Dify_AI.yml

app:
  description: ''
  icon: 🤖
  icon_background: '#FFEAD5'
  mode: workflow
  name: 北京市水体抗生素污染空间分布及生态风险评估： 人工智能增效关键数据提取与分析
  use_icon_as_answer_icon: false
dependencies:
- current_identifier: null
  type: marketplace
  value:
    marketplace_plugin_unique_identifier: langgenius/openai:0.1.0@539dc0dde8ecaa84e61b8eeb154c3bb7fbc9caa80afd83763dd5860a725d2742
- current_identifier: null
  type: marketplace
  value:
    marketplace_plugin_unique_identifier: langgenius/deepseek:0.0.6@dd589dc093c8084925858034ab5ec1fdf0d33819f43226c2f8c4a749a9acbbb2
kind: app
version: 0.3.0
workflow:
  conversation_variables: []
  environment_variables: []
  features:
    file_upload:
      allowed_file_extensions:
      - .JPG
      - .JPEG
      - .PNG
      - .GIF
      - .WEBP
      - .SVG
      allowed_file_types:
      - image
      allowed_file_upload_methods:
      - local_file
      - remote_url
      enabled: false
      fileUploadConfig:
        audio_file_size_limit: 50
        batch_count_limit: 5
        file_size_limit: 15
        image_file_size_limit: 10
        video_file_size_limit: 100
        workflow_file_upload_limit: 10
      image:
        enabled: false
        number_limits: 3
        transfer_methods:
        - local_file
        - remote_url
      number_limits: 3
    opening_statement: ''
    retriever_resource:
      enabled: true
    sensitive_word_avoidance:
      enabled: false
    speech_to_text:
      enabled: false
    suggested_questions: []
    suggested_questions_after_answer:
      enabled: false
    text_to_speech:
      enabled: false
      language: ''
      voice: ''
  graph:
    edges:
    - data:
        isInIteration: false
        isInLoop: false
        sourceType: start
        targetType: iteration
      id: 1750466107735-source-1750466144404-target
      selected: false
      source: '1750466107735'
      sourceHandle: source
      target: '1750466144404'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1750466144404'
        sourceType: document-extractor
        targetType: llm
      id: 1750466165834--1750466363404-target
      selected: false
      source: '1750466165834'
      sourceHandle: source
      target: '1750466363404'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1750466144404'
        sourceType: llm
        targetType: code
      id: 1750466363404-source-1750466433066-target
      selected: false
      source: '1750466363404'
      sourceHandle: source
      target: '1750466433066'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: false
        isInLoop: false
        sourceType: iteration
        targetType: template-transform
      id: 1750466144404-source-1750466469749-target
      selected: false
      source: '1750466144404'
      sourceHandle: source
      target: '1750466469749'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: false
        isInLoop: false
        sourceType: template-transform
        targetType: end
      id: 1750466469749-source-1750466481626-target
      selected: false
      source: '1750466469749'
      sourceHandle: source
      target: '1750466481626'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1750466144404'
        sourceType: document-extractor
        targetType: llm
      id: 1750466165834-source-1751007101693-target
      selected: false
      source: '1750466165834'
      sourceHandle: source
      target: '1751007101693'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1750466144404'
        sourceType: llm
        targetType: code
      id: 1751007101693-source-1750466433066-target
      selected: false
      source: '1751007101693'
      sourceHandle: source
      target: '1750466433066'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1750466144404'
        sourceType: iteration-start
        targetType: document-extractor
      id: 1750466144404start-source-1750466165834-target
      selected: false
      source: 1750466144404start
      sourceHandle: source
      target: '1750466165834'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: false
        isInLoop: false
        sourceType: start
        targetType: iteration
      id: 1750466107735-source-1751117358810-target
      selected: false
      source: '1750466107735'
      sourceHandle: source
      target: '1751117358810'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751117358810'
        sourceType: iteration-start
        targetType: document-extractor
      id: 1751117358810start-source-1751117642024-target
      selected: false
      source: 1751117358810start
      sourceHandle: source
      target: '1751117642024'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInLoop: false
        sourceType: iteration
        targetType: template-transform
      id: 1751117358810-source-1750466469749-target
      selected: false
      source: '1751117358810'
      sourceHandle: source
      target: '1750466469749'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751117358810'
        sourceType: llm
        targetType: code
      id: 1751117652660-source-1751121668795-target
      selected: false
      source: '1751117652660'
      sourceHandle: source
      target: '1751121668795'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751117358810'
        sourceType: llm
        targetType: code
      id: 1751117823361-source-1751121668795-target
      selected: false
      source: '1751117823361'
      sourceHandle: source
      target: '1751121668795'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751117358810'
        sourceType: document-extractor
        targetType: llm
      id: 1751117642024-source-1751117652660-target
      selected: false
      source: '1751117642024'
      sourceHandle: source
      target: '1751117652660'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751117358810'
        sourceType: document-extractor
        targetType: llm
      id: 1751117642024-source-1751117823361-target
      selected: false
      source: '1751117642024'
      sourceHandle: source
      target: '1751117823361'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: false
        isInLoop: false
        sourceType: start
        targetType: iteration
      id: 1750466107735-source-1751373621053-target
      selected: false
      source: '1750466107735'
      sourceHandle: source
      target: '1751373621053'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751373621053'
        sourceType: iteration-start
        targetType: document-extractor
      id: 1751373621053start-source-1751373640405-target
      selected: false
      source: 1751373621053start
      sourceHandle: source
      target: '1751373640405'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751373621053'
        sourceType: document-extractor
        targetType: llm
      id: 1751373640405-source-1751373650354-target
      selected: false
      source: '1751373640405'
      sourceHandle: source
      target: '1751373650354'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751373621053'
        sourceType: document-extractor
        targetType: llm
      id: 1751373640405-source-1751373851802-target
      selected: false
      source: '1751373640405'
      sourceHandle: source
      target: '1751373851802'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751373621053'
        sourceType: llm
        targetType: code
      id: 1751373650354-source-1751374441285-target
      selected: false
      source: '1751373650354'
      sourceHandle: source
      target: '1751374441285'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751373621053'
        sourceType: llm
        targetType: code
      id: 1751373851802-source-1751374441285-target
      selected: false
      source: '1751373851802'
      sourceHandle: source
      target: '1751374441285'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInLoop: false
        sourceType: iteration
        targetType: template-transform
      id: 1751373621053-source-1750466469749-target
      selected: false
      source: '1751373621053'
      sourceHandle: source
      target: '1750466469749'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: false
        isInLoop: false
        sourceType: start
        targetType: iteration
      id: 1750466107735-source-1751376385930-target
      selected: false
      source: '1750466107735'
      sourceHandle: source
      target: '1751376385930'
      targetHandle: target
      type: custom
      zIndex: 0
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751376385930'
        sourceType: iteration-start
        targetType: document-extractor
      id: 1751376385930start-source-1751378023089-target
      selected: false
      source: 1751376385930start
      sourceHandle: source
      target: '1751378023089'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751376385930'
        sourceType: document-extractor
        targetType: llm
      id: 1751378023089-source-1751378035165-target
      selected: false
      source: '1751378023089'
      sourceHandle: source
      target: '1751378035165'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751376385930'
        sourceType: document-extractor
        targetType: llm
      id: 1751378023089-source-1751378038707-target
      selected: false
      source: '1751378023089'
      sourceHandle: source
      target: '1751378038707'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751376385930'
        sourceType: llm
        targetType: code
      id: 1751378035165-source-1751378358745-target
      selected: false
      source: '1751378035165'
      sourceHandle: source
      target: '1751378358745'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInIteration: true
        isInLoop: false
        iteration_id: '1751376385930'
        sourceType: llm
        targetType: code
      id: 1751378038707-source-1751378358745-target
      selected: false
      source: '1751378038707'
      sourceHandle: source
      target: '1751378358745'
      targetHandle: target
      type: custom
      zIndex: 1002
    - data:
        isInLoop: false
        sourceType: iteration
        targetType: template-transform
      id: 1751376385930-source-1750466469749-target
      selected: false
      source: '1751376385930'
      sourceHandle: source
      target: '1750466469749'
      targetHandle: target
      type: custom
      zIndex: 0
    nodes:
    - data:
        desc: 请提交要读取抗生素浓度信息的文件
        selected: false
        title: 开始
        type: start
        variables:
        - allowed_file_extensions: []
          allowed_file_types:
          - document
          - image
          allowed_file_upload_methods:
          - local_file
          label: object_file
          max_length: 10
          options: []
          required: true
          type: file-list
          variable: object_file
      height: 117
      id: '1750466107735'
      position:
        x: 30
        y: 427
      positionAbsolute:
        x: 30
        y: 427
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
    - data:
        desc: ''
        error_handle_mode: terminated
        height: 454
        is_parallel: true
        iterator_selector:
        - '1750466107735'
        - object_file
        output_selector:
        - '1750466433066'
        - deepseek_chat
        output_type: array[object]
        parallel_nums: 10
        selected: false
        start_node_id: 1750466144404start
        title: 迭代
        type: iteration
        width: 1116
      height: 454
      id: '1750466144404'
      position:
        x: 334
        y: 417.6196808987317
      positionAbsolute:
        x: 334
        y: 417.6196808987317
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 1116
      zIndex: 1
    - data:
        desc: ''
        isInIteration: true
        selected: false
        title: ''
        type: iteration-start
      draggable: false
      height: 48
      id: 1750466144404start
      parentId: '1750466144404'
      position:
        x: 60
        y: 109.5
      positionAbsolute:
        x: 394
        y: 527.1196808987318
      selectable: false
      sourcePosition: right
      targetPosition: left
      type: custom-iteration-start
      width: 44
      zIndex: 1002
    - data:
        desc: ''
        isInIteration: true
        isInLoop: false
        is_array_file: true
        iteration_id: '1750466144404'
        selected: false
        title: 文档提取
        type: document-extractor
        variable_selector:
        - '1750466107735'
        - object_file
      height: 91
      id: '1750466165834'
      parentId: '1750466144404'
      position:
        x: 204
        y: 88
      positionAbsolute:
        x: 538
        y: 505.6196808987317
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        context:
          enabled: true
          variable_selector:
          - '1750466165834'
          - text
        desc: 话术1
        isInIteration: true
        isInLoop: false
        iteration_id: '1750466144404'
        model:
          completion_params: {}
          mode: chat
          name: deepseek-chat
          provider: langgenius/deepseek/deepseek
        prompt_template:
        - id: bf9b5d70-f601-4f4d-9dac-f07ed861ea55
          role: system
          text: '阶段1：抗生素与水体类型识别：

