DataMate/backend/openapi/specs/data-cleaning.yaml

openapi: 3.0.3
info:
  title: Data Cleaning Service API
  description: 数据清洗服务API - 策略/规则、流程编排对接
  version: 1.0.0
  contact:
    name: Data Mate Platform Team

servers:
  - url: http://localhost:8084
    description: Development server

tags:
  - name: CleaningTask
    description: 数据清洗任务管理
  - name: CleaningTemplate
    description: 数据清洗模板管理

paths:
  /ray/log:
    get:
      summary: 获取ray日志文件
      deprecated: false
      description: ''
      tags: [ ]
      parameters: [ ]
      responses:
        '200':
          description: ''
          content:
            application/json:
              schema:
                type: object
                properties: { }
          headers: { }
      security: [ ]
  /cleaning/tasks:
    get:
      summary: 查询数据清洗任务列表
      deprecated: false
      description: 获取所有数据清洗任务或根据查询参数筛选任务。
      tags:
        - CleaningTask
      parameters:
        - name: status
          in: query
          description: 根据任务状态筛选 (e.g., pending, running, completed, failed)
          required: false
          schema:
            type: string
        - name: keywords
          in: query
          description: 关键字
          required: false
          schema:
            type: string
        - name: page
          in: query
          description: 分页数
          required: true
          schema:
            type: integer
        - name: size
          in: query
          description: 分页单页数
          required: true
          schema:
            type: integer
      responses:
        '200':
          description: 成功获取任务列表
          content:
            application/json:
              schema:
                type: array
                items: &ref_1
                  $ref: '#/components/schemas/CleaningTask'
          headers: { }
      security: [ ]
    post:
      summary: 创建新的数据清洗任务
      deprecated: false
      description: 可以直接创建任务或基于现有模板创建任务。
      tags:
        - CleaningTask
      parameters: [ ]
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateCleaningTaskRequest'
            examples: { }
      responses:
        '201':
          description: 任务创建成功
          content:
            application/json:
              schema: *ref_1
          headers: { }
      security: [ ]
  /cleaning/tasks/{taskId}:
    get:
      summary: 获取单个数据清洗任务详情
      deprecated: false
      description: 根据任务ID获取任务的详细信息。
      tags:
        - CleaningTask
      parameters:
        - name: taskId
          in: path
          description: 任务的唯一标识符
          required: true
          example: ''
          schema:
            type: string
      responses:
        '200':
          description: 成功获取任务详情
          content:
            application/json:
              schema: *ref_1
          headers: { }
      security: [ ]
    delete:
      summary: 删除数据清洗任务
      deprecated: false
      description: 根据任务ID删除指定的任务。
      tags:
        - CleaningTask
      parameters:
        - name: taskId
          in: path
          description: 任务的唯一标识符
          required: true
          example: ''
          schema:
            type: string
      responses:
        '204':
          description: 任务删除成功
          headers: { }
      security: [ ]
  /cleaning/templates:
    get:
      summary: 查询数据清洗模板列表
      deprecated: false
      description: 获取所有可用的数据清洗模板。
      tags:
        - CleaningTemplate
      parameters: [ ]
      responses:
        '200':
          description: 成功获取模板列表
          content:
            application/json:
              schema:
                type: array
                items: &ref_2
                  $ref: '#/components/schemas/CleaningTemplate'
          headers: { }
      security: [ ]
    post:
      summary: 创建新的数据清洗模板
      deprecated: false
      description: 定义一个新的数据清洗模板。
      tags:
        - CleaningTemplate
      parameters: [ ]
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateCleaningTemplateRequest'
      responses:
        '201':
          description: 模板创建成功
          content:
            application/json:
              schema: *ref_2
          headers: { }
      security: [ ]
  /cleaning/templates/{templateId}:
    get:
      summary: 获取单个数据清洗模板详情
      deprecated: false
      description: 根据模板ID获取模板的详细信息。
      tags:
        - CleaningTemplate
      parameters:
        - name: templateId
          in: path
          description: 模板的唯一标识符
          required: true
          example: ''
          schema:
            type: string
      responses:
        '200':
          description: 成功获取模板详情
          content:
            application/json:
              schema: *ref_2
          headers: { }
      security: [ ]
    put:
      summary: 更新数据清洗模板
      deprecated: false
      description: 根据模板ID更新模板的全部信息。
      tags:
        - CleaningTemplate
      parameters:
        - name: templateId
          in: path
          description: 模板的唯一标识符
          required: true
          example: ''
          schema:
            type: string
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/UpdateCleaningTemplateRequest'
      responses:
        '200':
          description: 模板更新成功
          content:
            application/json:
              schema: *ref_2
          headers: { }
      security: [ ]
    delete:
      summary: 删除数据清洗模板
      deprecated: false
      description: 根据模板ID删除指定的模板。
      tags:
        - CleaningTemplate
      parameters:
        - name: templateId
          in: path
          description: 模板的唯一标识符
          required: true
          example: ''
          schema:
            type: string
      responses:
        '204':
          description: 模板删除成功
          headers: { }
      security: [ ]

