### Start Client using Setup Configuration

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/text-agent-to-strands-voice-agent/references/deployment-reference.md

Helper script to start the client using the saved setup configuration after deployment. Assumes the configuration is saved in strands/setup_config.json.

```bash
./deployment/start_client.sh strands
```

--------------------------------

### Install React Dependencies and Start App

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/workshops/livekit/README.md

Install the necessary Node.js dependencies for the React application and start the development server.

```bash
npm install
npm start
```

--------------------------------

### Start UI Client

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/sample-codes/websocket-dotnet/README.md

Navigate to the UI directory, install dependencies, and start the development server for the web client. This command prepares and launches the user interface.

```bash
cd ui-stream
npm install
npm run dev
```

--------------------------------

### Setup and Client Initialization

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/repeatable-patterns/07-function-call-with-structured-json/function_call.ipynb

Installs necessary libraries, imports modules, and initializes AWS Bedrock clients. Ensure API_ENDPOINT is set to your specific endpoint.

```python
!pip3 install -q aws-requests-auth

from aws_requests_auth.aws_auth import AWSRequestsAuth
import boto3
import datetime
import json
import os
import requests
import urllib

API_ENDPOINT="<API_ENDPOINT from cdk>"
AWS_REGION="us-east-1"
BEDROCK_MODEL="us.amazon.nova-pro-v1:0"
BEDROCK_TOOL_MODEL="auto"
# switch to "any" if you want to force the model to "only" use the tools you provide 

aws_session_credentials = boto3.Session().get_credentials()
aws_client_bedrock = boto3.client(
    service_name="bedrock-runtime"
)
```

--------------------------------

### Setup Environment and AWS Clients

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/customization/bedrock-distillation/distillation_recipes/01_citations/03_batch_inference.ipynb

Installs and upgrades necessary libraries like boto3. Restarts the kernel to apply changes. Loads model IDs from previous notebooks.

```python
# upgrade boto3 
%pip install --upgrade pip --quiet
%pip install boto3 --upgrade --quiet

# restart kernel
from IPython.core.display import HTML
HTML("<script>Jupyter.notebook.kernel.restart()</script>")
```

```python
# load PT model id from previous notebook
%store -r provisioned_model_id
%store -r custom_model_id
```

```python
import json
import sys
import os
import time

current_dir = os.getcwd()
parent_dir = os.path.dirname(current_dir)
skip_dir = os.path.dirname(parent_dir)
sys.path.append(skip_dir)

import boto3
from datetime import datetime
from botocore.exceptions import ClientError
from utils import create_s3_bucket

# Create Bedrock client
bedrock_client = boto3.client(service_name="bedrock", region_name='us-east-1')

# Create runtime client for inference
bedrock_runtime = boto3.client(service_name='bedrock-runtime', region_name='us-east-1')

# Region and accountID
session = boto3.session.Session(region_name='us-east-1')
region = 'us-east-1'
sts_client = session.client(service_name='sts', region_name='us-east-1')
account_id = sts_client.get_caller_identity()['Account']

# Define bucket and prefixes (using the same bucket as in distillation)
BUCKET_NAME = '<BUCKET_NAME>' # Same bucket used in distillation notebook
DATA_PREFIX = 'citations_distillation'  # Same prefix used in distillation notebook
batch_inference_prefix = f"{DATA_PREFIX}/batch_inference"  # New prefix for batch inference
```

--------------------------------

### Install Dependencies

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/customization/Nova_2.0/workshop/EndToEnd_SFT_RFT_Workshop/SFT_RFT/01_Data_prep.ipynb

Installs necessary packages from a requirements file. Use this at the beginning of your environment setup.

```python
!pip install -r requirements.txt --quiet
```

--------------------------------

### Setup Python Virtual Environment and Install Dependencies

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/workshops/livekit/README.md