            “请从文献{{#1750466165834.text#}}

            中提取以下信息：

            PaperID是{{#1750466144404.index#}}

            ①Name:抗生素名称（如‘氧氟沙星’，需区分异构体，如‘R-氧氟沙星’）；②对应的水体类型Water（如‘河水’‘海水’‘雨水’），若未明确标注‘水体’，需根据上下文推断（如‘湖泊沉积物’不属于水体）。”

            阶段2：浓度与采样信息整合：

            “针对上一步识别的抗生素-水体对，提取：①浓度数值Value（保留科学计数法，如‘3.2×10⁻³ μg/L’，注意标注“浓度范围”“平均值”“中位数”“未检出”等情况）；②采样地点Location（精确到流域/城市，如‘长江三角洲’）；③采样时间Date（精确到年月，如‘2023年5月’）；④若浓度低于检测限（LOD），标注‘<LOD’及数值（如‘<0.5
            ng/L’）。”

            阶段3：时空数据结构化：

            “将前两步结果生成环境监测数据记录，以JSON数组格式输出要求：①包含‘文献PaperID{{#1750466144404.index#}}-抗生素Name-浓度Value-水体Water-地点Location-时间Date’五要素；②地点需关联流域（如‘太湖梅梁湾’属于‘太湖流域’）；③时间信息缺失时标注‘[时间未知]’。

