Skip to content
Nappai Documentation

Astra Vectorize

This component helps Nappai understand the text in your data by choosing the best way to create “embeddings” – special codes that represent the meaning of your text. It uses Astra DB, a powerful database, to do this efficiently.

Relationship with Astra DB

This component connects to Astra DB to generate text embeddings. It lets you select different embedding models and providers offered by Astra DB, ensuring Nappai uses the most suitable method for your data.

Inputs

  • Provider: Choose the company that provides the embedding model (e.g., OpenAI, Cohere). This is required.
  • Model Name: Select the specific embedding model from the provider you chose. This is required. The available options will change depending on the provider you select.
  • Model Parameters: (Advanced) Allows you to fine-tune the embedding model with additional settings. This is optional.
  • Credential: Enter your credentials to access your Astra DB account. This is required for Nappai to connect to your Astra DB and generate embeddings.

Outputs

  • Vectorize: This output provides the configuration settings Nappai needs to create embeddings using your chosen provider and model. This configuration is then used by other Nappai components to process your data.

Usage Example

Imagine you want Nappai to analyze customer reviews. You would use this component to:

  1. Select a provider (e.g., OpenAI).
  2. Choose a specific model from OpenAI (e.g., text-embedding-ada-002).
  3. (Optional) Adjust any advanced model parameters.
  4. Enter your Astra DB credentials.

Nappai will then use this configuration to create embeddings for your customer reviews, allowing other Nappai components to analyze the sentiment and topics discussed.

Templates

This component is not currently used in any predefined templates.

  • Astra DB: Implementation of Vector Store using Astra DB with search capabilities: This component uses the embeddings generated by Astra Vectorize to create a searchable database of your text data.

Tips and Best Practices

  • Start with a readily available model from a reputable provider to ensure good performance.
  • Only use advanced model parameters if you have experience with embedding models. Incorrect settings can impact the quality of your embeddings.
  • Ensure your Astra DB credentials are correct to avoid connection errors.

Security Considerations

Always protect your Astra DB credentials. Do not share them with unauthorized individuals. Nappai uses secure methods to handle your credentials, but it’s crucial to maintain good security practices.