ChatGPT/FastGPT
1
0
Fork 0
mirror of https://github.com/labring/FastGPT.git synced 2025-07-23 21:13:50 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
c031e6dcc9a2e16abe6e0cd37784fdc0dc38cba4
FastGPT/docSite/content/docs/development/openapi/dataset.md
Archer c031e6dcc9 4.6.7-alpha commit (#743) 
Co-authored-by: Archer <545436317@qq.com>
Co-authored-by: heheer <71265218+newfish-cmyk@users.noreply.github.com>
2024-01-19 11:17:28 +08:00

35 KiB
Raw Blame History

title, description, icon, draft, toc, weight
title description icon draft toc weight
知识库接口 FastGPT OpenAPI 知识库接口 dataset false true 853
如何获取知识库ID（datasetId） 如何获取文件集合ID（collection_id）

创建训练订单

请求示例

curl --location --request POST 'https://api.fastgpt.in/api/support/wallet/bill/createTrainingBill' \
--header 'Authorization: Bearer {{apikey}}' \
--header 'Content-Type: application/json' \
--data-raw '{
    "name": "可选，自定义订单名称，例如：文档训练-fastgpt.docx"
}'

响应结果

data 为 billId，可用于添加知识库数据时进行账单聚合。

{
  "code": 200,
  "statusText": "",
  "message": "",
  "data": "65112ab717c32018f4156361"
}

知识库添加数据

{{< tabs tabTotal="4" >}} {{< tab tabName="请求示例" >}} {{< markdownify >}}

curl --location --request POST 'https://api.fastgpt.in/api/core/dataset/data/pushData' \
--header 'Authorization: Bearer apikey' \
--header 'Content-Type: application/json' \
--data-raw '{
    "collectionId": "64663f451ba1676dbdef0499",
    "trainingMode": "chunk",
    "prompt": "可选。qa 拆分引导词，chunk 模式下忽略",
    "billId": "可选。如果有这个值，本次的数据会被聚合到一个订单中，这个值可以重复使用。可以参考 [创建训练订单] 获取该值。",
    "data": [
        {
            "q": "你是谁？",
            "a": "我是FastGPT助手"
        },
        {
            "q": "你会什么？",
            "a": "我什么都会",
            "indexes": [{
                "type":"custom",
                "text":"你好"
            }]
        }
    ]
}'

{{< /markdownify >}} {{< /tab >}}

{{< tab tabName="参数说明" >}} {{< markdownify >}}

需要先了解 FastGPT 的多路索引概念：

在 FastGPT 中，你可以为一组数据创建多个索引，如果不指定索引，则系统会自动取对应的 chunk 作为索引。例如前面的请求示例中：

q：你是谁？a:我是FastGPT助手 它的indexes属性为空，意味着不自定义索引，而是使用默认的索引（你是谁？\n我是FastGPT助手）。

在第二组数据中q:你会什么？a:我什么都会指定了一个你好的索引，因此这组数据的索引为你好。

{
    "collectionId": "文件集合的ID，参考上面的第二张图",
    "mode": "chunk | qa ", //  chunk 模式: 可自定义索引。qa 模型：无法自定义索引，会自动取 data 中的 q 作为数据，让模型自动生成问答对和索引。
    "prompt": "QA 拆分提示词，需严格按照模板，建议不要传入。",
    "data": [
        {
            "q": "生成索引的内容，index 模式下最大 tokens 为3000，建议不超过 1000",
            "a": "预期回答/补充",
            "indexes": "自定义索引",
        },
        {
            "q": "xxx",
            "a": "xxxx"
        }
    ],
    
}

{{< /markdownify >}} {{< /tab >}}

{{< tab tabName="响应例子" >}} {{< markdownify >}}

{
    "code": 200,
    "statusText": "",
    "data": {
        "insertLen": 1, // 最终插入成功的数量
        "overToken": [], // 超出 token 的
       