            '
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        structured_output:
          schema:
            additionalProperties: false
            properties:
              Context:
                type: string
              Date:
                type: string
              Location:
                type: string
              Name:
                type: string
              PaperID:
                type: string
              Unit:
                type: string
              Value:
                type: number
              Water:
                type: string
            required:
            - PaperID
            - Name
            - Value
            - Unit
            - Water
            - Location
            - Date
            - Context
            type: object
        structured_output_enabled: true
        title: deepseek_chat1
        type: llm
        variables: []
        vision:
          enabled: false
      height: 147
      id: '1750466363404'
      parentId: '1750466144404'
      position:
        x: 508
        y: 60
      positionAbsolute:
        x: 842
        y: 477.6196808987317
      selected: true
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        code: "def main(arg1: object, arg2: object) -> dict:\n    return {\n     \
          \   \"deepseek_chat\": {\n            \"话术一\": arg1,\n            \"话术二\"\
          : arg2\n        }\n    }\n"
        code_language: python3
        desc: ''
        isInIteration: true
        isInLoop: false
        iteration_id: '1750466144404'
        outputs:
          deepseek_chat:
            children: null
            type: object
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        title: 代码执行
        type: code
        variables:
        - value_selector:
          - '1750466363404'
          - structured_output
          variable: arg1
        - value_selector:
          - '1751007101693'
          - structured_output
          variable: arg2
      height: 83
      id: '1750466433066'
      parentId: '1750466144404'
      position:
        x: 812
        y: 92
      positionAbsolute:
        x: 1146
        y: 509.6196808987317
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        desc: ''
        selected: false
        template: "[\r\n  {% for item in arg1 %}\r\n    {\r\n      \"deepseek_chat\"\
          : {{ item }},\r\n      \"source\": \"iteration_1\"\r\n    }{% if not loop.last\
          \ %},{% endif %}\r\n  {% endfor %}\r\n]\r\n[\r\n  {% for item in arg2 %}\r\
          \n    {\r\n      \"deepseek_coder\": {{ item }},\r\n      \"source\": \"\
          iteration_2\"\r\n    }{% if not loop.last %},{% endif %}\r\n  {% endfor\
          \ %}\r\n]\r\n[\r\n  {% for item in arg3 %}\r\n    {\r\n      \"ChatGPT_4o_mini\"\
          : {{ item }},\r\n      \"source\": \"iteration_3\"\r\n    }{% if not loop.last\
          \ %},{% endif %}\r\n  {% endfor %}\r\n]\r\n[\r\n  {% for item in arg4 %}\r\
          \n    {\r\n      \"ChatGPT_41_mini\": {{ item }},\r\n      \"source\": \"\
          iteration_4\"\r\n    }{% if not loop.last %},{% endif %}\r\n  {% endfor\
          \ %}\r\n]"
        title: 模板转换
        type: template-transform
        variables:
        - value_selector:
          - '1750466144404'
          - output
          variable: arg1
        - value_selector:
          - '1751117358810'
          - output
          variable: arg2
        - value_selector:
          - '1751373621053'
          - output
          variable: arg3
        - value_selector:
          - '1751376385930'
          - output
          variable: arg4
      height: 53
      id: '1750466469749'
      position:
        x: 1518.6612375229165
        y: 427
      positionAbsolute:
        x: 1518.6612375229165
        y: 427
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
    - data:
        desc: ''
        outputs:
        - value_selector:
          - '1750466469749'
          - output
          variable: 抗生素浓度
        selected: false
        title: 结束
        type: end
      height: 89
      id: '1750466481626'
      position:
        x: 1805.3387624770835
        y: 427
      positionAbsolute:
        x: 1805.3387624770835
        y: 427
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
    - data:
        context:
          enabled: true
          variable_selector:
          - '1750466165834'
          - text
        desc: 话术2
        isInIteration: true
        isInLoop: false
        iteration_id: '1750466144404'
        model:
          completion_params: {}
          mode: chat
          name: deepseek-chat
          provider: langgenius/deepseek/deepseek
        prompt_template:
        - id: e6e46307-9337-4ad8-8ec8-45474802bc6a
          role: system
          text: '请从我提供的科学文献{{#1750466165834.text#}}

            中提取所有抗生素及其浓度相关信息，以JSON数组格式输出。要求包含：

            PaperID是{{#1750466144404.index#}}

            ① Name:抗生素名称（如‘磺胺甲恶唑’，保留完整化学名）；

            ② Value and Unit:浓度数值及单位（如‘12.5 ng/L’，环境领域常见单位：ng/L、μg/L、μg/kg（沉积物），注意标注“浓度范围”“平/均值”“中位数”“未检出”等情况）

            ③ Water: 水体类型（如‘地表水’‘饮用水’‘污水处理厂出水’）

            ④ Location & Date: 采样信息（采样地点/河流名称、采样时间，如‘长江南京段，2023年7月’）

            ⑤ Context: 原文上下文（段落编号或句子片段，如‘摘要第3段’‘图 2 caption’）。'
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        structured_output:
          schema:
            additionalProperties: false
            properties:
              Context:
                type: string
              Date:
                type: string
              Location:
                type: string
              Name:
                type: string
              PaperID:
                type: string
              Unit:
                type: string
              Value:
                type: number
              Water:
                type: string
            required:
            - PaperID
            - Name
            - Value
            - Unit
            - Water
            - Location
            - Date
            - Context
            type: object
        structured_output_enabled: true
        title: deepseek_chat2
        type: llm
        variables: []
        vision:
          enabled: false
      height: 147
      id: '1751007101693'
      parentId: '1750466144404'
      position:
        x: 509.42857142857156
        y: 247
      positionAbsolute:
        x: 843.4285714285716
        y: 664.6196808987318
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        desc: ''
        error_handle_mode: terminated
        height: 454
        is_parallel: true
        iterator_selector:
        - '1750466107735'
        - object_file
        output_selector:
        - '1751121668795'
        - deepseek_coder
        output_type: array[object]
        parallel_nums: 10
        selected: false
        start_node_id: 1751117358810start
        title: 迭代 2
        type: iteration
        width: 1116
      height: 454
      id: '1751117358810'
      position:
        x: 334
        y: 921
      positionAbsolute:
        x: 334
        y: 921
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 1116
      zIndex: 1
    - data:
        desc: ''
        isInIteration: true
        selected: false
        title: ''
        type: iteration-start
      draggable: false
      height: 48
      id: 1751117358810start
      parentId: '1751117358810'
      position:
        x: 60
        y: 109.5
      positionAbsolute:
        x: 394
        y: 1030.5
      selectable: false
      sourcePosition: right
      targetPosition: left
      type: custom-iteration-start
      width: 44
      zIndex: 1002
    - data:
        desc: ''
        isInIteration: true
        isInLoop: false
        is_array_file: true
        iteration_id: '1751117358810'
        selected: false
        title: 文档提取2
        type: document-extractor
        variable_selector:
        - '1750466107735'
        - object_file
      height: 91
      id: '1751117642024'
      parentId: '1751117358810'
      position:
        x: 204
        y: 88
      positionAbsolute:
        x: 538
        y: 1009
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        context:
          enabled: true
          variable_selector:
          - '1751117642024'
          - text
        desc: '话术1 '
        isInIteration: true
        isInLoop: false
        iteration_id: '1751117358810'
        model:
          completion_params: {}
          mode: chat
          name: deepseek-coder
          provider: langgenius/deepseek/deepseek
        prompt_template:
        - id: 963865c5-ce94-42a7-b118-2f82b32563ba
          role: system
          text: '请从我提供的科学文献{{#1751117642024.text#}}

