283 lines
13 KiB
Python
283 lines
13 KiB
Python
"""pipeline_llm.accounting — 异步出账循环（owner 模型三方账）。
|
||
|
||
扫 llm_usage 中 accounting_status='created' 且 status='SUCCEEDED' 的用量流水，
|
||
逐条映射模型→产品，调产品层 product_accounting_generic 记账
|
||
（产品层内部按 is_self_use 分流：自用只记 PAY* 采购成本；跨机构另记 PAY 客户应付，
|
||
折扣精确到产品），成功置 'accounted'，失败置 'failed'（三态：created/accounted/failed）。
|
||
2026-09-06 用户定夺：status 统一枚举 SUCCEEDED/FAILED/PENDING/RUNNING；
|
||
只有 SUCCEEDED 的行才写 accounting_status（其余 NULL），出账循环自然只拾取 SUCCEEDED。
|
||
|
||
与账号/存储的同步记账分工明确（2026-09 用户定夺）：
|
||
- 账号/存储：purchase_realtime 同步记账
|
||
- 模型（本模块）：调用时只记成本侧 + created 流水，出账走本循环（异步）
|
||
|
||
单循环守卫（多进程部署时防重复出账）：Redis SETNX llm_acc:lock（db4，与限流同库）。
|
||
"""
|
||
|
||
import asyncio
|
||
import json
|
||
import logging
|
||
import time
|
||
from datetime import datetime
|
||
|
||
from ahserver.serverenv import ServerEnv
|
||
|
||
from .gateway import _get_db, _redis
|
||
|
||
logger = logging.getLogger("pipeline_llm.accounting")
|
||
|
||
_SCAN_INTERVAL = 60 # 扫描间隔（秒）
|
||
_BATCH = 200 # 每批处理条数
|
||
_LOCK_KEY = 'llm_acc:lock'
|
||
_LOCK_TTL = 90 # 锁 TTL（略大于单批预算耗时）
|
||
|
||
_trace_cleanup_day = '' # trace 过期清理已跑日期（进程内每日一次）
|
||
|
||
|
||
async def _get_pending(sor):
|
||
recs = await sor.sqlExe(
|
||
"SELECT id, org_id, user_id, model_id, account_id, cost, ppid, usages, "
|
||
"created_at "
|
||
"FROM llm_usage WHERE accounting_status='created' AND status='SUCCEEDED' "
|
||
"ORDER BY created_at LIMIT %d" % _BATCH, {})
|
||
await sor.sqlExe("COMMIT", {})
|
||
return recs or []
|
||
|
||
|
||
async def _deduct_wallet(sor, row, supplier_cost):
|
||
"""产品记账完成后扣减调用账号的钱包（2026-09-11 用户定夺）。
|
||
|
||
钱包 = llm_account.balance。1 个供应商财务账户(accounting.account) : n 个
|
||
llm_account 钱包——一个供应商可配多账号(钱包)，扣减主体按流水 account_id
|
||
定位「**调用模型的那个钱包**」，扣减金额 = accounting 对该供应商的记账额
|
||
（product_accounting_generic 返回的 supplier_cost，即 PAY* 分录的采购成本）。
|
||
|
||
幂等守卫：流水 cost 列在出账成功时回填 supplier_cost（govern_settle 创建时
|
||
cost=0）。cost≠0 ⟺ 本行已扣过钱包——accounted 行被手动重置回 created 重跑时
|
||
跳过扣减，防重复扣。（failed→created 追回流程安全：failed 行在扣减前就
|
||
return False，从未回填 cost，重跑正常扣。）
|
||
|
||
允许扣成负值：调用→出账有 ~60s 窗口且门禁阈值(缺省5元)是事前闸，并发超扣
|
||
是真实成本，宁可透支为负（待办 provider 派生待办催充值），不可丢失钱流或把
|
||
已记账流水打回 failed（账务不可逆）。扣减失败只 warning 不回滚客户侧记账。
|
||
"""
|
||
try:
|
||
prior_cost = float(getattr(row, 'cost', 0) or 0)
|
||
except (TypeError, ValueError):
|
||
prior_cost = 0.0
|
||
if prior_cost != 0:
|
||
logger.info("pipeline_llm.accounting: 流水 %s 已回填 cost=%.6f（钱包已扣），"
|
||
"幂等跳过重复扣减", getattr(row, 'id', ''), prior_cost)
|
||
