Hello guys, in this blog we will learn how can we create a simple chat application using AWS Bedrock and the AWS Boto3 python library.
We will be creating a simple script which will take the prompt provided by the user as an input and will provide with relavant output.
By following along with the blog you will also learn how you can make a basic API call directly to the AWS Bedrock service.
Prerequisite
In this example we will be using the AWS Bedrock service along with the models provided by the AI21 Labs So before getting started with the example you must request access for the AI21 Labs models.
Also since the bedrock service is not available in all the AWS regions we will be using the us-east-1 region for this example.
Action Items
Access for the AI21 Labs models in the us-east-1 region.
Generate and configure access keys to execute the below python script. You will need the Bedrock service access.
Lets get the python script ready
- First install the AWS Boto3 library
pip3 install boto3
First we will add the import statement.
These statements allow us to use the AWS Boto3 library to call Amazon Bedrock.
import json
import boto3
- Initialize the Bedrock client library.
session = boto3.Session()
bedrock = session.client(service_name='bedrock-runtime')
Build the payload for the API call.
Here we are identifying the model to use, the prompt, and the inference parameters for the specified model.
bedrock_model_id = "ai21.j2-ultra-v1"
prompt = "Which is the largest city in India?"
body = json.dumps({
"prompt": prompt,
"maxTokens": 1024,
"temperature": 0,
"topP": 0.5,
"stopSequences": [],
"countPenalty": {"scale": 0 },
"presencePenalty": {"scale": 0 },
"frequencyPenalty": {"scale": 0 }
})
Call the Bedrock API.
We use Bedrock's invoke_model function to make the call.
response = bedrock.invoke_model(body=body, modelId=bedrock_model_id, accept='application/json', contentType='application/json')
Display the response.
This extracts & prints the returned text from the model's response JSON.
response_body = json.loads(response.get('body').read())
response_text = response_body.get("completions")[0].get("data").get("text")
print(response_text)
Save the file. Great! Now you are ready to run the script!
Run the script using the below command
python3 bedrock_api.py
Finally Here is the code for complete script
import json
import boto3
session = boto3.Session()
bedrock = session.client(service_name='bedrock-runtime')
bedrock_model_id = "ai21.j2-ultra-v1"
prompt = "What is the largest city in New Hampshire?"
body = json.dumps({
"prompt": prompt,
"maxTokens": 1024,
"temperature": 0,
"topP": 0.5,
"stopSequences": [],
"countPenalty": {"scale": 0 },
"presencePenalty": {"scale": 0 },
"frequencyPenalty": {"scale": 0 }
})
response = bedrock.invoke_model(body=body, modelId=bedrock_model_id, accept='application/json', contentType='application/json')
response_body = json.loads(response.get('body').read())
response_text = response_body.get("completions")[0].get("data").get("text")
print(response_text)
Conclusion
In conclusion, creating a chat app using AWS Bedrock and the Boto3 library is a streamlined process that leverages the robust capabilities of AWS's AI and ML services.
By following the steps outlined in this blog, you can efficiently set up and deploy a chat application using Python that harnesses the power of GenAI for intelligent interactions. AWS Bedrock provides a flexible and scalable foundation, while Boto3 simplifies the integration and management of AWS services.
Whether you're a seasoned developer or new to cloud-based AI, this combination offers a powerful solution for building sophisticated chat applications. Explore further possibilities and continue innovating with AWS Bedrock, Boto3, Python, and GenAI to enhance your app's functionality and user experience.
