xsl/meida_backend
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
53 Commits 1 Branch 0 Tags
cb8a1e3c0f3c500470106009a0454fea15122b0d
Commit Graph
5 Commits
Author SHA1 Message Date
xsl cb8a1e3c0f fix(report): FE 问卷键名双侧兼容 + worker 持久事件循环（live E2E 两处修复）
CI / lint / typecheck / import-linter / tests / secrets (push) Canceled after 0s
Details 
FE↔后端联调实测暴露并修复：
1. 问卷键名三处错位（FE height_cm/weight_kg/psych_style vs 文档
   height/weight/psyche_style）——vision._bmi 多键取值、diagnosis 维度
   装配与规则上下文双键或运算，两侧谁改键名都不再炸
2. celery worker asyncio.run 每投递起新 loop + 模块级引擎缓存 →
   asyncpg 连接跨 loop（同进程第 3 个报告任务必现 RuntimeError:
   attached to a different loop；集成测试单循环掩盖）。新增
   apps/worker/loop.py 持久 loop，report/ping 两任务接入

E2E 终态： psyche_style/height_weight 恢复、BMI 体型确定值、ok 26/32、
真实 LLM 文案通过校验 degraded=False。
 2026-08-15 01:05:10 +08:00
xsl da41e9ab0d chore(worker): register report_tasks in celery include（T15 合并补齐） 
T15 按纪律禁改 celery_app.py，在偏差说明中移交：真 worker 进程需加载
apps.worker.report_tasks，否则 meida.report.generate 只能靠 send_task
按名投递而无人消费。
 2026-08-14 22:19:55 +08:00
xsl e8208d2972 feat(report): orchestration skeleton with staged pipeline and refund-on-failure 
POST /reports 骨架：资格校验→终身配额预扣→tk_task→cpu 队列；动态装配
流水线（dotted-path 注册、缺阶段即 CONFIG_ERROR）；中间态逐段落库、失败
退款、derived_features 跨重试/重生成复用（T-P3-07/09）。

实现：packages/report 的 pipeline.py（Stage 契约保持不变，新增 assemble_stages/
run_pipeline/StageFailureError，ReportContext 增可选 task_id 参数、向后兼容）、
repo.py、schemas.py、service.py（重放优先于资格校验、reserve+建行+建任务同事务、
提交后才投递、并发同键落败回滚）；apps/api/routers/reports.py 填桩（send_task 按
任务名投递，不 import apps.worker）；apps/worker/report_tasks.py（worker 侧自建
gateway 装配、状态归一 QUEUED→RUNNING、成功 commit 配额+成本聚合写 model_info、
基础设施异常重试一次、重试耗尽兜底失败退款）。测试：单测 6 个（装配缺失/空配置/
协议违约/顺序执行/失败保留产物），集成 6 场景（预扣 lifetime/PENDING、成功
committed+两阶段键、同 request_key 幂等不重复扣、二次首生成 40920、阶段失败
refund+derived_features 保留、regenerate 走 m: 月配额 version+1 复用视觉中间态）。

偏差说明：
1) apps/worker/celery_app.py 的 include 只含 apps.worker.tasks（禁改文件），真
worker 进程注册 meida.report.generate 需协调者合并时把 include 补上
apps.worker.report_tasks；API 侧已按任务名 send_task，两侧只靠名字衔接，测试
不受影响。2) ReportContext 增加可选 task_id 关键字参数（失败结案要推动
tk_task），Stage 协议/DEFAULT_STAGE_PATHS/stage_names 未动，T16/T17 兼容。
3) JSONB 深变更用 flag_modified 显式标脏（重新赋相等新 dict 会被 flush 的按值
比较跳过，集成测试抓到过该坑）。4) POST 后 tk_task 保持 PENDING（任务文档
§4 场景 1 口径），QUEUED→RUNNING 由 worker 入口归一。
 2026-08-14 22:18:14 +08:00
xsl f46ecc5eb0 fix(llm): price by call time and stop paying for thinking tokens 
Two defects that only surface against the real DeepSeek API, both of which
break the cost model rather than the code:

deepseek-v4-flash enables thinking mode by default. The same single-turn
question took 13.5s with it on and 1.8s with it off, and the thinking tokens
are billed at the output rate. That misses NFR-PERF-02b (15s per turn) and
inflates every conversational call. SceneConfig now carries an explicit
`thinking` flag, all four V1 scenes set it to false, and the provider always
sends the parameter rather than relying on the upstream default. Reasoning
tokens are recorded separately in Usage.tokens_reasoning so the cost of ever
turning it back on is visible.

DeepSeek moves to peak/off-peak pricing on 2026-08-17, with peak (Beijing
09:00-12:00 and 14:00-18:00) at twice the off-peak rate. Reports are generated
mostly during peak hours, so a single blended rate would leave ob_ai_call --
the only input to V2 pricing -- off by nearly a factor of two.
PeakAwareRateTable resolves the rate at call time; cost is computed and stored
then, never recomputed later at whatever the current rate happens to be.

Also here:
- Replace the live API key in the redaction test with a synthetic one. The
  test that proves we redact secrets was itself a secret leak waiting for its
  first commit.
- Correct LLM_COST_RATES_JSON, which held placeholder rates.
- Make ids.SnowflakeGenerator take an injectable clock so rollover and
  clock-rollback are testable without patching a slots class.
- Rename exceptions to the -Error suffix (N818) and clear the remaining ruff,
  mypy and import-linter findings, including two dead lines: a duplicate
  pytestmark assignment and a no-op assert on a None return.

Gates: ruff, mypy (96 files incl. tests), 4/4 import contracts,
detect-secrets, 166 unit + 3 integration tests, migration up/down round-trip.
 2026-08-14 00:54:40 +08:00
xsl 746b66da01 feat(api): vertical slice POST /v1/agent/chat and the Celery ping path 
One request exercises every layer: JWT auth, quota reservation, a versioned
prompt through the LLM gateway, a cost row, quota commit or refund, and the
unified envelope. The trivial ping task proves the queue routing and the
tk_task transitions before any real long-running task is written.

Routes are mounted at the /api/v1 base path of 架构文档 §9.1; /v1/* is
rewritten to it at the ASGI layer so both spellings work from one route table.
 2026-08-14 00:26:39 +08:00