return
|
||
account_id = getattr(row, 'account_id', '') or ''
|
||
if not account_id:
|
||
return
|
||
try:
|
||
amount = float(supplier_cost or 0)
|
||
except (TypeError, ValueError):
|
||
amount = 0.0
|
||
if amount <= 0:
|
||
return
|
||
try:
|
||
# 只发 UPDATE 不单独 COMMIT——与调用方回填 cost/accounted 的 UPDATE 同一
|
||
# 事务提交（原子）：钱包扣减与 cost 回填要么都落要么都不落，杜绝
|
||
# "已扣钱但 cost 未回填→重跑重复扣" 的窗口。
|
||
await sor.sqlExe(
|
||
"UPDATE llm_account SET balance=balance-${c}$ WHERE id=${i}$",
|
||
{"c": round(amount, 6), "i": account_id})
|
||
except Exception as e:
|
||
logger.warning("pipeline_llm.accounting: 账号钱包扣减失败（不回滚记账，对账修正）"
|
||
"account=%s cost=%.6f: %s", account_id, amount, str(e)[:160])
|
||
|
||
|
||
async def _find_product_id(sor, model_id):
|
||
"""llm_model.id → product.id（产品层导入时按 resource_ref_id 建立）。"""
|
||
recs = await sor.sqlExe(
|
||
"SELECT id FROM product WHERE resource_ref_id=${r}$ AND status='1' LIMIT 1",
|
||
{"r": model_id})
|
||
await sor.sqlExe("COMMIT", {})
|
||
return (getattr(recs[0], 'id', '') or '') if recs else ''
|
||
|
||
|
||
async def _settle_one(sor, row):
|
||
"""单条出账。成功置 accounted；失败置 failed 并记录原因（不静默重试死循环）。"""
|
||
env = ServerEnv()
|
||
model_id = getattr(row, 'model_id', '') or ''
|
||
product_id = await _find_product_id(sor, model_id)
|
||
if not product_id:
|
||
await sor.U('llm_usage', {
|
||
'id': row.id, 'accounting_status': 'failed',
|
||
'note': '出账失败：模型(%s)未映射到产品，请先在产品管理导入产线模型' % model_id,
|
||
})
|
||
await sor.sqlExe("COMMIT", {})
|
||
logger.warning("pipeline_llm.accounting: model %s 无产品映射，流水 %s 置 failed",
|
||
model_id, row.id)
|
||
return False
|
||
# 计费唯一事实源 = usages JSON 全量原文（2026-09-10 用户定夺，req/resp_tokens
|
||
# 冗余列已删 m0021）：token 计数（prompt_tokens/completion_tokens/
|
||
# prompt_tokens_details）与按量因子（时长/分辨率/张数）都在其中。
|
||
usage_data: dict = {}
|
||
usages_raw = getattr(row, 'usages', '') or ''
|
||
if usages_raw:
|
||
try:
|
||
extra = json.loads(usages_raw)
|
||
if isinstance(extra, dict):
|
||
usage_data.update(extra)
|
||
except Exception:
|
||
# usages 是计费事实源，解析失败必须显式记账失败（写明原因），
|
||
# 禁止静默空 dict 兜底成 0 元账（历史上 [:2000] 截断即此病根）
|
||
await sor.U('llm_usage', {
|
||
'id': row.id, 'accounting_status': 'failed',
|
||
'note': '出账失败：usages 不是合法 JSON（长度%d），无法计费——'
|
||
'请核查流水写入方是否截断' % len(usages_raw),
|
||
})
|
||
await sor.sqlExe("COMMIT", {})
|
||
logger.warning("pipeline_llm.accounting: 流水 %s usages 非法 JSON，置 failed", row.id)
|
||
return False
|
||
if not usage_data:
|
||
await sor.U('llm_usage', {
|
||
'id': row.id, 'accounting_status': 'failed',
|
||