            中提取所有抗生素及其浓度相关信息，以JSON数组格式输出。要求包含：

            PaperID是{{#1751117358810.index#}}

            ① Name:抗生素名称（如‘磺胺甲恶唑’，保留完整化学名）；

            ② Value and Unit:浓度数值及单位（如‘12.5 ng/L’，环境领域常见单位：ng/L、μg/L、μg/kg（沉积物），注意标注“浓度范围”“平/均值”“中位数”“未检出”等情况）

            ③ Water: 水体类型（如‘地表水’‘饮用水’‘污水处理厂出水’）

            ④ Location & Date: 采样信息（采样地点/河流名称、采样时间，如‘长江南京段，2023年7月’）

            ⑤ Context: 原文上下文（段落编号或句子片段，如‘摘要第3段’‘图 2 caption’）。'
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        structured_output:
          schema:
            additionalProperties: false
            properties:
              Context:
                type: string
              Date:
                type: string
              Location:
                type: string
              Name:
                type: string
              PaperID:
                type: string
              Unit:
                type: string
              Value:
                type: number
              Water:
                type: string
            required:
            - PaperID
            - Name
            - Value
            - Unit
            - Water
            - Location
            - Date
            - Context
            type: object
        structured_output_enabled: true
        title: deepseek_coder1
        type: llm
        variables: []
        vision:
          enabled: false
      height: 147
      id: '1751117652660'
      parentId: '1751117358810'
      position:
        x: 508
        y: 60
      positionAbsolute:
        x: 842
        y: 981
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        context:
          enabled: true
          variable_selector:
          - '1751117642024'
          - text
        desc: 话术2
        isInIteration: true
        isInLoop: false
        iteration_id: '1751117358810'
        model:
          completion_params: {}
          mode: chat
          name: deepseek-coder
          provider: langgenius/deepseek/deepseek
        prompt_template:
        - id: 2587d97a-5f7e-4990-8c6c-3a6dbdb11e0e
          role: system
          text: '阶段1：抗生素与水体类型识别：

            请从文献{{#1751117642024.text#}}

            中提取以下信息：

            PaperID是{{#1751117358810.index#}}

            ①Name:抗生素名称（如‘氧氟沙星’，需区分异构体，如‘R-氧氟沙星’）；②对应的水体类型Water（如‘河水’‘海水’‘雨水’），若未明确标注‘水体’，需根据上下文推断（如‘湖泊沉积物’不属于水体）。”

            阶段2：浓度与采样信息整合：

            “针对上一步识别的抗生素-水体对，提取：①浓度数值Value（保留科学计数法，如‘3.2×10⁻³ μg/L’，注意标注“浓度范围”“平均值”“中位数”“未检出”等情况）；②采样地点Location（精确到流域/城市，如‘长江三角洲’）；③采样时间Date（精确到年月，如‘2023年5月’）；④若浓度低于检测限（LOD），标注‘<LOD’及数值（如‘<0.5
            ng/L’）。”

            阶段3：时空数据结构化：

            “将前两步结果生成环境监测数据记录，以JSON数组格式输出要求：①包含‘文献PaperID{{#1751117358810.index#}}-抗生素Name-浓度Value-水体Water-地点Location-时间Date’五要素；②地点需关联流域（如‘太湖梅梁湾’属于‘太湖流域’）；③时间信息缺失时标注‘[时间未知]’。'
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        structured_output:
          schema:
            additionalProperties: false
            properties:
              Context:
                type: string
              Date:
                type: string
              Location:
                type: string
              Name:
                type: string
              PaperID:
                type: string
              Unit:
                type: string
              Value:
                type: number
              Water:
                type: string
            required:
            - PaperID
            - Name
            - Value
            - Unit
            - Water
            - Location
            - Date
            - Context
            type: object
        structured_output_enabled: true
        title: deepseek_coder2
        type: llm
        variables: []
        vision:
          enabled: false
      height: 147
      id: '1751117823361'
      parentId: '1751117358810'
      position:
        x: 508
        y: 247
      positionAbsolute:
        x: 842
        y: 1168
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        code: "def main(arg1: object, arg2: object) -> dict:\n    return {\n     \
          \   \"deepseek_coder\": {\n            \"话术一\": arg1,\n            \"话术二\"\
          : arg2\n        }\n    }\n"
        code_language: python3
        desc: ''
        isInIteration: true
        isInLoop: false
        iteration_id: '1751117358810'
        outputs:
          deepseek_coder:
            children: null
            type: object
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        title: 代码执行 2
        type: code
        variables:
        - value_selector:
          - '1751117652660'
          - structured_output
          variable: arg1
        - value_selector:
          - '1751117823361'
          - structured_output
          variable: arg2
      height: 83
      id: '1751121668795'
      parentId: '1751117358810'
      position:
        x: 812
        y: 92
      positionAbsolute:
        x: 1146
        y: 1013
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        desc: ''
        error_handle_mode: terminated
        height: 454
        is_parallel: true
        iterator_selector:
        - '1750466107735'
        - object_file
        output_selector:
        - '1751374441285'
        - ChatGPT_4o_mini
        output_type: array[object]
        parallel_nums: 10
        selected: false
        start_node_id: 1751373621053start
        title: 迭代 3
        type: iteration
        width: 1116
      height: 454
      id: '1751373621053'
      position:
        x: 334
        y: 1415
      positionAbsolute:
        x: 334
        y: 1415
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 1116
      zIndex: 1
    - data:
        desc: ''
        isInIteration: true
        selected: false
        title: ''
        type: iteration-start
      draggable: false
      height: 48
      id: 1751373621053start
      parentId: '1751373621053'
      position:
        x: 60
        y: 109.5
      positionAbsolute:
        x: 394
        y: 1524.5
      selectable: false
      sourcePosition: right
      targetPosition: left
      type: custom-iteration-start
      width: 44
      zIndex: 1002
    - data:
        desc: ''
        isInIteration: true
        isInLoop: false
        is_array_file: true
        iteration_id: '1751373621053'
        selected: false
        title: 文档提取3
        type: document-extractor
        variable_selector:
        - '1750466107735'
        - object_file
      height: 91
      id: '1751373640405'
      parentId: '1751373621053'
      position:
        x: 204
        y: 88
      positionAbsolute:
        x: 538
        y: 1503
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        context:
          enabled: true
          variable_selector:
          - '1751373640405'
          - text
        desc: 话术1
        isInIteration: true
        isInLoop: false
        iteration_id: '1751373621053'
        model:
          completion_params: {}
          mode: chat
          name: gpt-4o-mini
          provider: langgenius/openai/openai
        prompt_template:
        - id: 2f081e41-44a8-4f18-a046-326e248766f4
          role: system
          text: '请从我提供的科学文献{{#1751373640405.text#}}

