feat(expenses): backfill historical claims into expense cases

This commit is contained in:
caoxiaozhu
2026-07-14 10:09:09 +08:00
parent 11275e4ba6
commit 5ed34c2b8f
11 changed files with 2053 additions and 3 deletions

View File

@@ -0,0 +1,237 @@
from __future__ import annotations
import re
from dataclasses import dataclass
from typing import Any
from sqlalchemy.engine import URL, make_url
from sqlalchemy.exc import ArgumentError
DEFAULT_POSTGRESQL_PORT = 5432
DISPOSABLE_TARGET_MARKERS = ("migration-probe", "disposable-probe")
SENSITIVE_QUERY_KEY_PARTS = (
"credential",
"passfile",
"password",
"secret",
"token",
)
FORBIDDEN_ROUTING_QUERY_KEYS = frozenset(
{
"database",
"dbname",
"host",
"hostaddr",
"options",
"port",
"service",
"servicefile",
"user",
"username",
}
)
class MaintenanceDatabaseTargetError(ValueError):
"""维护命令数据库目标不满足安全约束。"""
def __init__(self, code: str, message: str) -> None:
super().__init__(message)
self.code = code
@dataclass(frozen=True, slots=True)
class MaintenanceDatabaseTarget:
host: str
port: int
database: str
username: str | None
sanitized_url: str
exact_target: str
is_disposable: bool
def _raise_target_error(code: str, message: str) -> None:
raise MaintenanceDatabaseTargetError(code, message)
def _normalize_host(value: str) -> str:
normalized = str(value or "").strip().lower()
if normalized.startswith("[") and normalized.endswith("]"):
normalized = normalized[1:-1]
return normalized.rstrip(".")
def _normalize_probe_component(value: str) -> str:
return re.sub(r"[^a-z0-9]+", "-", str(value or "").lower()).strip("-")
def _matches_disposable_marker(value: str) -> bool:
normalized = _normalize_probe_component(value)
return any(
normalized == marker
or normalized.startswith(f"{marker}-")
or normalized.startswith(f"x-financial-{marker}-")
for marker in DISPOSABLE_TARGET_MARKERS
)
def is_disposable_maintenance_target(*, host: str, database: str) -> bool:
"""只有主机名和数据库名都带显式 probe 标记时才视为一次性目标。"""
return _matches_disposable_marker(host) and _matches_disposable_marker(database)
def _sanitize_query(url: URL) -> URL:
sanitized_query: dict[str, Any] = {}
for key, value in url.query.items():
normalized_key = str(key).lower()
if any(part in normalized_key for part in SENSITIVE_QUERY_KEY_PARTS):
if isinstance(value, tuple):
sanitized_query[key] = tuple("***" for _ in value)
else:
sanitized_query[key] = "***"
continue
sanitized_query[key] = value
return url.set(query=sanitized_query)
def _format_target_host(host: str) -> str:
return f"[{host}]" if ":" in host else host
def parse_maintenance_database_target(
database_url: str,
*,
expected_host: str,
expected_database: str,
) -> MaintenanceDatabaseTarget:
"""解析显式数据库 URL并核对操作人声明的目标主机和数据库。"""
raw_url = str(database_url or "").strip()
if not raw_url:
_raise_target_error(
"database_url_required",
"维护命令必须显式提供 DATABASE_URL禁止回退到环境文件或默认配置。",
)
try:
parsed_url = make_url(raw_url)
except (ArgumentError, TypeError, ValueError) as exc:
raise MaintenanceDatabaseTargetError(
"invalid_database_url",
"DATABASE_URL 不是有效的 SQLAlchemy 数据库 URL。",
) from exc
if parsed_url.get_backend_name() != "postgresql":
_raise_target_error(
"postgresql_required",
"维护命令只允许连接 PostgreSQL 数据库。",
)
routing_query_keys = sorted(
str(key).lower()
for key in parsed_url.query
if str(key).lower() in FORBIDDEN_ROUTING_QUERY_KEYS
)
if routing_query_keys:
_raise_target_error(
"database_routing_query_forbidden",
f"DATABASE_URL 查询参数不得覆盖连接目标或 schema{', '.join(routing_query_keys)}",
)
host = _normalize_host(parsed_url.host or "")
if not host:
_raise_target_error(
"database_host_required",
"DATABASE_URL 必须包含显式 PostgreSQL 主机名。",
)
database = str(parsed_url.database or "").strip()
if not database:
_raise_target_error(
"database_name_required",
"DATABASE_URL 必须包含显式数据库名。",
)
normalized_expected_host = _normalize_host(expected_host)
if not normalized_expected_host:
_raise_target_error(
"expected_host_required",
"必须通过 expected_host 声明预期数据库主机。",
)
if host != normalized_expected_host:
_raise_target_error(
"expected_host_mismatch",
f"DATABASE_URL 主机与预期不一致actual={host}; expected={normalized_expected_host}",
)
normalized_expected_database = str(expected_database or "").strip()
if not normalized_expected_database:
_raise_target_error(
"expected_database_required",
"必须通过 expected_database 声明预期数据库名。",
)
if database != normalized_expected_database:
_raise_target_error(
"expected_database_mismatch",
"DATABASE_URL 数据库名与预期不一致:"
f"actual={database}; expected={normalized_expected_database}",
)
port = int(parsed_url.port or DEFAULT_POSTGRESQL_PORT)
exact_target = f"{_format_target_host(host)}:{port}/{database}"
sanitized_url = _sanitize_query(parsed_url).render_as_string(hide_password=True)
return MaintenanceDatabaseTarget(
host=host,
port=port,
database=database,
username=parsed_url.username,
sanitized_url=sanitized_url,
exact_target=exact_target,
is_disposable=is_disposable_maintenance_target(host=host, database=database),
)
def validate_maintenance_database_target(
database_url: str,
*,
expected_host: str,
expected_database: str,
apply: bool = False,
allow_non_disposable: bool = False,
confirm_target: str | None = None,
) -> MaintenanceDatabaseTarget:
"""校验维护目标;非一次性数据库 apply 必须显式放行并精确确认目标。"""
target = parse_maintenance_database_target(
database_url,
expected_host=expected_host,
expected_database=expected_database,
)
if not apply:
return target
normalized_confirmation = str(confirm_target or "").strip()
if target.is_disposable:
if normalized_confirmation and normalized_confirmation != target.exact_target:
_raise_target_error(
"confirm_target_mismatch",
"confirm_target 与解析后的数据库目标不一致:"
f"actual={normalized_confirmation}; expected={target.exact_target}",
)
return target
if not allow_non_disposable:
_raise_target_error(
"non_disposable_apply_forbidden",
"非一次性数据库 apply 必须显式启用 allow_non_disposable。",
)
if normalized_confirmation != target.exact_target:
_raise_target_error(
"confirm_target_mismatch",
"非一次性数据库 apply 必须通过 confirm_target 精确确认:"
f"expected={target.exact_target}",
)
return target