Skip to content
Nappai Documentation

CSVAgent

This component, CSVAgent, lets you build a helpful assistant that can analyze and process information stored in CSV (Comma Separated Values) files. Think of it as a way to give your data a voice and automate tasks related to it. Nappai uses this to connect your CSV data with powerful language models, making it easy to extract insights and perform actions.

Relationship with Language Models

This component works with various language models (like those from OpenAI or others). You choose the language model you want to use, and the CSVAgent uses it to understand and process the data in your CSV file. The better the language model, the more accurately and intelligently the assistant will work with your data.

Inputs

  • Language Model: Select the language model you want your assistant to use. This is crucial; the model’s capabilities will directly impact how well the assistant understands and processes your CSV data. You must select a language model.

  • File Path: Specify the location of your CSV file. Nappai needs this to know where to find the data your assistant will work with. You must provide a valid file path.

  • Agent Type (Advanced): This setting lets you fine-tune how your assistant interacts with the data. It offers different pre-defined approaches. If you’re unsure, leave it at the default setting (“openai-tools”). This is an optional advanced setting.

Outputs

The CSVAgent doesn’t directly produce outputs you see in the dashboard. Instead, it creates a powerful “agent” behind the scenes. This agent is then used by other Nappai components to perform actions based on your CSV data. For example, you might use this agent to ask questions about your data, generate reports, or automate data entry tasks.

Usage Example

Let’s say you have a CSV file containing sales data. You can use the CSVAgent to create an assistant that can answer questions like: “What were the total sales in March?”, “Which product sold the most?”, or “What is the average sale price?”. You would select your preferred language model, specify the path to your sales data CSV file, and then use other Nappai components (like a “Chat Output” component) to interact with the created agent and see the results.

Templates

[This section will be populated with links to relevant flow templates once available.]

Many components in Nappai can work with the CSVAgent to create powerful workflows. Here are a few examples:

  • Chat Output: Displays the answers from your CSV data assistant in a chat-like interface.
  • Explain SQL results: If your analysis involves SQL queries, this component helps interpret the results.
  • SQL Query Constructor: Helps you build SQL queries to interact with your data.
  • Text Output: Displays the results of your analysis as plain text.

Tips and Best Practices

  • Choose the right language model: Select a language model appropriate for the complexity of your data and the tasks you want to perform. More powerful models generally provide better results but may be more expensive.
  • Ensure your CSV file is well-formatted: A clean, well-structured CSV file will lead to better results.
  • Start with simple tasks: Begin by asking simple questions about your data to test your setup before moving on to more complex analyses.

Security Considerations

  • Ensure that the CSV file you upload does not contain sensitive information.
  • Choose a secure and trusted language model provider. Nappai will handle data according to its security policies. Refer to Nappai’s security documentation for details.