Simplifying API Integration with Natural Language: A Guide to Gorilla OpenFunctions

25 Apr 2025
Simplifying API Integration with Natural Language: A Guide to Gorilla OpenFunctions

In today's fast-paced development environment, integrating APIs efficiently can be a daunting task, especially when it comes to generating accurate requests. The Gorilla OpenFunctions allows developers to leverage large language models (LLMs) to translate natural language instructions into executable API calls. This article will explore how to utilize the Cognitive Action Generate API Calls from Natural Language to enhance your application's API management capabilities.

Prerequisites

Before you dive into using the Cognitive Actions, ensure you have the following:

  • Access to the Gorilla OpenFunctions platform, including an API key for authentication.
  • Familiarity with JSON for structuring input and output data.
  • Basic knowledge of making HTTP requests in your programming language of choice.

Conceptually, authentication typically involves passing your API key in the headers of your requests.

Cognitive Actions Overview

Generate API Calls from Natural Language

The Generate API Calls from Natural Language action empowers developers to extend the capabilities of LLMs in formulating executable API calls based on natural language input. This action is categorized under API Management and significantly improves the accuracy of API requests, reducing the manual effort involved in crafting them.

Input

The input for this action is structured as a JSON object with the following schema:

{
  "prompt": "USER: <<question>> Call me an Uber ride type \"Plus\" in Berkeley at zipcode 94704 in 10 minutes <<function>> [{\"name\": \"Uber Carpool\", \"api_name\": \"uber.ride\", \"description\": \"Find suitable ride for customers given the location, type of ride, and the amount of time the customer is willing to wait as parameters\", \"parameters\": [{\"name\": \"loc\", \"description\": \"Location of the starting place of the Uber ride\"}, {\"name\": \"type\", \"enum\": [\"plus\", \"comfort\", \"black\"], \"description\": \"Types of Uber ride user is ordering\"}, {\"name\": \"time\", \"description\": \"The amount of time in minutes the customer is willing to wait\"}]}]\nASSISTANT: ",
  "maxTokens": 128,
  "temperature": 0.8,
  "topK": -1,
  "topP": 0.95,
  "presencePenalty": 0,
  "frequencyPenalty": 0
}
  • Required Field:
    • prompt: The natural language instruction that includes the necessary context for generating an API call.
  • Optional Fields:
    • maxTokens: Limits the number of tokens generated in the output (default is 128).
    • temperature: Adjusts randomness in token selection (default is 0.8).
    • topK: Specifies how many of the top scoring tokens to consider (default is -1, meaning all).
    • topP: Controls the cumulative probability for token selection (default is 0.95).
    • presencePenalty: Applies a penalty based on token presence (default is 0).
    • frequencyPenalty: Penalizes tokens based on their frequency (default is 0).

Example Input

Here’s an example of a structured input payload:

{
  "topK": -1,
  "topP": 0.95,
  "prompt": "USER: <<question>> Call me an Uber ride type \"Plus\" in Berkeley at zipcode 94704 in 10 minutes <<function>> [{\"name\": \"Uber Carpool\", \"api_name\": \"uber.ride\", \"description\": \"Find suitable ride for customers given the location, type of ride, and the amount of time the customer is willing to wait as parameters\", \"parameters\": [{\"name\": \"loc\", \"description\": \"Location of the starting place of the Uber ride\"}, {\"name\": \"type\", \"enum\": [\"plus\", \"comfort\", \"black\"], \"description\": \"Types of Uber ride user is ordering\"}, {\"name\": \"time\", \"description\": \"The amount of time in minutes the customer is willing to wait\"}]}]\nASSISTANT: ",
  "maxTokens": 128,
  "temperature": 0.8,
  "presencePenalty": 0,
  "frequencyPenalty": 0
}

Output

The output will be a generated API call based on the provided prompt. For instance, a successful execution may return:

uber.ride(loc="94704", type="plus", time=10)

Conceptual Usage Example (Python)

Here’s a conceptual Python code snippet to demonstrate how you might invoke this action:

import requests
import json

# Replace with your Cognitive Actions API key and endpoint
COGNITIVE_ACTIONS_API_KEY = "YOUR_COGNITIVE_ACTIONS_API_KEY"
COGNITIVE_ACTIONS_EXECUTE_URL = "https://api.cognitiveactions.com/actions/execute" # Hypothetical endpoint

action_id = "357923a8-731a-47d0-8617-eb9063bec22c"  # Action ID for Generate API Calls from Natural Language

# Construct the input payload based on the action's requirements
payload = {
    "topK": -1,
    "topP": 0.95,
    "prompt": "USER: <<question>> Call me an Uber ride type \"Plus\" in Berkeley at zipcode 94704 in 10 minutes <<function>> [{\"name\": \"Uber Carpool\", \"api_name\": \"uber.ride\", \"description\": \"Find suitable ride for customers given the location, type of ride, and the amount of time the customer is willing to wait as parameters\", \"parameters\": [{\"name\": \"loc\", \"description\": \"Location of the starting place of the Uber ride\"}, {\"name\": \"type\", \"enum\": [\"plus\", \"comfort\", \"black\"], \"description\": \"Types of Uber ride user is ordering\"}, {\"name\": \"time\", \"description\": \"The amount of time in minutes the customer is willing to wait\"}]}]\nASSISTANT: ",
    "maxTokens": 128,
    "temperature": 0.8,
    "presencePenalty": 0,
    "frequencyPenalty": 0
}

headers = {
    "Authorization": f"Bearer {COGNITIVE_ACTIONS_API_KEY}",
    "Content-Type": "application/json"
}

try:
    response = requests.post(
        COGNITIVE_ACTIONS_EXECUTE_URL,
        headers=headers,
        json={"action_id": action_id, "inputs": payload}  # Hypothetical structure
    )
    response.raise_for_status()  # Raise an exception for bad status codes (4xx or 5xx)

    result = response.json()
    print("Action executed successfully:")
    print(json.dumps(result, indent=2))

except requests.exceptions.RequestException as e:
    print(f"Error executing action {action_id}: {e}")
    if e.response is not None:
        print(f"Response status: {e.response.status_code}")
        try:
            print(f"Response body: {e.response.json()}")
        except json.JSONDecodeError:
            print(f"Response body: {e.response.text}")

In this Python snippet, replace the placeholder API key and endpoint with your actual values. The payload structure aligns with the input requirements for the action.

Conclusion

The Generate API Calls from Natural Language action provides a powerful way to simplify API integration by translating user instructions into executable calls. By leveraging this Cognitive Action, developers can save time and reduce errors in API request formulation. Explore further use cases and consider integrating this action into your applications to enhance their interaction capabilities with APIs.

Happy coding!

← Back to All Articles