Skip to content
Nappai Documentation

Scorer Retriever

The Scorer Retriever component is designed to help users find relevant documents by searching through a vector store and applying a score threshold to filter results. It allows users to perform searches based on queries and returns documents along with their scores.

Relationship with Vector Store Technology

The Scorer Retriever component interacts with a vector store, which is a specialized database that stores documents in a way that makes them easy to search and retrieve. By using this technology, the component can efficiently find documents that match a user’s search query.

Inputs

  • Search Query: This is the search term or phrase you want to use to find documents.
  • Parent Document Vectorstore: This is the database where the documents are stored and from which the search will be conducted.
  • Score Threshold: This sets the minimum score a document must have to be included in the results. A higher threshold means fewer, but more relevant, results.
  • Number of Results: This determines how many documents will be returned from the search. The default is 10, but you can adjust it as needed.

Outputs

The Scorer Retriever component produces two main outputs:

  • Retriever: This is the tool configured to perform the search. It helps in executing the search operation.
  • Search Results: A list of documents that match the search criteria. These documents are the ones you can review and use in your workflow.

Usage Example

Imagine you are managing a large database of customer feedback. You can use the Scorer Retriever to search for feedback related to a specific product feature by entering a relevant search query. The component will return the most relevant feedback documents, allowing you to quickly assess customer opinions.

Templates

Currently, there are no specific templates where this component is pre-configured. However, you can easily integrate it into your custom workflows.

  • Document Uploader: Allows you to add new documents to the vector store.
  • Feedback Analyzer: Helps in analyzing the sentiment of the retrieved documents.

Tips and Best Practices

  • Use specific and clear search queries to get the most relevant results.
  • Adjust the score threshold to balance between the quantity and quality of the results.
  • Start with the default number of results and adjust based on your needs.

Security Considerations

Ensure that the vector store containing sensitive documents is secured and access is controlled to prevent unauthorized searches and data retrieval.