        "repeat": [], // 重复的数量
        "error": [] // 其他错误
    }
}

{{< /markdownify >}} {{< /tab >}}

{{< tab tabName="QA Prompt 模板" >}} {{< markdownify >}}

{{theme}} 里的内容可以换成数据的主题。默认为：它们可能包含多个主题内容

我会给你一段文本，{{theme}}，学习它们，并整理学习成果，要求为：
1. 提出最多 25 个问题。
2. 给出每个问题的答案。
3. 答案要详细完整，答案可以包含普通文字、链接、代码、表格、公示、媒体链接等 markdown 元素。
4. 按格式返回多个问题和答案:

Q1: 问题。
A1: 答案。
Q2:
A2:
……

我的文本："""{{text}}"""

{{< /markdownify >}} {{< /tab >}}

{{< /tabs >}}

搜索测试

{{< tabs tabTotal="3" >}} {{< tab tabName="请求示例" >}} {{< markdownify >}}

curl --location --request POST 'https://api.fastgpt.in/api/core/dataset/searchTest' \
--header 'Authorization: Bearer fastgpt-xxxxx' \
--header 'Content-Type: application/json' \
--data-raw '{
    "datasetId": "知识库的ID",
    "text": "导演是谁",
    "limit": 5000,
    "similarity": 0,
    "searchMode": "embedding",
    "usingReRank": false
}'

{{< /markdownify >}} {{< /tab >}}

{{< tab tabName="参数说明" >}} {{< markdownify >}}

  • datasetId - 知识库ID
  • text - 需要测试的文本
  • limit - 最大 tokens 数量
  • similarity - 最低相关度（0~1，可选）
  • searchMode - 搜索模式：embedding | fullTextRecall | mixedRecall
  • usingReRank - 使用重排

{{< /markdownify >}} {{< /tab >}}

{{< tab tabName="响应示例" >}} {{< markdownify >}}

返回 top k 结果， limit 为最大 Tokens 数量，最多 20000 tokens。

{
  "code": 200,
  "statusText": "",
  "data": [
    {
        "id": "65599c54a5c814fb803363cb",
        "q": "你是谁",
        "a": "我是FastGPT助手",
        "datasetId": "6554684f7f9ed18a39a4d15c",
        "collectionId": "6556cd795e4b663e770bb66d",
        "sourceName": "GBT 15104-2021 装饰单板贴面人造板.pdf",
        "sourceId": "6556cd775e4b663e770bb65c",
        "score": 0.8050316572189331
    },
    ......
  ]
}

{{< /markdownify >}} {{< /tab >}}

{{< /tabs >}}

更多接口

目前未整理，简陋导出：

POST 知识库搜索测试

POST /core/dataset/searchTest

Body Parameters

{
  "datasetId": "656c2ccff7f114064daa72f6",
  "text": "导演是谁",
  "limit": 1500,
  "searchMode": "embedding",
  "usingReRank": true,
  "similarity": 0.5
}

Params

Name Location Type Required Description
Authorization header string no none
body body object no none
» datasetId body string yes none
» text body string yes none
» limit body integer no none
» searchMode body search mode yes none
» usingReRank body boolean no none
» similarity body similary no none

