Commit Graph

10 Commits

Author SHA1 Message Date
39338df808 feat(kg): 实现 Phase 2 GraphRAG 融合功能
核心功能:
- 三层检索策略:向量检索(Milvus)+ 图检索(KG 服务)+ 融合排序
- LLM 生成:支持同步和流式(SSE)响应
- 知识库访问控制:knowledge_base_id 归属校验 + collection_name 绑定验证

新增模块(9个文件):
- models.py: 请求/响应模型(GraphRAGQueryRequest, RetrievalStrategy, GraphContext 等)
- milvus_client.py: Milvus 向量检索客户端(OpenAI Embeddings + asyncio.to_thread)
- kg_client.py: KG 服务 REST 客户端(全文检索 + 子图导出,fail-open)
- context_builder.py: 三元组文本化(10 种关系模板)+ 上下文构建
- generator.py: LLM 生成(ChatOpenAI,支持同步和流式)
- retriever.py: 检索编排(并行检索 + 融合排序)
- kb_access.py: 知识库访问校验(归属验证 + collection 绑定,fail-close)
- interface.py: FastAPI 端点(/query, /retrieve, /query/stream)
- __init__.py: 模块入口

修改文件(3个):
- app/core/config.py: 添加 13 个 graphrag_* 配置项
- app/module/__init__.py: 注册 kg_graphrag_router
- pyproject.toml: 添加 pymilvus 依赖

测试覆盖(79 tests):
- test_context_builder.py: 13 tests(三元组文本化 + 上下文构建)
- test_kg_client.py: 14 tests(KG 响应解析 + PagedResponse + 边字段映射)
- test_milvus_client.py: 8 tests(向量检索 + asyncio.to_thread)
- test_retriever.py: 11 tests(并行检索 + 融合排序 + fail-open)
- test_kb_access.py: 18 tests(归属校验 + collection 绑定 + 跨用户负例)
- test_interface.py: 15 tests(端点级回归 + 403 short-circuit)

关键设计:
- Fail-open: Milvus/KG 服务失败不阻塞管道,返回空结果
- Fail-close: 访问控制失败拒绝请求,防止授权绕过
- 并行检索: asyncio.gather() 并发运行向量和图检索
- 融合排序: Min-max 归一化 + 加权融合(vector_weight/graph_weight)
- 延迟初始化: 所有客户端在首次请求时初始化
- 配置回退: graphrag_llm_* 为空时回退到 kg_llm_*

安全修复:
- P1-1: KG 响应解析(PagedResponse.content)
- P1-2: 子图边字段映射(sourceEntityId/targetEntityId)
- P1-3: collection_name 越权风险(归属校验 + 绑定验证)
- P1-4: 同步 Milvus I/O(asyncio.to_thread)
- P1-5: 测试覆盖(79 tests,包括安全负例)

测试结果:79 tests pass 
2026-02-20 09:41:55 +08:00
5a553ddde3 feat(knowledge-graph): 实现知识图谱基础设施搭建
实现功能:
- Neo4j Docker Compose 配置(社区版,端口 7474/7687,数据持久化)
- Makefile 新增 Neo4j 命令(neo4j-up/down/logs/shell)
- knowledge-graph-service Spring Boot 服务(完整的 DDD 分层架构)
- kg_extraction Python 模块(基于 LangChain LLMGraphTransformer)

技术实现:
- Neo4j 配置:环境变量化密码,统一默认值 datamate123
- Java 服务:
  - Domain: GraphEntity, GraphRelation 实体模型
  - Repository: Spring Data Neo4j,支持 graphId 范围查询
  - Service: 业务逻辑,graphId 双重校验,查询限流
  - Controller: REST API,UUID 格式校验
  - Exception: 实现 ErrorCode 接口,统一异常体系
- Python 模块:
  - KnowledgeGraphExtractor 类
  - 支持异步/同步/批量抽取
  - 支持 schema-guided 模式
  - 兼容 OpenAI 及自部署模型

关键设计:
- graphId 权限边界:所有实体操作都在正确的 graphId 范围内
- 查询限流:depth 和 limit 参数受配置约束
- 异常处理:统一使用 BusinessException + ErrorCode
- 凭据管理:环境变量化,避免硬编码
- 双重防御:Controller 格式校验 + Service 业务校验