            中提取所有抗生素及其浓度相关信息，以JSON数组格式输出。要求包含：

            PaperID是{{#1751373621053.index#}}

            ① Name:抗生素名称（如‘磺胺甲恶唑’，保留完整化学名）；

            ② Value and Unit:浓度数值及单位（如‘12.5 ng/L’，环境领域常见单位：ng/L、μg/L、μg/kg（沉积物），注意标注“浓度范围”“平/均值”“中位数”“未检出”等情况）

            ③ Water: 水体类型（如‘地表水’‘饮用水’‘污水处理厂出水’）

            ④ Location & Date: 采样信息（采样地点/河流名称、采样时间，如‘长江南京段，2023年7月’）

            ⑤ Context: 原文上下文（段落编号或句子片段，如‘摘要第3段’‘图 2 caption’）。'
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        structured_output:
          schema:
            additionalProperties: false
            properties:
              Context:
                type: string
              Date:
                type: string
              Location:
                type: string
              Name:
                type: string
              PaperID:
                type: string
              Unit:
                type: string
              Value:
                type: number
              Water:
                type: string
            required:
            - PaperID
            - Name
            - Value
            - Unit
            - Water
            - Location
            - Date
            - Context
            type: object
        structured_output_enabled: true
        title: ChatGPT-4o-mini1
        type: llm
        variables: []
        vision:
          enabled: false
      height: 147
      id: '1751373650354'
      parentId: '1751373621053'
      position:
        x: 508
        y: 60
      positionAbsolute:
        x: 842
        y: 1475
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        context:
          enabled: true
          variable_selector:
          - '1751373640405'
          - text
        desc: 话术2
        isInIteration: true
        isInLoop: false
        iteration_id: '1751373621053'
        model:
          completion_params: {}
          mode: chat
          name: gpt-4o-mini
          provider: langgenius/openai/openai
        prompt_template:
        - id: c06f5b15-b700-4c40-9447-5ea53d03afdb
          role: system
          text: 'PaperID是{{#1751373621053.index#}}

            阶段1：抗生素与水体类型识别：

            请从文献{{#1751373640405.text#}}

            中提取以下信息：①Name:抗生素名称（如‘氧氟沙星’，需区分异构体，如‘R-氧氟沙星’）；②对应的水体类型Water（如‘河水’‘海水’‘雨水’），若未明确标注‘水体’，需根据上下文推断（如‘湖泊沉积物’不属于水体）。”

            阶段2：浓度与采样信息整合：

            “针对上一步识别的抗生素-水体对，提取：①浓度数值Value（保留科学计数法，如‘3.2×10⁻³ μg/L’，注意标注“浓度范围”“平均值”“中位数”“未检出”等情况）；②采样地点Location（精确到流域/城市，如‘长江三角洲’）；③采样时间Date（精确到年月，如‘2023年5月’）；④若浓度低于检测限（LOD），标注‘<LOD’及数值（如‘<0.5
            ng/L’）。”

            阶段3：时空数据结构化：

            “将前两步结果生成环境监测数据记录，以JSON数组格式输出要求：①包含‘抗生素Name-浓度Value-水体Water-地点Location-时间Date’五要素；②地点需关联流域（如‘太湖梅梁湾’属于‘太湖流域’）；③时间信息缺失时标注‘[时间未知]’。'
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        structured_output:
          schema:
            additionalProperties: false
            properties:
              Context:
                type: string
              Date:
                type: string
              Location:
                type: string
              Name:
                type: string
              PaperID:
                type: string
              Unit:
                type: string
              Value:
                type: number
              Water:
                type: string
            required:
            - PaperID
            - Name
            - Value
            - Unit
            - Water
            - Location
            - Date
            - Context
            type: object
        structured_output_enabled: true
        title: ChatGPT-4o-mini2
        type: llm
        variables: []
        vision:
          enabled: false
      height: 147
      id: '1751373851802'
      parentId: '1751373621053'
      position:
        x: 508
        y: 247
      positionAbsolute:
        x: 842
        y: 1662
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        code: "def main(arg1: object, arg2: object) -> dict:\n    return {\n     \
          \   \"ChatGPT_4o_mini\": {\n            \"话术一\": arg1,\n            \"话术二\"\
          : arg2\n        }\n    }\n\n"
        code_language: python3
        desc: ''
        isInIteration: true
        isInLoop: false
        iteration_id: '1751373621053'
        outputs:
          ChatGPT_4o_mini:
            children: null
            type: object
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        title: 代码执行 3
        type: code
        variables:
        - value_selector:
          - '1751373650354'
          - structured_output
          variable: arg1
        - value_selector:
          - '1751373851802'
          - structured_output
          variable: arg2
      height: 83
      id: '1751374441285'
      parentId: '1751373621053'
      position:
        x: 812
        y: 92
      positionAbsolute:
        x: 1146
        y: 1507
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        desc: ''
        error_handle_mode: terminated
        height: 454
        is_parallel: true
        iterator_selector:
        - '1750466107735'
        - object_file
        output_selector:
        - '1751378358745'
        - ChatGPT_41_mini
        output_type: array[object]
        parallel_nums: 10
        selected: false
        start_node_id: 1751376385930start
        title: 迭代 4
        type: iteration
        width: 1116
      height: 454
      id: '1751376385930'
      position:
        x: 334
        y: 1909
      positionAbsolute:
        x: 334
        y: 1909
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 1116
      zIndex: 1
    - data:
        desc: ''
        isInIteration: true
        selected: false
        title: ''
        type: iteration-start
      draggable: false
      height: 48
      id: 1751376385930start
      parentId: '1751376385930'
      position:
        x: 60
        y: 109.5
      positionAbsolute:
        x: 394
        y: 2018.5
      selectable: false
      sourcePosition: right
      targetPosition: left
      type: custom-iteration-start
      width: 44
      zIndex: 1002
    - data:
        desc: ''
        isInIteration: true
        isInLoop: false
        is_array_file: true
        iteration_id: '1751376385930'
        selected: false
        title: 文档提取4
        type: document-extractor
        variable_selector:
        - '1750466107735'
        - object_file
      height: 91
      id: '1751378023089'
      parentId: '1751376385930'
      position:
        x: 204
        y: 88
      positionAbsolute:
        x: 538
        y: 1997
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        context:
          enabled: true
          variable_selector:
          - '1751378023089'
          - text
        desc: 话术1
        isInIteration: true
        isInLoop: false
        iteration_id: '1751376385930'
        model:
          completion_params: {}
          mode: chat
          name: gpt-4.1-mini
          provider: langgenius/openai/openai
        prompt_template:
        - id: cd69cc4d-8c8b-4481-8421-e21ba1f3686d
          role: system
          text: '请从我提供的科学文献pdf{{#1751378023089.text#}}