Response Examples

成功

{
  "code": 200,
  "statusText": "",
  "message": "",
  "data": {
    "list": [
      {
        "id": "65962b23f5fac58e46330dfd",
        "q": "# 快速了解 FastGPT\nFastGPT 的能力与优势\n\nFastGPT 是一个基于 LLM 大语言模型的知识库问答系统，提供开箱即用的数据处理、模型调用等能力。同时可以通过 Flow 可视化进行工作流编排，从而实现复杂的问答场景！\n\n🤖\n\nFastGPT 在线使用：[https://fastgpt.in](https://fastgpt.in)\n\n|     |     |\n| --- | --- |\n| ![](https://doc.fastgpt.in/imgs/intro1.png) | ![](https://doc.fastgpt.in/imgs/intro2.png) |\n| ![](https://doc.fastgpt.in/imgs/intro3.png) | ![](https://doc.fastgpt.in/imgs/intro4.png) |\n\n",
        "a": "",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65962b2089642fd209da3b03",
        "sourceName": "https://doc.fastgpt.in/docs/intro/",
        "sourceId": "https://doc.fastgpt.in/docs/intro/",
        "score": [
          {
            "type": "embedding",
            "value": 0.8036568760871887,
            "index": 20
          },
          {
            "type": "fullText",
            "value": 1.168349443855932,
            "index": 2
          },
          {
            "type": "reRank",
            "value": 0.9870296135626316,
            "index": 0
          },
          {
            "type": "rrf",
            "value": 0.04366449476962486,
            "index": 0
          }
        ]
      },
      {
        "id": "65962b24f5fac58e46330dff",
        "q": "# 快速了解 FastGPT\n## FastGPT 能力\n### 2. 简单易用的可视化界面\nFastGPT 采用直观的可视化界面设计，为各种应用场景提供了丰富实用的功能。通过简洁易懂的操作步骤，可以轻松完成 AI 客服的创建和训练流程。\n\n![](https://doc.fastgpt.in/imgs/ability5.png)\n\n",
        "a": "",
        "chunkIndex": 2,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65962b2089642fd209da3b03",
        "sourceName": "https://doc.fastgpt.in/docs/intro/",
        "sourceId": "https://doc.fastgpt.in/docs/intro/",
        "score": [
          {
            "type": "embedding",
            "value": 0.8152669668197632,
            "index": 3
          },
          {
            "type": "fullText",
            "value": 1.0511363636363635,
            "index": 8
          },
          {
            "type": "reRank",
            "value": 0.9287972729281414,
            "index": 14
          },
          {
            "type": "rrf",
            "value": 0.04265696347031964,
            "index": 1
          }
        ]
      },
      {
        "id": "65962b25f5fac58e46330e00",
        "q": "# 快速了解 FastGPT\n## FastGPT 能力\n### 3. 自动数据预处理\n提供手动输入、直接分段、LLM 自动处理和 CSV 等多种数据导入途径，其中“直接分段”支持通过 PDF、WORD、Markdown 和 CSV 文档内容作为上下文。FastGPT 会自动对文本数据进行预处理、向量化和 QA 分割，节省手动训练时间，提升效能。\n\n![](https://doc.fastgpt.in/imgs/ability2.png)\n\n",
        "a": "",
        "chunkIndex": 3,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65962b2089642fd209da3b03",
        "sourceName": "https://doc.fastgpt.in/docs/intro/",
        "sourceId": "https://doc.fastgpt.in/docs/intro/",
        "score": [
          {
            "type": "embedding",
            "value": 0.8158369064331055,
            "index": 2
          },
          {
            "type": "fullText",
            "value": 1.014030612244898,
            "index": 20
          },
          {
            "type": "reRank",
            "value": 0.9064876908461501,
            "index": 17
          },
          {
            "type": "rrf",
            "value": 0.04045823457588163,
            "index": 2
          }
        ]
      },
      {
        "id": "65a7e1e8fc13bdf20fd46d41",
        "q": "# 快速了解 FastGPT\n## FastGPT 能力\n### 5. 强大的 API 集成\nFastGPT 对外的 API 接口对齐了 OpenAI 官方接口，可以直接接入现有的 GPT 应用，也可以轻松集成到企业微信、公众号、飞书等平台。\n\n![](https://doc.fastgpt.in/imgs/ability4.png)",
        "a": "",
        "chunkIndex": 66,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7e1d4fc13bdf20fd46abe",
        "sourceName": "dataset - 2024-01-04T151625.388.csv",
        "sourceId": "65a7e1d2fc13bdf20fd46abc",
        "score": [
          {
            "type": "embedding",
            "value": 0.803692102432251,
            "index": 18
          },
          {
            "type": "fullText",
            "value": 1.0511363636363635,
            "index": 7
          },
          {
            "type": "reRank",
            "value": 0.9177460552422909,
            "index": 15
          },
          {
            "type": "rrf",
            "value": 0.03970501147383226,
