@ def _run_durable_governance(case: ScaleCase, workdir: Path) -> dict[str, Any]:
|
||||
|
||||
|
||||
def run_acceptance() -> dict[str, Any]:
|
||||
"""运行两档最终矩阵并返回全部指标与门禁结论。"""
|
||||
"""
|
||||
运行中、大两档固定矩阵，核对匹配集合、并发限制和持久化恢复。
|
||||
|
||||
外部服务使用固定边界，数据库只使用临时目录内的 SQLite；
|
||||
结果衡量本地执行治理，不代表生产网络时延。
|
||||
|
||||
:return: 包含采样环境、各场景指标、逐项门禁和总体通过标记的报告
|
||||
"""
|
||||
cases = (
|
||||
ScaleCase("medium", 100, 10, 1000, 100),
|
||||
ScaleCase("large", 200, 20, 1200, 100),
|
||||
@@ -1292,7 +1474,9 @@ def run_acceptance() -> dict[str, Any]:
|
||||
|
||||
|
||||
def main() -> int:
|
||||
"""解析输出路径，运行门禁并以进程状态反映验收结果。"""
|
||||
"""
|
||||
解析输出路径，运行门禁并以进程状态反映验收结果。
|
||||
"""
|
||||
parser = argparse.ArgumentParser(
|
||||
description=(
|
||||
"运行 100/10/1000 与 200/20/1200 两档离线订阅治理验收；"
|
||||
|
||||
Reference in New Issue
Block a user