            中提取所有抗生素及其浓度相关信息，以JSON数组格式输出。要求包含：

            PaperID是{{#1751376385930.index#}}

            ① Name:抗生素名称（如‘磺胺甲恶唑’，保留完整化学名）；

            ② Value and Unit:浓度数值及单位（如‘12.5 ng/L’，环境领域常见单位：ng/L、μg/L、μg/kg（沉积物），注意标注“浓度范围”“平均值”“中位数”“未检出”等情况）

            ③ Water: 水体类型（如‘地表水’‘饮用水’‘污水处理厂出水’）

            ④ Location & Date: 采样信息（采样地点/河流名称、采样时间，如‘长江南京段，2023年7月’）

            ⑤ Context: 原文上下文（段落编号或句子片段，如‘摘要第3段’‘图 2 caption’）。'
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        structured_output:
          schema:
            additionalProperties: false
            properties:
              Context:
                type: string
              Date:
                type: string
              Location:
                type: string
              Name:
                type: string
              PaperID:
                type: string
              Unit:
                type: string
              Value:
                type: number
              Water:
                type: string
            required:
            - PaperID
            - Name
            - Value
            - Unit
            - Water
            - Location
            - Date
            - Context
            type: object
        structured_output_enabled: true
        title: ChatGPT-4.1-mini1
        type: llm
        variables: []
        vision:
          enabled: false
      height: 147
      id: '1751378035165'
      parentId: '1751376385930'
      position:
        x: 508
        y: 60
      positionAbsolute:
        x: 842
        y: 1969
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        context:
          enabled: true
          variable_selector:
          - '1751378023089'
          - text
        desc: 话术2
        isInIteration: true
        isInLoop: false
        iteration_id: '1751376385930'
        model:
          completion_params: {}
          mode: chat
          name: gpt-4.1-mini
          provider: langgenius/openai/openai
        prompt_template:
        - id: dd367ad0-f361-4ea0-921d-57d6720f31cd
          role: system
          text: 'PaperID是{{#1751376385930.index#}}

            阶段1：抗生素与水体类型识别：

            请从文献{{#1751378023089.text#}}

            中提取以下信息：①Name:抗生素名称（如‘氧氟沙星’，需区分异构体，如‘R-氧氟沙星’）；②对应的水体类型Water（如‘河水’‘海水’‘雨水’），若未明确标注‘水体’，需根据上下文推断（如‘湖泊沉积物’不属于水体）。”

            阶段2：浓度与采样信息整合：

            “针对上一步识别的抗生素-水体对，提取：①浓度数值Value（保留科学计数法，如‘3.2×10⁻³ μg/L’，注意标注“浓度范围”“平均值”“中位数”“未检出”等情况）；②采样地点Location（精确到流域/城市，如‘长江三角洲’）；③采样时间Date（精确到年月，如‘2023年5月’）；④若浓度低于检测限（LOD），标注‘<LOD’及数值（如‘<0.5
            ng/L’）。”