            "index": 3
          }
        ]
      },
      {
        "id": "65a7be319d96e21823f69c9b",
        "q": "FastGPT Flow 的工作流设计方案提供了哪些操作？",
        "a": "FastGPT Flow 的工作流设计方案提供了数据预处理、各类 AI 应用设置、调试测试及结果反馈等操作。",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7be059d96e21823f69af5",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7be059d96e21823f69ae8",
        "score": [
          {
            "type": "embedding",
            "value": 0.8283981680870056,
            "index": 0
          },
          {
            "type": "reRank",
            "value": 0.9620363047907355,
            "index": 4
          },
          {
            "type": "rrf",
            "value": 0.03177805800756621,
            "index": 4
          }
        ]
      },
      {
        "id": "65a7be389d96e21823f69d58",
        "q": "FastGPT Flow 的实验室预约示例中使用了哪些参数？",
        "a": "FastGPT Flow 的实验室预约示例中使用了姓名、时间和实验室名称等参数。",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7be059d96e21823f69af5",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7be059d96e21823f69ae8",
        "score": [
          {
            "type": "embedding",
            "value": 0.8143455386161804,
            "index": 9
          },
          {
            "type": "reRank",
            "value": 0.9806919138043485,
            "index": 1
          },
          {
            "type": "rrf",
            "value": 0.0304147465437788,
            "index": 5
          }
        ]
      },
      {
        "id": "65a7be309d96e21823f69c78",
        "q": "FastGPT Flow 是什么？",
        "a": "FastGPT Flow 是一款基于大型语言模型的知识库问答系统，通过引入 Flow 可视化工作流编排技术，提供了一个即插即用的解决方案。",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7be059d96e21823f69af5",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7be059d96e21823f69ae8",
        "score": [
          {
            "type": "embedding",
            "value": 0.8115077018737793,
            "index": 11
          },
          {
            "type": "reRank",
            "value": 0.9686195704870232,
            "index": 3
          },
          {
            "type": "rrf",
            "value": 0.029513888888888888,
            "index": 6
          }
        ]
      },
      {
        "id": "65a7be389d96e21823f69d5e",
        "q": "FastGPT Flow 的实验室预约示例中的代码实现了哪些功能？",
        "a": "FastGPT Flow 的实验室预约示例中的代码实现了预约实验室、修改预约、查询预约和取消预约等功能。",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7be059d96e21823f69af5",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7be059d96e21823f69ae8",
        "score": [
          {
            "type": "embedding",
            "value": 0.8166953921318054,
            "index": 1
          },
          {
            "type": "reRank",
            "value": 0.8350804533361768,
            "index": 20
          },
          {
            "type": "rrf",
            "value": 0.028474711270410194,
            "index": 8
          }
        ]
      },
      {
        "id": "65a7be389d96e21823f69d4f",
        "q": "FastGPT Flow 的联网搜索示例中使用了哪些参数？",
        "a": "FastGPT Flow 的联网搜索示例中使用了搜索关键词、Google 搜索的 API 密钥和自定义搜索引擎 ID。",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7be059d96e21823f69af5",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7be059d96e21823f69ae8",
        "score": [
          {
            "type": "embedding",
            "value": 0.8025297522544861,
            "index": 21
          },
          {
            "type": "reRank",
            "value": 0.9730876959261983,
            "index": 2
          },
          {
            "type": "rrf",
            "value": 0.028068137824235385,
            "index": 10
          }
        ]
      },
      {
        "id": "65a7e1e8fc13bdf20fd46d55",
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7e1d4fc13bdf20fd46abe",
        "sourceName": "dataset - 2024-01-04T151625.388.csv",
        "sourceId": "65a7e1d2fc13bdf20fd46abc",
        "q": "# 快速了解 FastGPT\n## FastGPT 特点\n1.  **项目开源**\n    \n    FastGPT 遵循附加条件 Apache License 2.0 开源协议，你可以 [Fork](https://github.com/labring/FastGPT/fork) 之后进行二次开发和发布。FastGPT 社区版将保留核心功能，商业版仅在社区版基础上使用 API 的形式进行扩展，不影响学习使用。\n    \n2.  **独特的 QA 结构**\n    \n    针对客服问答场景设计的 QA 结构，提高在大量数据场景中的问答准确性。\n    \n3.  **可视化工作流**\n    \n    通过 Flow 模块展示了从问题输入到模型输出的完整流程，便于调试和设计复杂流程。\n    \n4.  **无限扩展**\n    \n    基于 API 进行扩展，无需修改 FastGPT 源码，也可快速接入现有的程序中。\n    \n5.  **便于调试**\n    \n    提供搜索测试、引用修改、完整对话预览等多种调试途径。\n    \n6.  **支持多种模型**\n    \n    支持 GPT、Claude、文心一言等多种 LLM 模型，未来也将支持自定义的向量模型。",
