description: >- Generate multiple queries from different perspectives for a given user input query.

Multi Query Retriever

The Multi Query Retriever is a specialized retriever that generates multiple queries from different perspectives for a given user input query. This approach helps overcome some limitations of distance-based similarity search in vector databases.

Input Parameters

Vector Store
- Label: Vector Store
- Name: vectorStore
- Type: VectorStore
- Description: The vector store to be used for document retrieval.
Language Model
- Label: Language Model
- Name: model
- Type: BaseLanguageModel
- Description: The language model used to generate alternative questions.
Prompt
- Label: Prompt
- Name: modelPrompt
- Type: string
- Description: The prompt template for the language model to generate alternative questions. Use {question} to refer to the original question.
- Default: A predefined prompt that instructs the AI to generate 3 different versions of the given user question.

Functionality

The node initializes with the provided vector store, language model, and prompt template.
When executed, it takes the user’s input query and uses the language model to generate multiple alternative questions based on the prompt.
These alternative questions are then used to query the vector store, potentially retrieving a more diverse and comprehensive set of relevant documents.

Usage

This retriever is particularly useful in scenarios where:

The user's initial query might not capture all aspects of their information need.
The desired information could be expressed in various ways in the document collection.
A broader exploration of the topic is beneficial.

Output

The node returns a MultiQueryRetriever instance, which can be used to retrieve documents based on the original query and its generated alternatives.

{% hint style="info" %} This section is a work in progress. We appreciate any help you can provide in completing this section. Please check our Contribution Guide to get started. {% endhint %}