Generating Text with Amazon Bedrock

In this module, we’ll explore how to use Amazon Bedrock Converse API to create conversational applications for:

  • ️ Summarization
  • Sentiment Analysis
  • Question and Answering (Q&A)

The Converse API provides a consistent API that works with all Amazon Bedrock models that support messages. This means you can write code once and use it with different models. Should a model have unique inference parameters, the Converse API also allows you to pass those unique parameters in a model-specific structure.

We’ll learn how to invoke various foundation models to perform these tasks. Be sure to leverage Amazon Q Developer to help you with the exercises.

️ Getting Started

To start, we will be updating a script to run the text generation examples. Inside your Visual Studio Code IDE, open the file:
text_examples/gen_text.py

genimage

Here we have code to invoke the different models, but we are missing logic to make the code functional. This is where we can leverage Amazon Q Developer to turn our ideas into workable code.

Currently, the code for text summarization is already filled out. The exercises will guide you to complete the sections for sentiment analysis and Q&A.

You can choose any of the models listed on line 58 to help you complete the exercises.

The complete reference code is available at:
full_code/gen_text.py

Exercise 1: Sentiment Analysis with Amazon Bedrock

Sentiment analysis is the tech version of “reading the room.” It gauges the mood or opinion embedded in a piece of text. Businesses use it to understand customer feedback, market trends, and social media conversations—turning qualitative data into actionable insights.

Right now, the function sentiment_analysis (line 93) is waiting for your input. You must fill out the function with a call to a model and a prompt to perform sentiment analysis.

You can view the summarize_text function (line 71) as an example.

Run the script using this command in the terminal:

python3 text_examples/gen_text.py

genimage

The text you’ll be working with is on line 113.

Exercise 2: Q&A with Amazon Bedrock

Question and Answer (Q&A) systems are like your own personal research assistant. They help you extract specific answers from a sea of information. This is incredibly useful in scenarios ranging from customer service bots to extracting insights from large datasets.

You must fill out the perform_qa function (line 103) with a call to a model and create a prompt that takes in the question and the source text.

Execute your script with by entering the following code in the Terminal and pressing enter.

python3 text_examples/gen_text.py

genimage

Wrap Up Now that you have gotten a taste of using Amazon Bedrock for processing text, let’s explore how we can use it for video tasks next!