        "a": "",
        "chunkIndex": 67,
        "score": [
          {
            "type": "fullText",
            "value": 1.0340073529411764,
            "index": 12
          },
          {
            "type": "reRank",
            "value": 0.9542227274192233,
            "index": 9
          },
          {
            "type": "rrf",
            "value": 0.027272727272727275,
            "index": 11
          }
        ]
      },
      {
        "id": "65a7be319d96e21823f69c8f",
        "q": "FastGPT Flow 的工作流设计中，模块之间如何进行组合和组装？",
        "a": "FastGPT Flow 允许用户在核心工作流模块中进行自由组合和组装，从而衍生出一个新的模块。",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7be059d96e21823f69af5",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7be059d96e21823f69ae8",
        "score": [
          {
            "type": "embedding",
            "value": 0.8098832368850708,
            "index": 13
          },
          {
            "type": "reRank",
            "value": 0.9478657435317039,
            "index": 12
          },
          {
            "type": "rrf",
            "value": 0.027212143650499815,
            "index": 12
          }
        ]
      },
      {
        "id": "65a7be359d96e21823f69ce0",
        "q": "FastGPT Flow 的模块的输入和输出如何连接？",
        "a": "FastGPT Flow 的模块的输入和输出通过连接点进行连接，连接点的颜色代表了不同的数据类型。",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7be059d96e21823f69af5",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7be059d96e21823f69ae8",
        "score": [
          {
            "type": "embedding",
            "value": 0.8060981035232544,
            "index": 16
          },
          {
            "type": "reRank",
            "value": 0.9530133603823691,
            "index": 10
          },
          {
            "type": "rrf",
            "value": 0.027071520029266508,
            "index": 13
          }
        ]
      },
      {
        "id": "65a7be319d96e21823f69c98",
        "q": "FastGPT Flow 的工作流设计方案能够满足哪些问答场景？",
        "a": "FastGPT Flow 的工作流设计方案能够满足基本的 AI 知识库问答需求，并适应各种复杂的问答场景，例如联网搜索、数据库操作、数据实时更新、消息通知等。",
        "chunkIndex": 0,
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7be059d96e21823f69af5",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7be059d96e21823f69ae8",
        "score": [
          {
            "type": "embedding",
            "value": 0.814436137676239,
            "index": 8
          },
          {
            "type": "reRank",
            "value": 0.8814109034236719,
            "index": 19
          },
          {
            "type": "rrf",
            "value": 0.026992753623188405,
            "index": 16
          }
        ]
      },
      {
        "id": "65a7e058fc13bdf20fd46577",
        "datasetId": "6593e137231a2be9c5603ba7",
        "collectionId": "65a7e01efc13bdf20fd45815",
        "sourceName": "FastGPT软著.pdf",
        "sourceId": "65a7e01dfc13bdf20fd457f3",
        "q": "FastGPT Flow 工作流设计112312 3123213123 232321312 21312 23一、介绍FastGPT 作为一款基于大型语言模型（LLM）的知识库问答系统，旨在为用户提供一个即插即用的解决方案。它集成了数据处理、模型调用等多项功能，通过引入 Flow 可视化工作流编排技术，进一步增强了对复杂问答场景的支持能力。本文将重点介绍 FastGPT Flow工作流的设计方案和应用优势。\nFastGPT Flow 工 作 流 采 用 了 React Flow 框 架 作 为 UI 底 座 ， 结 合 自 研 的 FlowController 实现工作流的运行。FastGPT 使用 Flow 模块为用户呈现了一个直观、可视化的界面，从而简化了 AI 应用工作流程的设计和管理方式。React Flow 的应用使得用户能够以图形化的方式组织和编排工作流，这不仅使得工作流的创建过程更为直观，同时也为用户提供了强大且灵活的工作流编辑器。在 FastGPT Flow 工作流设计中，核心工作流模块包括用户引导、问题输入、知识库检索、AI 文本生成、问题分类、结构化内容提取、指定回复、应用调用和 HTTP 扩展，并允许用户在这类模块中进行自由组合和组装，从而衍生出一个新的模块。",
        "a": "",
        "chunkIndex": 0,
        "score": [
          {
            "type": "fullText",
            "value": 1.0229779411764706,
            "index": 15
          },
          {
            "type": "reRank",
            "value": 0.9577545043363116,
            "index": 8
          },
          {
            "type": "rrf",
            "value": 0.026992753623188405,
            "index": 17
          }
        ]
      }
    ],
    "duration": "2.978s",
    "searchMode": "mixedRecall",
    "limit": 1500,
    "similarity": 0.1,
    "usingReRank": true,
    "usingSimilarityFilter": true
  }
}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