Creates a Python virtual environment, activates it, and installs project dependencies from 'requirements.txt'. This ensures a clean and isolated environment for the LiveKit agent.

```bash
cd nova-s2s-workshop/livekit
python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt
```

--------------------------------

### Python CLI Client: Installation and Execution

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/text-agent-to-strands-voice-agent/references/client-reference.md

Provides commands for installing necessary Python packages (`pyaudio`, `websockets`) and examples for running the client against a local server or a deployed AgentCore Runtime using a presigned URL.

```bash
pip install pyaudio websockets

# Local server
python client.py --ws-url ws://localhost:8080/ws

# Deployed via AgentCore Runtime (presigned URL)
python client.py --runtime-arn arn:aws:bedrock-agentcore:us-east-1:123456:runtime/RTID
```

--------------------------------

### Install AgentCore Starter Toolkit

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/text-agent-to-strands-voice-agent/references/deployment-reference.md

Installs the necessary toolkit for deploying the AgentCore voice agent. Ensure Python 3 and AWS CLI are installed.

```bash
pip install bedrock-agentcore-starter-toolkit
```

--------------------------------

### Install PyAudio on Windows (using pipwin)

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/repeatable-patterns/token-usage-logger/README.md

Install PyAudio on Windows by first installing pipwin and then the PyAudio binary.

```bash
# Install PyAudio binary directly using pip
pip install pipwin
pipwin install pyaudio
```

--------------------------------

### Run Automated Setup Script

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/tutorials/path-to-production/00-setup/0_setup_environment.md

Execute the provided shell script to automate the creation of a virtual environment, its activation, and the installation of all required packages from requirements.txt. Ensure you run this from the 'path-to-production' directory.

```bash
chmod +x 00-setup/0_setup_environment.sh && ./00-setup/0_setup_environment.sh
```

--------------------------------

### Install gemini-to-nova-migration Plugin

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/gemini-to-nova-migration/README.md

Add the marketplace and install the plugin to begin migrating Gemini code. After installation, invoke the skill using the provided command.

```bash
/plugin marketplace add https://github.com/aws-samples/amazon-nova-samples
/plugin install gemini-to-nova-migration@aws-samples-amazon-nova-samples
```

--------------------------------

### Run the HyperPod CLI Installer

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/customization/nova-forge-hyperpod-cli-installation/INSTALL_USAGE.md

Execute the installation script for the HyperPod CLI. You will be prompted for an installation directory.

```bash
bash install_hp_cli.sh
```

--------------------------------

### Clone Repository and Navigate

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/repeatable-patterns/bedrock-knowledge-base/README.md

Clone the repository and navigate into the project directory to begin setup.

```bash
git clone <repository-url>
cd <repository-name>
```

--------------------------------

### Set up Environment File

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/customization/bedrock-distillation/distillation_recipes/02_function_calling/01_prepare_data.ipynb

Copies an example .env file to the project root and prompts the user to fill in necessary AWS credentials. Uncomment to use.

```python
# set up environment file
# %cp $(python -c "import bfcl_eval; print(bfcl_eval.__path__[0])")/.env.example $BFCL_PROJECT_ROOT/.env
# Fill in necessary values in `.env`
```

--------------------------------

### Install Required Packages

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/repeatable-patterns/28-email-assistant-with-strands-agents/email_assistant_strands_agents.ipynb

Installs necessary packages from a requirements file. Use this at the beginning of your project setup.

```python
!pip install -r requirements.txt -q --no-cache-dir
```

--------------------------------

### Video Timestamps - Event with Example Format

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/nova-prompter/plugins/nova-prompting/skills/nova2-prompt/SKILL.md

Asks for the start and end timestamps of an event, providing an example format.

```text
When does "{event_description}" in the video? Specify the start and end timestamps, e.g. [[9, 14]]
```

--------------------------------

### Voice Agent Setup and Run Instructions

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/text-agent-to-strands-voice-agent/SKILL.md