代码审查:
- 经过 3 轮 Codex 审查和 2 轮 Claude 修复
- 所有 P0 和 P1 问题已解决
- 编译通过,无阻塞性问题

文件变更:
- 新增:Neo4j 配置、knowledge-graph-service(11 个 Java 文件)、kg_extraction(3 个 Python 文件)
- 修改:Makefile、pom.xml、application.yml、pyproject.toml
2026-02-17 20:42:55 +08:00
f8f9faaa06 feat(runtime): 添加 Pillow 图像处理库依赖
- 在 poetry.lock 中新增 pillow 包及其所有版本的依赖信息
- 支持多种操作系统平台包括 macOS、Linux、Windows 和 iOS
- 支持多种架构包括 x86_64、arm64、aarch64 和 win32
- 支持 Python 版本从 3.9 到 3.14 的完整兼容性
- 包含多种文件格式如 wheel 和 tar.gz 校验信息
- 添加了文档、测试和类型检查相关的可选依赖配置
2026-02-06 13:21:01 +08:00
b5d7c66240 feat(data-management): 扩展源文档排除功能支持Excel文件类型
- 在后端服务中扩展源文档类型检查,新增对XLS和XLSX文件的支持
- 修改DatasetFileApplicationService中的过滤逻辑,统一处理所有源文档类型
- 新增isSourceDocument和isDerivedFile辅助方法进行文件类型判断
- 更新前端DatasetFileTransfer组件中的注释说明
- 在Python运行时依赖中添加openpyxl和xlrd库以支持Excel文件处理
- 修改标注项目接口中源文档类型的集合定义
- 更新文件操作钩子中的派生文件排除逻辑
2026-01-31 11:30:55 +08:00
Dallas98
cbb146d3d7 feat(chart): add Helm chart for deploying Label Studio with PostgreSQL (#152)
* feat(chart): add Helm chart for deploying Label Studio with PostgreSQL

* feat(milvus): update Milvus configuration to use URI and remove deprecated host/port settings
2025-12-10 17:46:12 +08:00
Dallas98
6ccbc8a02f feat(dependencies): update greenlet to version 3.3.0 and adjust Python version compatibility (#145) 2025-12-10 10:17:48 +08:00
hefanli
1d19cd3a62 feature: add data-evaluation
* feature: add evaluation task management function

* feature: add evaluation task detail page

* fix: delete duplicate definition for table t_model_config

* refactor: rename package synthesis to ratio

* refactor: add eval file table and  refactor related code

* fix: calling large models in parallel during evaluation
2025-12-04 09:23:54 +08:00
Dallas98
bcd1bc1534 feat: Update Python version requirements and refine dependency constraints in configuration (#123)
* feat: Implement data synthesis task management with database models and API endpoints

* feat: Update Python version requirements and refine dependency constraints in configuration
2025-12-02 16:34:22 +08:00
Dallas98
8b164cb012 feat: Implement data synthesis task management with database models and API endpoints (#122) 2025-12-02 15:23:58 +08:00
Jason Wang
45743f39f5 feat: add labeling template. refactor: switch to Poetry, build and deploy of backend Python (#79)
* feat: Enhance annotation module with template management and validation

- Added DatasetMappingCreateRequest and DatasetMappingUpdateRequest schemas to handle dataset mapping requests with camelCase and snake_case support.
- Introduced Annotation Template schemas including CreateAnnotationTemplateRequest, UpdateAnnotationTemplateRequest, and AnnotationTemplateResponse for managing annotation templates.
- Implemented AnnotationTemplateService for creating, updating, retrieving, and deleting annotation templates, including validation of configurations and XML generation.
- Added utility class LabelStudioConfigValidator for validating Label Studio configurations and XML formats.
- Updated database schema for annotation templates and labeling projects to include new fields and constraints.
- Seeded initial annotation templates for various use cases including image classification, object detection, and text classification.

* feat: Enhance TemplateForm with improved validation and dynamic field rendering; update LabelStudio config validation for camelCase support

* feat: Update docker-compose.yml to mark datamate dataset volume and network as external

* feat: Add tag configuration management and related components

- Introduced new components for tag selection and browsing in the frontend.
- Added API endpoint to fetch tag configuration from the backend.
- Implemented tag configuration management in the backend, including loading from YAML.
- Enhanced template service to support dynamic tag rendering based on configuration.
- Updated validation utilities to incorporate tag configuration checks.
- Refactored existing code to utilize the new tag configuration structure.

* feat: Refactor LabelStudioTagConfig for improved configuration loading and validation

* feat: Update Makefile to include backend-python-docker-build in the build process

* feat: Migrate to poetry for better deps management

* Add pyyaml dependency and update Dockerfile to use Poetry for dependency management

- Added pyyaml (>=6.0.3,<7.0.0) to pyproject.toml dependencies.
- Updated Dockerfile to install Poetry and manage dependencies using it.
- Improved layer caching by copying only dependency files before the application code.
- Removed unnecessary installation of build dependencies to keep the final image size small.

* feat: Remove duplicated backend-python-docker-build target from Makefile

* fix: airflow is not ready for adding yet

* feat: update Python version to 3.12 and remove project installation step in Dockerfile
2025-11-13 15:32:30 +08:00