HTTP Status Code 200

Name Type Required Restrictions Title description
» code integer true none none
» statusText string true none none
» message string true none none
» data object true none none
»» list [object] true none none
»»» id string true none none
»»» q string true none none
»»» a string true none none
»»» chunkIndex integer true none none
»»» datasetId string true none none
»»» collectionId string true none none
»»» sourceName string true none none
»»» sourceId string true none none
»»» score [object] true none none
»»»» type string true none none
»»»» value number true none none
»»»» index integer true none none
»» duration string true none none
»» searchMode string true none none
»» limit integer true none none
»» similarity number true none none
»» usingReRank boolean true none none
»» usingSimilarityFilter boolean true none none

openapi/知识库/知识库crud

GET 获取知识库列表

GET /core/dataset/list

Params

Name Location Type Required Description
parentId query string no 父级的ID
Authorization header string no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

GET 获取知识库详情

GET /core/dataset/detail

Params

Name Location Type Required Description
id query string no 知识库id
Authorization header string no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

openapi/知识库/集合crud

POST 获取知识库集合列表

POST /core/dataset/collection/list

Body Parameters

{
  "pageNum": 1,
  "pageSize": 10,
  "datasetId": "6597ca43e26f2a90a1501414",
  "parentId": null,
  "searchText": "",
  "simple": true
}

Params

Name Location Type Required Description
Authorization header string no none
body body object no none
» pageNum body integer no none
» pageSize body integer no none
» datasetId body string yes none
» parentId body null no none
» searchText body string no none
» simple body boolean no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

GET 获取集合详情

GET /core/dataset/collection/detail

Params

Name Location Type Required Description
id query string no 知识库id
Authorization header string no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

PUT 更新集合

PUT /core/dataset/collection/update

Body Parameters

{
  "id": "6597ce094e10ee661f0891c8",
  "parentId": null,
  "name": "222"
}

Params

Name Location Type Required Title Description
Authorization header string no none
body body object no none
» id body string yes none
» parentId body null no 父级的id none
» name body string no 名称 none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

POST 创建空集合（文件夹或者一个空集合）

POST /core/dataset/collection/create

Body Parameters

{
  "datasetId": "6597ca43e26f2a90a1501414",
  "parentId": null,
  "name": "集合名",
  "type": "folder",
  "metadata": {}
}

Params

Name Location Type Required Title Description
Authorization header string no none
body body object no none
» datasetId body string yes none
» parentId body null no none
» name body string yes none
» type body collection type yes none
» metadata body object no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

POST 创建文本集合

POST /core/dataset/collection/create/text

Body Parameters

{
  "text": "xxxxxxxxxxxxxx",
  "datasetId": "6593e137231a2be9c5603ba7",
  "parentId": null,
  "name": "测试",
  "trainingType": "qa",
  "chunkSize": 8000,
  "chunkSplitter": "",
  "qaPrompt": "",
  "metadata": {}
}