'note': '出账失败：usages 为空，无计费事实源——请核查上游 usage 返回',
|
||
})
|
||
await sor.sqlExe("COMMIT", {})
|
||
logger.warning("pipeline_llm.accounting: 流水 %s usages 为空，置 failed", row.id)
|
||
return False
|
||
# derived 依赖归一（2026-09-07 根治）：token 定价的 cached_tokens/uncache_tokens
|
||
# 靠 YAML derived 变量引用 prompt_tokens_details.cached_tokens 计算——该键缺失时
|
||
# DictObject 属性链抛 AttributeError，引擎兜底成 0，输入 token 永不计费
|
||
# （实测 qwen3.8-max uncache_tokens=0 记账失败根因之一）。缺失补 0 结构，
|
||
# derived 正常求值：无缓存时 uncache=prompt_tokens、cached=0。
|
||
ptd = usage_data.get('prompt_tokens_details')
|
||
if not isinstance(ptd, dict):
|
||
usage_data['prompt_tokens_details'] = {'cached_tokens': 0}
|
||
elif ptd.get('cached_tokens') is None:
|
||
ptd['cached_tokens'] = 0
|
||
# 忙闲时计价维度（2026-09-07）：价目表存在忙时(8点-22点)/闲时(22点-次日8点)
|
||
# 两档定价（deepseek-v4-pro-0813），按调用时刻 llm_usage.created_at 的小时数注入
|
||
# hour——YAML 用 hour 区间 filters 选档。多余键对不用它的定价方案无影响
|
||
# （引擎只严格检查定价项里出现的维度键）。
|
||
# 2026-09-10：注入浮点小时（时+分/60+秒/3600），支持 pricing 引擎 peak_times
|
||
# 时段串（如 "9:00 =~= 12:00"）的精确时刻语义；对存量整数小时区间 YAML
|
||
# （8 =~ 22 等）边界语义不变（整点值相同）。
|
||
created_at = getattr(row, 'created_at', None)
|
||
if created_at is not None:
|
||
try:
|
||
usage_data['hour'] = (int(created_at.hour)
|
||
+ int(created_at.minute) / 60.0
|
||
+ int(created_at.second) / 3600.0)
|
||
except Exception:
|
||
pass
|
||
fn = getattr(env, 'product_accounting_generic', None)
|
||
if fn is None:
|
||
# 产品模块未加载 → 保留 created 下轮再试
|
||
return False
|
||
result = await fn(
|
||
product_id=product_id,
|
||
usage_data=usage_data,
|
||
userorgid=getattr(row, 'org_id', '') or '',
|
||
userid=getattr(row, 'user_id', '') or '',
|
||
)
|
||
if not result or not result.get('success', result.get('orderid')):
|
||
msg = str((result or {}).get('message', result))[:180]
|
||
await sor.U('llm_usage', {
|
||
'id': row.id, 'accounting_status': 'failed',
|
||
'note': '出账失败：%s' % msg,
|
||
})
|
||
await sor.sqlExe("COMMIT", {})
|
||
logger.error("pipeline_llm.accounting: 流水 %s 出账失败: %s", row.id, msg)
|
||
return False
|
||
# 产品记账完成（2026-09-11 用户定夺）：记账成功之后扣减调用账号的钱包——
|
||
# supplier_cost 是定价引擎算出的真实进货成本，扣 llm_account.balance（流水
|
||
# account_id 定位"调用模型的钱包"）；同时把 cost 回填进流水（对账可查每笔
|
||
# 钱包扣了多少）。扣减失败只 warning，不回滚客户侧账务。
|
||
supplier_cost = float((result or {}).get('supplier_cost', 0) or 0)
|
||
await _deduct_wallet(sor, row, supplier_cost)
|
||
# 回填客户应付 + 成本侧金额 + 置已出账
|
||
await sor.sqlExe(
|
||
"UPDATE llm_usage SET charge=${c}$, cost=${k}$, accounting_status='accounted', "
|
||