            阶段3：时空数据结构化：

            “将前两步结果生成环境监测数据记录，以JSON数组格式输出要求：①包含‘抗生素Name-浓度Value-水体Water-地点Location-时间Date’五要素；②地点需关联流域（如‘太湖梅梁湾’属于‘太湖流域’）；③时间信息缺失时标注‘[时间未知]’。'
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        structured_output:
          schema:
            additionalProperties: false
            properties:
              Context:
                type: string
              Date:
                type: string
              Location:
                type: string
              Name:
                type: string
              PaperID:
                type: string
              Unit:
                type: string
              Value:
                type: number
              Water:
                type: string
            required:
            - PaperID
            - Name
            - Value
            - Unit
            - Water
            - Location
            - Date
            - Context
            type: object
        structured_output_enabled: true
        title: ChatGPT-4.1-mini2
        type: llm
        variables: []
        vision:
          enabled: false
      height: 147
      id: '1751378038707'
      parentId: '1751376385930'
      position:
        x: 508
        y: 247
      positionAbsolute:
        x: 842
        y: 2156
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        code: "def main(arg1: object, arg2: object) -> dict:\n    return {\n     \
          \   \"ChatGPT_41_mini\": {\n            \"话术一\": arg1,\n            \"话术二\"\
          : arg2\n        }\n    }"
        code_language: python3
        desc: ''
        isInIteration: true
        isInLoop: false
        iteration_id: '1751376385930'
        outputs:
          ChatGPT_41_mini:
            children: null
            type: object
        retry_config:
          max_retries: 3
          retry_enabled: true
          retry_interval: 1000
        selected: false
        title: 代码执行 4
        type: code
        variables:
        - value_selector:
          - '1751378035165'
          - structured_output
          variable: arg1
        - value_selector:
          - '1751378038707'
          - structured_output
          variable: arg2
      height: 83
      id: '1751378358745'
      parentId: '1751376385930'
      position:
        x: 812
        y: 92
      positionAbsolute:
        x: 1146
        y: 2001
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom
      width: 244
      zIndex: 1002
    - data:
        author: user
        desc: ''
        height: 260
        selected: false
        showAuthor: true
        text: '{"root":{"children":[{"children":[],"direction":null,"format":"","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          16px;","textFormat":0},{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"本工作流程围绕科学文献中抗生素浓度数据获取展开，通过调用多种大语言模型，以多样化话术对文献进行询问，经文档抽取、代码执行、模板转换等环节，最终实现数据提取。","type":"text","version":1}],"direction":"ltr","format":"","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          16px;","textFormat":0}],"direction":"ltr","format":"","indent":0,"type":"root","version":1,"textStyle":"font-size:
          16px;"}}'
        theme: pink
        title: ''
        type: ''
        width: 249
      height: 260
      id: '1753182127826'
      position:
        x: -236.30923651380604
        y: 427
      positionAbsolute:
        x: -236.30923651380604
        y: 427
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom-note
      width: 249
    - data:
        author: user
        desc: ''
        height: 129
        selected: false
        showAuthor: true
        text: '{"root":{"children":[{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"工作流程将由此开始。系统会提示用户上传含有抗生素浓度的科学文献。","type":"text","version":1}],"direction":"ltr","format":"","indent":0,"type":"paragraph","version":1,"textFormat":0,"textStyle":""}],"direction":"ltr","format":"","indent":0,"type":"root","version":1,"textStyle":"font-size:
          16px;"}}'
        theme: blue
        title: ''
        type: ''
        width: 240
      height: 129
      id: '1753182276524'
      position:
        x: 30
        y: 558.1123945553131
      positionAbsolute:
        x: 30
        y: 558.1123945553131
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom-note
      width: 240
    - data:
        author: user
        desc: ''
        height: 335
        selected: false
        showAuthor: true
        text: '{"root":{"children":[{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"本流程以包含抗生素浓度信息挖掘需求的文件（","type":"text","version":1},{"detail":0,"format":16,"mode":"normal","style":"font-size:
          16px;","text":"object_file","type":"text","version":1},{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"）作为输入，通过双并行迭代模式提升数据处理效率。具体而言：","type":"text","version":1}],"direction":"ltr","format":"","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          16px;","textFormat":0},{"children":[],"direction":null,"format":"","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          16px;","textFormat":0},{"children":[{"children":[{"detail":0,"format":1,"mode":"normal","style":"font-size:
          16px;","text":"并行执行","type":"text","version":1},{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"：同步触发四个独立迭代流程（迭代1、迭代2、迭代3、迭代4）分别采用不同的大语言模型（deepseek_chat,deepseek-coder,ChatGPT-4o-mini,ChatGPT-4.1-mini）进行数据提取；","type":"text","version":1}],"direction":"ltr","format":"","indent":0,"type":"link","version":1,"textStyle":"font-size:
          16px;","rel":"noreferrer","target":null,"title":null,"url":""}],"direction":"ltr","format":"start","indent":0,"type":"listitem","version":1,"textFormat":1,"textStyle":"font-size:
          16px;","value":1},{"children":[{"detail":0,"format":1,"mode":"normal","style":"font-size:
          16px;","text":"差异化分析","type":"text","version":1},{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"：通过模型多样性覆盖不同文献分析场景（如结构化数据解析、非结构化文本挖掘等）；","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"listitem","version":1,"textFormat":1,"textStyle":"font-size:
          16px;","value":2},{"children":[{"detail":0,"format":1,"mode":"normal","style":"font-size:
          16px;","text":"效率优化","type":"text","version":1},{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"：并行架构显著提升数据提取吞吐量，确保快速完成大规模文献处理","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"listitem","version":1,"textFormat":1,"textStyle":"font-size:
          16px;","value":3}],"direction":"ltr","format":"","indent":0,"type":"list","version":1,"textFormat":1,"textStyle":"font-size:
          16px;","listType":"bullet","start":1,"tag":"ul"}],"direction":"ltr","format":"","indent":0,"type":"root","version":1,"textFormat":1,"textStyle":"font-size:
          16px;"}}'
        theme: blue
        title: ''
        type: ''
        width: 505
      height: 335
      id: '1753182324388'
      position:
        x: -236.30923651380604
        y: 706.2131430558675
      positionAbsolute:
        x: -236.30923651380604
        y: 706.2131430558675
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom-note
      width: 505
    - data:
        author: user
        desc: ''
        height: 233
        selected: false
        showAuthor: true
        text: '{"root":{"children":[{"children":[{"detail":0,"format":1,"mode":"normal","style":"font-size:
          16px;","text":"文档提取节点","type":"text","version":1},{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"：","type":"text","version":1}],"direction":"ltr","format":"","indent":0,"type":"paragraph","version":1,"textFormat":1,"textStyle":"font-size:
