feat(p1): complete P1 foundation stage
Some checks failed
CI Go / test (push) Has been cancelled
CI Proto / lint (push) Has been cancelled
CI Python / test (push) Has been cancelled
CI TypeScript / test (push) Has been cancelled

- monorepo: pnpm workspace + go.work + pyproject.toml + commitlint/husky
- infra: docker-compose (minimal + full profiles) + init-sql + prometheus
- arch-scan: multi-language scanner skeleton (TS/Go/Python/Proto)
- shared-proto: buf v2 + classes.proto (ClassService CRUD contract)
- api-gateway: Go/Gin + JWT HS256 auth + reverse proxy + request ID
- classes: NestJS golden template (error system + observability + middleware + CRUD + tests)
- teacher-portal: Next.js + paper-feel UI design system
- CI/CD: 4 workflows (go/ts/py/proto)
- docs: migration guide + project_rules + coding-standards + git-workflow + ui-design-system + 004 + 9 module READMEs + known-issues + spec/plan migration + roadmap
This commit is contained in:
SpecialX
2026-07-07 23:39:37 +08:00
commit 2ba4250165
100 changed files with 15242 additions and 0 deletions

View File

@@ -0,0 +1,104 @@
# data-ana 数据分析服务
> 版本0.1(骨架)
> 日期2026-07-07
> 状态P4 待交付
> 关联文档:[架构影响地图](../../architecture/004_architecture_impact_map.md)
---
## 1. 模块职责
data-ana 是数据分析限界上下文Python 实现),管理 Report、Mastery、Dashboard 聚合。
消费 core-edu 与 content 的领域事件,构建 ClickHouse 学情宽表,计算知识点掌握度。
对外提供学情仪表盘查询、班级/年级/学校维度报表、个性化推荐数据支撑。
---
## 2. 技术栈
| 类别 | 技术 | 版本 | 用途 |
|------|------|------|------|
| 语言 | Python | 3.12+ | 数据分析 |
| 框架 | FastAPI | 0.110+ | 异步 API |
| ORM | SQLAlchemy | 2.x | ClickHouse 查询 |
| 分析库 | ClickHouse | - | 学情宽表、实时聚合 |
| 任务 | Celery | 5.x | 异步计算任务 |
| 校验 | pydantic | 2.x | 运行时校验 |
| 追踪 | OpenTelemetry | - | 分布式追踪 |
---
## 3. 架构图
```mermaid
graph TB
BFF[BFF 聚合层] -->|gRPC| ANA[data-ana 服务]
K[(Kafka)] -->|消费事件| ANA
ANA --> PROJ[Projection<br/>事件 → ClickHouse]
ANA --> CALC[MasteryCalculator<br/>掌握度计算]
ANA --> RPT[ReportGenerator<br/>报表生成]
ANA -->[(ClickHouse<br/>学情宽表)]
ANA -.事件.-> K2[(Kafka<br/>edu.insight.mastery.updated)]
```
> 待 P4 交付时补充完整的分层架构图与事件消费管线图。
---
## 4. 核心流程图
```mermaid
sequenceDiagram
participant K as Kafka
participant ANA as data-ana
participant CH as ClickHouse
participant BFF as BFF
K->>ANA: edu.teaching.grade.recorded
ANA->>ANA: Projection 写入宽表
ANA->>ANA: 触发掌握度计算Celery
ANA->>CH: 更新 mastery 表
ANA->>K: 发布 edu.insight.mastery.updated
BFF->>ANA: gRPC GetDashboard(studentId)
ANA->>CH: 查询宽表聚合
ANA-->>BFF: DashboardDTO
```
> 待 P4 交付时补充报表生成流程、掌握度算法流程、CDC 同步流程。
---
## 5. 对外契约
- **gRPC 服务**DashboardService、MasteryService、ReportService待 P4 定义 protobuf
- **消费事件**`edu.teaching.*``edu.content.*`
- **发布事件**`edu.insight.mastery.updated`
- **ClickHouse 宽表**student_wide_table、mastery_table、exam_analysis_table
> 待 P4 交付时补充完整 protobuf 定义与宽表 schema。
---
## 6. 依赖关系
- **上游**teacher-bff、student-bff、parent-bff、aigRPC 查询学情)
- **下游**ClickHouse、Kafka消费与发布、Celery异步任务
- **跨服务**:消费 core-edu/content 事件;发布掌握度事件被 core-edu 消费触发推荐
---
## 7. 架构约束
1. 查询必须走 ClickHouse 宽表,禁止实时计算
2. 事件消费必须幂等,基于 event_id 去重
3. 掌握度计算必须异步Celery不阻塞事件消费
4. 宽表查询响应时间 ≤ 5 秒
> 待 P4 交付时补充完整约束清单。
---
## 8. 架构决策
> 待 P4 交付时补充 ADR 记录预计包括Python+FastAPI 选型、ClickHouse 选型、Celery 异步框架选型、掌握度算法选型。