components:
  schemas:
    OperatorInstance:
      type: object
      properties:
        id:
          type: string
        overrides:
          type: object
          properties: { }
          additionalProperties:
            type: object
            properties: { }
      required:
        - id
        - overrides
    CleaningProcess:
      type: object
      properties:
        process:
          type: number
          format: float
          description: 进度百分比
        totalFileNum:
          type: integer
          description: 总文件数量
        finishedFileNum:
          type: integer
          description: 已完成文件数量
      required:
        - process
        - totalFileNum
        - finishedFileNum
    OperatorResponse:
      type: object
      properties:
        id:
          type: string
          description: 算子ID
        name:
          type: string
          description: 算子名称
        description:
          type: string
          description: 算子描述
        version:
          type: string
          description: 算子版本
        inputs:
          type: string
          description: 输入类型
        outputs:
          type: string
          description: 输入类型
        runtime:
          type: string
          description: 运行时设置
        settings:
          type: string
          description: 算子参数
        isStar:
          type: boolean
          description: 是否收藏
        createdAt:
          type: string
          format: date-time
          description: 创建时间
        updatedAt:
          type: string
          format: date-time
          description: 更新时间
      required:
        - inputs
        - outputs
        - runtime
        - settings
        - isStar
    UpdateCleaningTemplateRequest:
      type: object
      required:
        - name
        - instance
        - id
      properties:
        id:
          type: string
        name:
          type: string
          description: 模板名称
        description:
          type: string
          description: 模板描述
        instance:
          type: array
          items: &ref_3
            $ref: '#/components/schemas/OperatorInstance'
          description: 模板定义的清洗规则和配置
    CreateCleaningTemplateRequest:
      type: object
      required:
        - name
        - instance
      properties:
        name:
          type: string
          description: 模板名称
        description:
          type: string
          description: 模板描述
        instance:
          type: array
          items: *ref_3
          description: 任务的具体配置（如果非模板创建，则直接定义）'
    CleaningTemplate:
      type: object
      required:
        - id
        - name
        - instance
        - createdAt
      properties:
        id:
          type: string
          description: 模板唯一标识符
        name:
          type: string
          description: 模板名称
        description:
          type: string
          description: 模板描述
        instance:
          type: array
          items: &ref_4
            $ref: '#/components/schemas/OperatorResponse'
          description: 模板定义的清洗规则和配置
        createdAt:
          type: string
          format: date-time
          description: 模板创建时间
        updatedAt:
          type: string
          format: date-time
          description: 模板最后更新时间
    CreateCleaningTaskRequest:
      type: object
      required:
        - name
        - instance
        - srcDatasetId
        - srcDatasetName
        - destDatasetName
        - destDatasetType
      properties:
        name:
          type: string
          description: 任务名称
        description:
          type: string
          description: 任务描述
        srcDatasetId:
          type: string
        srcDatasetName:
          type: string
        destDatasetName:
          type: string
        destDatasetType:
          type: string
        instance:
          type: array
          items: *ref_3
          description: 任务的具体配置（如果非模板创建，则直接定义）
    ErrorResponse:
      type: object
      properties:
        error:
          type: string
          description: 错误类型
        message:
          type: string
          description: 错误详细信息
    CleaningTask:
      type: object
      required:
        - id
        - name
        - status
        - createdAt
        - startedAt
      properties:
        id:
          type: string
          description: 任务唯一标识符
        name:
          type: string
          description: 任务名称
        description:
          type: string
          description: 任务描述
        srcDatasetId:
          type: string
          description: 源数据集id
        srcDatasetName:
          type: string
          description: 源数据集名称
        destDatasetId:
          type: string
          description: 目标数据集id
        destDatasetName:
          type: string
          description: 目标数据集名称
        status:
          type: string
          description: 任务当前状态
          enum:
            - pending
            - running
            - completed
            - failed
        templateId:
          type: string
          description: 关联的模板ID（如果基于模板创建）
        instance:
          type: array
          items: *ref_4
          description: 任务的具体配置（如果非模板创建，则直接定义）
        progress:
          $ref: '#/components/schemas/CleaningProcess'
        createdAt:
          type: string
          description: 任务创建时间
          format: date-time
        startedAt:
          type: string
          format: date-time
          description: 任务开始时间
        finishedAt:
          type: string
          format: date-time
          description: 任务最后更新时间
    securitySchemes: { }