          16px;"},{"children":[],"direction":"ltr","format":"","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          16px;","textFormat":0},{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"在每轮迭代中，由此对输入文件开展解析，精准定位、抽取与抗生素浓度相关文本内容，为后续模型分析提供基础素材。迭代1、迭代2、迭代3、迭代4分别代表不同的大语言模型，其内部的两种LLM分别代表同一种大语言模型的两种询问话术。","type":"text","version":1}],"direction":"ltr","format":"","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          16px;","textFormat":0}],"direction":"ltr","format":"","indent":0,"type":"root","version":1,"textFormat":1,"textStyle":"font-size:
          16px;"}}'
        theme: blue
        title: ''
        type: ''
        width: 506
      height: 233
      id: '1753182396900'
      position:
        x: -236.30923651380604
        y: 1071.4799907166887
      positionAbsolute:
        x: -236.30923651380604
        y: 1071.4799907166887
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom-note
      width: 506
    - data:
        author: user
        desc: ''
        height: 255
        selected: false
        showAuthor: true
        text: '{"root":{"children":[{"children":[{"detail":0,"format":1,"mode":"normal","style":"font-size:
          16px;","text":"大语言模型：","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"paragraph","version":1,"textFormat":1,"textStyle":"font-size:
          16px;"},{"children":[],"direction":"ltr","format":"start","indent":0,"type":"paragraph","version":1,"textFormat":1,"textStyle":"font-size:
          14px;"},{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"解析后的文本，流转至不同大语言模型：","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          16px;","textFormat":0},{"children":[{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"借助多模型差异化能力，以多样化话术对文献内容深度询问，覆盖 “数据准确性校验”“浓度单位换算”“实验条件关联分析”
          等多维度需求；","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"listitem","version":1,"textStyle":"font-size:
          16px;","value":1},{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"为保障流程稳定性，设置 “失败时重试 3 次” 机制 ，应对模型调用异常、文本理解偏差等问题，提升数据提取成功率。","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"listitem","version":1,"textStyle":"font-size:
          16px;","value":2}],"direction":"ltr","format":"","indent":0,"type":"list","version":1,"textStyle":"font-size:
          16px;","listType":"bullet","start":1,"tag":"ul"}],"direction":"ltr","format":"","indent":0,"type":"root","version":1,"textFormat":1,"textStyle":"font-size:
          16px;"}}'
        theme: blue
        title: ''
        type: ''
        width: 508
      height: 255
      id: '1753182434877'
      position:
        x: -236.30923651380604
        y: 1330.0790319880284
      positionAbsolute:
        x: -236.30923651380604
        y: 1330.0790319880284
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom-note
      width: 508
    - data:
        author: user
        desc: ''
        height: 256
        selected: false
        showAuthor: true
        text: '{"root":{"children":[{"children":[{"detail":0,"format":1,"mode":"normal","style":"font-size:
          16px;","text":"代码执行节点","type":"text","version":1},{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"：","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"paragraph","version":1,"textFormat":1,"textStyle":"font-size:
          16px;"},{"children":[],"direction":null,"format":"start","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          14px;","textFormat":0},{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"经模型交互输出的中间结果，交由进一步处理：","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          16px;","textFormat":0},{"children":[],"direction":"ltr","format":"","indent":0,"type":"paragraph","version":1,"textFormat":0,"textStyle":""},{"children":[{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"执行数据清洗（如剔除无效格式、修正识别错误）、逻辑校验（匹配实验逻辑，过滤矛盾数据 ）；","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"listitem","version":1,"textStyle":"font-size:
          16px;","value":1},{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          16px;","text":"同样配置 “失败时重试 3 次”，确保数据处理质量，为最终输出夯实基础。","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"listitem","version":1,"textStyle":"font-size:
          16px;","value":2}],"direction":"ltr","format":"","indent":0,"type":"list","version":1,"textStyle":"font-size:
          16px;","listType":"bullet","start":1,"tag":"ul"}],"direction":"ltr","format":"","indent":0,"type":"root","version":1,"textFormat":1,"textStyle":"font-size:
          16px;"}}'
        theme: blue
        title: ''
        type: ''
        width: 507
      height: 256
      id: '1753182505676'
      position:
        x: -236.30923651380604
        y: 1619.8759503936867
      positionAbsolute:
        x: -236.30923651380604
        y: 1619.8759503936867
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom-note
      width: 507
    - data:
        author: user
        desc: ''
        height: 169
        selected: false
        showAuthor: true
        text: '{"root":{"children":[{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          14px;","text":"迭代流程产出的结构化数据，可通过","type":"text","version":1},{"detail":0,"format":1,"mode":"normal","style":"font-size:
          14px;","text":"模板转换","type":"text","version":1},{"detail":0,"format":0,"mode":"normal","style":"font-size:
          14px;","text":"环节，将数据映射至标准化格式（如科研报表、数据库字段 ），便于后续统计分析、可视化呈现等应用。","type":"text","version":1}],"direction":"ltr","format":"","indent":0,"type":"paragraph","version":1,"textFormat":0,"textStyle":""}],"direction":"ltr","format":"","indent":0,"type":"root","version":1,"textStyle":"font-size:
          14px;"}}'
        theme: blue
        title: ''
        type: ''
        width: 245
      height: 169
      id: '1753182543484'
      position:
        x: 1524.3306206868185
        y: 545.1117190211345
      positionAbsolute:
        x: 1524.3306206868185
        y: 545.1117190211345
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom-note
      width: 245
    - data:
        author: user
        desc: ''
        height: 292
        selected: false
        showAuthor: true
        text: '{"root":{"children":[{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          14px;","text":"通过","type":"text","version":1},{"detail":0,"format":1,"mode":"normal","style":"font-size:
          14px;","text":"并行迭代 + 多模型协同 + 重试机制","type":"text","version":1},{"detail":0,"format":0,"mode":"normal","style":"font-size:
          14px;","text":"，攻克科学文献数据分散、理解难度高的痛点，高效获取精准抗生素浓度数据；既适配 “大规模文献批量分析” 场景，也支持
          “单篇文献深度挖掘” 需求，可灵活拓展至 “多类型实验数据提取”“跨文献数据对比” 等衍生任务，为科研分析、数据治理提供强支撑。","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          14px;","textFormat":0},{"children":[],"direction":null,"format":"start","indent":0,"type":"paragraph","version":1,"textFormat":0,"textStyle":""},{"children":[{"detail":0,"format":0,"mode":"normal","style":"font-size:
          14px;","text":"简言之，本流程以自动化、智能化链路，打通 “文献输入→内容解析→模型交互→数据规整” 全环节，让抗生素浓度数据提取更高效、更精准，助力科研工作聚焦核心分析环节
          。","type":"text","version":1}],"direction":"ltr","format":"start","indent":0,"type":"paragraph","version":1,"textStyle":"font-size:
          14px;","textFormat":0}],"direction":"ltr","format":"","indent":0,"type":"root","version":1,"textStyle":"font-size:
          14px;"}}'
        theme: blue
        title: ''
        type: ''
        width: 349
      height: 292
      id: '1753182565421'
      position:
        x: 1805.3387624770835
        y: 545.1117190211345
      positionAbsolute:
        x: 1805.3387624770835
        y: 545.1117190211345
      selected: false
      sourcePosition: right
      targetPosition: left
      type: custom-note
      width: 349
    viewport:
      x: 318.0973915049858
      y: 23.561257152410008
      zoom: 0.2720843778535487