Provides instructions for setting up the voice agent environment, including system dependencies, Python virtual environment, and running the server. Also details how to connect the client to the WebSocket server.

```markdown
# [Agent Name] Voice Agent

Real-time voice agent using Strands BidiAgent with Amazon Nova Sonic.

## Setup

### System Dependencies

macOS:
```bash
brew install portaudio
```

Ubuntu/Debian:
```bash
sudo apt-get install -y portaudio19-dev gcc g++ make
```

### Python Environment

```bash
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
```

### Environment Variables

```bash
# Set any required tool/service ARNs or API keys
export AWS_DEFAULT_REGION=us-east-1
```

### Run

```bash
source .venv/bin/activate
python server.py
```

The WebSocket server starts on port 8080 by default. To use a different port:

```bash
python server.py --port 9000
# or
PORT=9000 python server.py
```

### Connect the Client

Open a second terminal:

```bash
source .venv/bin/activate
python client.py --ws-url ws://localhost:8080/ws
```

This opens a browser with mic/speaker controls. If you changed the server port, update the URL accordingly:

```bash
python client.py --ws-url ws://localhost:9000/ws
```
```

--------------------------------

### Clone Repository and Navigate

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/workshops/livekit/README.md

Clone the Amazon Nova Samples repository and set up the workshop directory. This command is for initial setup and assumes you haven't cloned the repository before.

```bash
git clone https://github.com/aws-samples/amazon-nova-samples
mv amazon-nova-samples/speech-to-speech/workshops nova-s2s-workshop
rm -rf amazon-nova-samples
cd nova-s2s-workshop/livekit
```

--------------------------------

### Install Required Libraries

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/repeatable-patterns/14-document-longcontext-idp/document_longcontext_idp.ipynb

Installs necessary AWS SDK and request libraries. Use this at the beginning of your environment setup.

```python
%pip install "boto3>=1.28.57" "awscli>=1.29.57" "botocore>=1.31.57" "requests>=2.32.3" --quiet
```

--------------------------------

### Install Hugging Face Hub

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/repeatable-patterns/02-key-info-extraction-to-sql/key_info_extraction_to_sql.ipynb

Installs the Hugging Face Hub library, which is required for downloading datasets. This is a one-time setup.

```python
## TO DO: uncomment this if never run this before
#! pip install huggingface_hub --quiet
```

--------------------------------

### Example RAG Question

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/workshops/README.md

An example of a question that can be asked when using the RAG integration with Amazon Nova User Guide data.

```text
What is Amazon Nova Sonic?
```

--------------------------------

### Start Sample Application Server

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/usecases/adaptive-ui-testing-with-nova-act/README.md

Change to the sample-app directory and start a Python HTTP server on port 8000. This application should remain running while you conduct the demo.

```bash
cd sample-app && python3 -m http.server 8000
```

--------------------------------

### Install Required Libraries

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/repeatable-patterns/08-lite-as-subagent-to-pro/lite_as_subagent_to_pro.ipynb

Installs necessary Python packages for the notebook, including IPython, PyMuPDF, and matplotlib. Use this at the beginning of your environment setup.

```python
%pip install IPython PyMuPDF matplotlib --quiet
```

--------------------------------

### Nova 1 Few-Shot Examples: System Prompt

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/nova-prompter/plugins/nova-prompting/skills/nova1-prompt/SKILL.md

For long or complex examples, place them within the system prompt using labeled blocks like `<Example 1>` to guide the model's response.

```text
Below are a few examples of well-formatted {artifact} to guide your response.

<Example 1>
{full example 1}

<Example 2>
{full example 2}
```

--------------------------------

### Install Dependencies and Verify Tooling

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/sample-codes/agentcore/README.md

Set up a Python virtual environment, install the required deployment dependencies, and verify that the `agentcore` and `deploy.py` tools are accessible.

```bash
# Create and activate virtual environment
python3 -m venv venv
source venv/bin/activate

# Install deployment tooling
pip install -r deployment/requirements.txt

# Verify
agentcore --help
python deployment/deploy.py --help
```