Params

Name Location Type Required Title Description
Authorization header string no none
body body object no none
» datasetId body string no none
» parentId body null no none
» name body string yes none
» text body string yes 原文本 none
» trainingType body training type yes none
» chunkSize body integer no 分块大小 none
» chunkSplitter body string no 自定义最高优先级的分段符号 none
» qaPrompt body string no none
» metadata body object no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

POST 创建网络链接集合

POST /core/dataset/collection/create/link

Body Parameters

{
  "link": "https://doc.fastgpt.in/docs/course/quick-start/",
  "datasetId": "6593e137231a2be9c5603ba7",
  "parentId": null,
  "trainingType": "chunk",
  "chunkSize": 512,
  "chunkSplitter": "",
  "qaPrompt": "",
  "metadata": {
    "webPageSelector": ".docs-content"
  }
}

Params

Name Location Type Required Title Description
Authorization header string no none
body body object no none
» datasetId body string yes none
» parentId body null no none
» link body string yes none
» trainingType body training type yes none
» chunkSize body integer no none
» chunkSplitter body string no none
» qaPrompt body string no none
» metadata body object no none
»» webPageSelector body string no web选择器 none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

DELETE 删除一个集合

DELETE /core/dataset/collection/delete

Params

Name Location Type Required Title Description
id query string no 知识库id
Authorization header string no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

openapi/知识库/数据crud

POST 获取数据列表

POST /core/dataset/data/list

Body Parameters

{
  "pageNum": 1,
  "pageSize": 10,
  "collectionId": "65a8d2700d70d3de0bf09186",
  "searchText": ""
}

Params

Name Location Type Required Title Description
Authorization header string no none
body body object no none
» pageNum body integer yes none
» pageSize body integer yes none
» searchText body string yes none
» collectionId body string yes none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

GET 获取数据详情

GET /core/dataset/data/detail

Params

Name Location Type Required Title Description
id query string yes none
Authorization header string no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

DELETE 删除一条数据

DELETE /core/dataset/data/delete

Params

Name Location Type Required Title Description
id query string no none
Authorization header string no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

PUT 更新数据

PUT /core/dataset/data/update

Body Parameters

{
  "id": "6597ce094e10ee661f0891c8",
  "parentId": null,
  "name": "222"
}

Params

Name Location Type Required Title Description
Authorization header string no none
body body object no none
» id body string yes none
» q body string yes none
» a body string no none
» indexes body [数据自定义向量] no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

POST 知识库插入记录（批量插入）

POST /core/dataset/data/pushData

Body Parameters

{
  "collectionId": "string",
  "data": [
    {
      "a": "string",
      "q": "string",
      "chunkIndex": 1
    }
  ],
  "trainingMode": "string",
  "promot": "string",
  "billId": ""
}

Params

Name Location Type Required Title Description
Authorization header string no none
body body object no none
» collectionId body string yes none
» data body [object] yes none
»» a body string no none
»» q body string no none
»» chunkIndex body integer no none
» trainingMode body training type no none
» promot body string no none
» billId body string no none

Response Examples

200 Response

{}

Responses

HTTP Status Code Meaning Description Data schema
200 OK 成功 Inline

Responses Data Schema

Data Schema

similary

1

Attribute

Name Type Required Restrictions Title Description
anonymous integer false none none

search mode

"embedding"

Attribute

Name Type Required Restrictions Title Description
anonymous string false none none

Enum

Name Value
anonymous embedding
anonymous fullTextRecall
anonymous mixedRecall

training type

"chunk"

Attribute

Name Type Required Restrictions Title Description
anonymous string false none none

Enum

Name Value
anonymous chunk
anonymous qa

collection type

"folder"

Attribute

Name Type Required Restrictions Title Description
anonymous string false none none

Enum

Name Value
anonymous folder
anonymous virtual
anonymous link
anonymous file

数据自定义向量

{
  "defaultIndex": true,
  "type": "string",
  "text": "string"
}

Attribute

Name Type Required Restrictions Title Description
defaultIndex boolean false none 是否为默认
type string true none none
text string true none 索引文本