Using Prompt Templates and Parameters

tip

Having problems? Don't worry, reach out on discord and we will help you out.

In this part of the tutorial series, we'll explore how to use prompt templates and parameters with llm-chain. Prompt templates allow you to create dynamic prompts, and parameters are the text strings you put into your templates.

Here's a simple Rust program demonstrating how to use prompt templates and parameters:

use llm_chain::{executor, parameters, prompt, step::Step};

#[tokio::main(flavor = "current_thread")]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Create a new ChatGPT executor
    let exec = executor!()?;
    // Create our step containing our prompt template
    let step = Step::for_prompt_template(prompt!(
        "You are a bot for making personalized greetings",
        "Make a personalized greeting tweet for {{text}}" // Text is the default parameter name, but you can use whatever you want
    ));

    // A greeting for emil!
    let res = step.run(&parameters!("Emil"), &exec).await?;
    println!("{}", res);

    // A greeting for you
    let res = step.run(&parameters!("Your Name Here"), &exec).await?;

    println!("{}", res.to_immediate().await?.as_content());

    Ok(())
}

Let's break down the different parts of the code:

1. We start with importing the necessary libraries, including the traits and structs required for our program.
2. The main async function is defined, using Tokio as the runtime.
3. We create a new Executor with the default settings.
4. A Step is created containing our prompt template with a placeholder ({{text}}) that will be replaced with a specific value later.
5. We create a Parameters object with the value "Emil" to replace the placeholder in the prompt template.
6. We execute the Step with the provided parameters and store the result in res, then print the response to the console.
7. We create another Parameters object, this time with the value "Your Name Here" to replace the placeholder.
8. We execute the Step again with the new parameters, store the result in res, and print the response to the console.

In the next tutorial, we will combine multiple LLM invocations to solve more complicated problems.