"note=CONCAT(COALESCE(note,''),'|accounted:',${o}$) WHERE id=${i}$",
|
||
{"c": float((result or {}).get('customer_amount', 0) or 0),
|
||
"k": round(supplier_cost, 6),
|
||
"o": str((result or {}).get('orderid', ''))[:40],
|
||
"i": row.id})
|
||
await sor.sqlExe("COMMIT", {})
|
||
return True
|
||
|
||
|
||
async def _run_one_pass():
|
||
"""一轮扫描-出账。返回处理条数。"""
|
||
db, dbname = _get_db()
|
||
async with db.sqlorContext(dbname) as sor:
|
||
rows = await _get_pending(sor)
|
||
if not rows:
|
||
return 0
|
||
ok_n = 0
|
||
for row in rows:
|
||
try:
|
||
db, dbname = _get_db()
|
||
async with db.sqlorContext(dbname) as sor:
|
||
if await _settle_one(sor, row):
|
||
ok_n += 1
|
||
except Exception as e:
|
||
logger.error("pipeline_llm.accounting: 流水 %s 出账异常: %s",
|
||
getattr(row, 'id', ''), e)
|
||
try:
|
||
db, dbname = _get_db()
|
||
async with db.sqlorContext(dbname) as sor:
|
||
await sor.U('llm_usage', {
|
||
'id': row.id, 'accounting_status': 'failed',
|
||
'note': '出账异常：%s' % str(e)[:180],
|
||
})
|
||
await sor.sqlExe("COMMIT", {})
|
||
except Exception:
|
||
pass
|
||
return ok_n
|
||
|
||
|
||
async def _maybe_cleanup_traces():
|
||
"""trace 保留期清理：每日最多跑一次（按日期字符串判定，循环间隔 60s 足够）。
|
||
|
||
trace.py 铁律 3 的落地点——cleanup_expired 此前无人调用（文档承诺
|
||
"记账 worker 循环每日调一次"但循环里没接，2026-09-11 补齐）。
|
||
永不抛出：清理失败不反噬出账主循环。
|
||
"""
|
||
global _trace_cleanup_day
|
||
today = datetime.now().strftime('%Y%m%d')
|
||
if today == _trace_cleanup_day:
|
||
return
|
||
_trace_cleanup_day = today
|
||
try:
|
||
from .trace import cleanup_expired
|
||
await cleanup_expired()
|
||
except Exception as e:
|
||
logger.warning("pipeline_llm.accounting: trace 过期清理失败（明日重试）: %s", e)
|
||
|
||
|
||
async def llm_usage_accounting_loop():
|
||
"""后台循环：抢分布式锁 → 出账一批 → 睡。永不抛出。
|
||
|
||
只由独立记账进程（app/pipeline_accounting_worker.py）运行，
|
||
不挂进 web 服务器的事件循环（同步记账在 web，异步记账在独立进程）。
|
||
"""
|
||
logger.info("[pipeline_llm] 异步出账循环已启动（间隔 %ds）", _SCAN_INTERVAL)
|
||
while True:
|
||
try:
|
||
# trace 过期清理（每日一次；放抢锁前——锁只护出账防重复记账，
|
||
# 清理是幂等操作（目录整删+按 created_at 删行），多进程同日跑无害）
|
||
await _maybe_cleanup_traces()
|
||
r = _redis()
|
||
if r is not None:
|
||
# 抢锁（多进程只跑一个）；Redis 故障则直接跑（与限流 fail-open 一致）
|
||
got = False
|
||
try:
|
||
got = bool(r.set(_LOCK_KEY, str(time.time()), nx=True, ex=_LOCK_TTL))
|
||
except Exception:
|
||
got = True
|
||
if not got:
|
||
await asyncio.sleep(_SCAN_INTERVAL)
|
||
continue
|
||
await _run_one_pass()
|
||
except Exception as e:
|
||
logger.error("pipeline_llm.accounting: 循环异常（继续）: %s", e)
|
||
await asyncio.sleep(_SCAN_INTERVAL)
|