Elsai NLI#

The Elsai NLI (Natural Language Interface) package enables users to ask natural language questions directly over structured data sources like CSV files using LLM-powered agents.

Prerequisites#

  • Python >= 3.9

Installation#

To install the elsai-nli package:

pip install --index-url https://elsai-core-package.optisolbusiness.com/root/elsai-nli/ elsai-nli==0.1.0

Component#

1. CSVAgentHandler#

CSVAgentHandler is a class that allows querying CSV files using natural language. It leverages an LLM and an agent interface to convert questions into structured operations.

from elsai_nli.natural_language_interface import CSVAgentHandler

agent = CSVAgentHandler(
    csv_files="path_to_csv",  # Can be a single file path or a list of paths
    model="llm_model",
    verbose=True,
    agent_type="openai-functions"
)

response = agent.ask_question("What is the total sales amount for each product?")

Required Parameters:

  • csv_files – Path or list of paths to CSV files to be queried

  • model – Identifier or instance of the LLM to be used

  • agent_type – Type of agent backend (e.g., openai-functions)

  • verbose – (Optional) Enables debug/log output during execution