语义查找协议如何结合向量数据库和LLM提升聊天机器人体验？

Introduction

This a protocol to enable users to offer writings of theirs so that others may search through them for semantically similar text. A typical use case envisioned is the server side implements a vectordb for text lookup, and the client side either returns the results as-is or feeds them into an LLM prompt for a chatbot experience. Strictly speaking though, neither use of a vectordb nor an LLM is required by the protocol.

这是一个协议，旨在让用户能够提供自己的文本，以便其他人可以搜索其中语义相似的文本。一个典型的设想用例是：服务器端实现一个用于文本查找的向量数据库A database system designed to store and perform high-dimensional semantic similarity searches on vector embeddings of data.，客户端则可以直接返回结果，或者将结果输入到LLM提示中以实现聊天机器人体验。然而，严格来说，该协议并不强制要求使用向量数据库A database system designed to store and perform high-dimensional semantic similarity searches on vector embeddings of data.或LLM。

Protocol Specification

To use, the client sends the server JSON over HTTP(S) of the form:

{"query", <user query>,
 "max_results": <max number of results returned>}

with additional fields as desired and receives a response of the form:

{"output": <list of outputs>}.

The demos in this repo illustrate using one additional field, max_cutoff.

使用时，客户端通过HTTP(S)向服务器发送以下格式的JSON：

{"query", <用户查询>,
 "max_results": <返回的最大结果数量>}

可以根据需要添加额外的字段，并接收以下格式的响应：

{"output": <输出列表>}.

本仓库中的演示示例说明了如何使用一个额外的字段：max_cutoff。

Also while the demos use HTTP for simplicitly, all of this works over HTTPS.

此外，虽然演示为了简便使用了HTTP，但所有功能在HTTPS上同样有效。

Demo Implementations

The repository provides two distinct demo implementations to showcase the protocol's flexibility, catering to different requirements and resource availability.

Full Demo (ChatGPT API required)

This demo allows both looking up text as-is and talking to ChatGPT using similar text as context. This does require access to the ChatGPT API, and one must set environmental variables OPENAI_ORG_ID and OPENAI_API_KEY according to one's OpenAI account.

此演示允许直接查找文本，以及使用相似的文本作为上下文与ChatGPT对话。这确实需要访问ChatGPT API，并且必须根据个人的OpenAI账户设置环境变量 OPENAI_ORG_ID 和 OPENAI_API_KEY。

In data, two example documents are provided - Aaron Swart'z Guerilla Open Access Manifesto and the US Declaration of Independence. However, you can replace them with whatever documents you like for semantic lookup.

在 data 目录中，提供了两个示例文档：Aaron Swartz 的《Guerilla Open Access Manifesto》和美国的《独立宣言》。但是，您可以将它们替换为您喜欢的任何文档以进行语义查找。

In config.yaml, there are three parameters that may be tuned as desired (though the default values should be reasonably good for general purpose use):

在 config.yaml 文件中，有三个参数可以根据需要进行调整（尽管默认值对于一般用途应该相当合适）：

• temp - The temperature used for querying ChatGPT. (用于查询ChatGPT的温度参数。)
• max_results - The maximum number of similar paragraphs that may be returned. (可能返回的相似段落的最大数量。)
• max_cutoff - The minimum embedding distance needed for a paragraph to possibly be returned. (一个段落可能被返回所需的最小嵌入距离。)

In one terminal, run server.py to generate a local web server at 127.0.0.1:5000. When queried via HTTP, this returns as JSON relevant paragraphs from text files in the data directory.

在一个终端中，运行 server.py 以在 127.0.0.1:5000 生成一个本地Web服务器。当通过HTTP查询时，它会以JSON格式返回 data 目录中文本文件的相关段落。

In another terminal, run full_demo.py to be prompted for instructions.

在另一个终端中，运行 full_demo.py，系统将提示您输入指令。

Minimal Demo (ChatGPT API not required)

This demo allows for looking up similar text.

此演示允许查找相似的文本。

Usage is similar to that of the full demo. Set parameters as desired in config.yaml (this time temp is unused). Run server.py to set up the server for querying text in the data directory, and run minimal_demo.py to start the demo.

使用方法与完整演示类似。在 config.yaml 中根据需要设置参数（这次 temp 参数未被使用）。运行 server.py 来设置用于查询 data 目录中文本的服务器，然后运行 minimal_demo.py 来启动演示。

Configuration Parameters Comparison

The behavior of both demos is controlled by a central config.yaml file. The following table outlines the key parameters and their applicability across the two demo types.

Further Directions

Allowing additional input parameters gives this protocol greater flexibility.

允许额外的输入参数为该协议提供了更大的灵活性。

One example not implemented here is instead of just querying top level files in data, one can query inside nested subdirectories, and a file path could be sent to restrict to a particular subdirectory.

一个未在此实现的例子是：不仅可以查询 data 目录中的顶级文件，还可以查询嵌套子目录内部，并且可以发送文件路径以将查询限制在特定的子目录。

This opens up possibilities for more complex document hierarchies and targeted search within large, organized corpora, making the Semantic Lookup Protocol a foundation for building sophisticated, context-aware information retrieval systems.

这为更复杂的文档层次结构和在大型、有组织的语料库中进行针对性搜索开辟了可能性，使语义查找协议一种允许用户共享文本并支持语义相似性搜索的通信协议，通过HTTP(S)传输JSON格式数据实现文本检索功能。成为构建复杂的、上下文感知的信息检索系统的基础。

常见问题（FAQ）

这个语义查找协议一种允许用户共享文本并支持语义相似性搜索的通信协议，通过HTTP(S)传输JSON格式数据实现文本检索功能。必须使用向量数据库A database system designed to store and perform high-dimensional semantic similarity searches on vector embeddings of data.和LLM吗？

不强制要求。协议本身只定义JSON通信格式，向量数据库A database system designed to store and perform high-dimensional semantic similarity searches on vector embeddings of data.用于高效检索，LLM用于增强对话体验，两者都是可选实现方式。

如何配置语义查找的返回结果数量和质量？

通过config.yaml中的max_results控制返回数量，max_cutoff控制相似度阈值。完整演示还可调整temperature参数影响LLM回答风格。

完整演示和最小演示的主要区别是什么？

完整演示需要ChatGPT API密钥，支持检索后与LLM对话；最小演示无需API，仅返回相似文本。两者共享相同的服务器和基础检索功能。