Amazon Bedrock Series

Overview

This project leverages AWS Bedrock Large Language Models (LLMs) for AI-driven applications. The setup process ensures a seamless environment for interaction with AWS services.

Initial Setup

To begin, use any IDE of your preference. This guide uses VSCode.

1. Create a Virtual Environment

A virtual environment is recommended to maintain dependencies separately.

Run the following commands in the terminal: For macOS/Linux users, activate the virtual environment using:

source venv/bin/activate 

For Windows users, activate the virtual environment using:

python -m venv venv
cd venv/Scripts
activate.ps1  # For PowerShell users

2. Creating a Git Repository

Create a new repository on GitHub or any other platform of your choice.

Errors I faced during repo creation due to previous git initializations

fatal: detected dubious ownership in repository at 'D:/'
'D:/' is owned by:
        'S-1-5-18'
but the current user is:
        'S-1-5-21-520853544-919835508-2353064755-1001'
To add an exception for this directory, call:

        git config --global --add safe.directory D:/

Solution By Git

git config --global --add safe.directory D:/

Initializing Git Repository

Reason:

I realized the issue was simply that I forgot to initialize Git. 😅

Steps Taken

1. Initialize Git in the project directory:

   git init
   

Here’s your updated documentation snippet with the Git initialization step:

Initializing Git Repository

Reason:

I realized the issue was simply that I forgot to initialize Git. 😅

Steps Taken

1. Initialize Git in the project directory:

git init 

2. Check for existing remote repositories:

git remote -v

3. Add the remote origin using your repository URL:

 git remote add origin <repo-url>

4. Verify that the remote origin is set correctly:

git remote -v

Another blunder:

I already had a main branch with gitignore file etc, in my terminal i accidently push it to a non-existing branch master, leading to its creation. now I dont want that.

I want to undo it

We can do that using git revert

Using git revert will create a new commit that undoes the changes made in the previous commit.

you just need to pass the old commit's hashcode to git revert command. you can get the hashcode using git log command. or previous push log eg: 4b0f231.

Now the commit it reverted, but what about the branch?

You need to delete the branch, cause it will create trouble in the future.

Just use git push origin --delete <branch-name> to delete the branch from remote repository.

Issue Still Exists

Once i have deleted it, I tried pushing it to the main branch but I got this fatal: refusing to merge unrelated histories

This happens when the local repository and remote repository have different histories, often when the repository was initialized separately. Due to creation of two branches and pushes to the new branch, the old one became outdated.

So we git fetch origin main to update the local repository with the latest changes from the remote repository. Then we can use git rebase origin/main to rebase our local branch onto the remote branch.

Then i got this Successfully rebased and updated refs/heads/main

The issue is now resolved. I can push my changes to the remote repository using cmd git push origin main

This worked for me!!!... I hope it works for you too!!!

3. Install Required Packages following essential packages are used:

• boto3 – AWS SDK for Python, enabling interaction with AWS services like S3 and EC2.
• awscli – AWS Command Line Interface, allowing users to interact with AWS via command-line.

Installing via requirements.txt It is best practice to install dependencies using a requirements.txt file.

Create the file

echo > requirements.txt 

Add the module names inside requirements.txt:

boto3
awscli

Now install the dependencies:

pip install -r requirements.txt

Verification After installation, confirm package availability

pip list | grep "boto3\|awscli"

4. Now Let's Configure AWS CLI:

Creating a User in IAM (AWS)

1. Navigate to IAM

• Sign in to the AWS Management Console.
• In the Services menu, go to IAM (Identity and Access Management).

2. Create a New User

• Click on Users in the left-hand menu.
• Select Add user.
• Enter the username as: testadmin.
• Under Access type, select AWS Management Console access (for console login) or Programmatic access (for API/CLI usage), depending on requirements.
• Click Next: Permissions.

3. Attach Administrative Permissions

• Under Set permissions, choose Attach existing policies directly.
• Search for and attach the policy AdministratorAccess.
• Click Next: Tags.

4. Add a Description Tag

• In the Tag key, enter a meaningful tag such as Description.
• Set the Tag value to: testuserkey.
• Click Next: Review.

5. Review and Create User

• Ensure the user details, permissions, and tags are correct.
• Click Create user.

6. Generate Access Keys

• Once the user is created, navigate to the user’s details page.
• Click Security credentials.
• Under Access keys, click Create access key.
• Select Use case based on your needs (e.g., Command Line Interface (CLI)).
• Click Create.
• Download the .csv file containing the access key ID and secret access key.

Things to Do:

• Download the .csv file containing the access key and secret key.
• Store it securely, either in your project folder or another safe location.

5. Configure AWS

To configure AWS credentials, run the following command in your VSCode terminal:

aws configure

Required Inputs:

You will be prompted to enter the following details:

1. AWS Access Key ID – Your unique Access Key ID.
2. AWS Secret Access Key – The corresponding Secret Access Key.
3. Default region name – Specify your AWS region (e.g., us-east-1, ap-south-1).
4. Default output format – Leave it blank or specify format (e.g., json, table, text), default format is json.

6. Running the Script

This experiment utilizes the Amazon Titan Text-to-Embedding model (amazon.titan-embed-text-v2:0) to explore how we can interact with LLM models via APIs using AWS services. By integrating Bedrock's capabilities into our workflow, we can seamlessly process text embeddings and incorporate them into various applications.

Through this implementation, we demonstrate how AI-powered solutions can be efficiently leveraged to enhance natural language understanding, improve search relevance, and streamline application integration with AWS Bedrock.

To execute the script, simply run:

python file_name.py