--------------------------------

### Install Dependencies

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/usecases/price_comparison/README.md

Navigate to the project directory, create and activate a virtual environment, and then install the required Python dependencies from the requirements.txt file.

```bash
# Navigate to the project directory
cd nova-act/usecases/price_comparison

# Create and activate virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt
```

--------------------------------

### Install Required Python Packages

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-embeddings/repeatable-patterns/multilingual-search/multilingual-utterance-search.ipynb

Installs essential Python libraries for AWS interaction, data manipulation, and similarity calculations. Use this at the beginning of your environment setup.

```python
# Install required packages
!pip install boto3 pandas numpy scikit-learn matplotlib seaborn tqdm langdetect
```

--------------------------------

### Copy Example Test Case IDs

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/customization/bedrock-distillation/distillation_recipes/02_function_calling/01_prepare_data.ipynb

Copies an example JSON file for test case IDs to the project root. This is a setup step for data preparation.

```shell
%cp $(python -c "import bfcl_eval, pathlib; print(pathlib.Path(bfcl_eval.__path__[0]) / 'test_case_ids_to_generate.json.example')") $BFCL_PROJECT_ROOT/test_case_ids_to_generate.json
```

--------------------------------

### Start UI Development Server

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/sample-codes/websocket-java/README.md

Start the UI development server to view the application in your browser.

```bash
cd ui-stream
npm run dev
```

--------------------------------

### Clone Repository

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/sample-codes/websocket-dotnet/README.md

Clone the repository to get the .NET WebSocket example code.

```bash
git clone <repository-url>
cd amazon-nova-samples/speech-to-speech/sample-codes/websocket-dotnet
```

--------------------------------

### Running the AgentCore Gateway Tutorial Script

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/tutorials/path-to-production/02-tool-use/2_agentcore_gateway.md

Provides the command to navigate to the tutorial directory and execute the Python script for setting up and running the AgentCore Gateway and workflows.

```bash
cd tutorials/path-to-production/03-tool-use
python 2_agentcore_gateway.py
```

--------------------------------

### Clone Repository and Install Dependencies

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/usecases/adaptive-ui-testing-with-nova-act/README.md

Clone the adaptive-ui-testing-with-nova-act repository and run the setup script to install Playwright, Chromium, a Python virtual environment with the Nova Act SDK, and all other dependencies.

```bash
git clone https://github.com/aws-samples/adaptive-ui-testing-with-nova-act
cd adaptive-ui-testing-with-nova-act
./setup.sh
```

--------------------------------

### Clone Repository and Install Dependencies

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/usecases/qa-testing/tests/README.md

Clone the repository, navigate to the tests directory, create and activate a virtual environment, and install project dependencies.

```bash
git clone <repository-url>
cd tests

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

pip install .
```

--------------------------------

### Start Websocket Server for Strands Agent

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/workshops/README.md

Starts the Python websocket server with the '--agent strands' flag, enabling the Strands Agent integration. Ensure uv is installed for managing Python dependencies.

```python
python server.py --agent strands
```

--------------------------------

### Setup Environment and Imports

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/customization/bedrock-distillation/distillation_recipes/01_citations/04_evaluate.ipynb

Imports necessary libraries and sets up the environment by adding a parent directory to the system path for utility functions.

```python
import boto3
import json
import os
import sys
from datetime import datetime
import re

print(boto3.__version__)
current_dir = os.getcwd()
parent_dir = os.path.dirname(current_dir)
skip_dir = os.path.dirname(parent_dir)
sys.path.append(skip_dir)
from utils import read_jsonl_to_dataframe, upload_training_data_to_s3
```

--------------------------------

### Set Up Virtual Environment

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/usecases/adaptive-ui-testing-with-nova-act/README.md

Commands to remove an existing virtual environment, create a new one, and install dependencies.

```bash
cd nova-act-tests
rm -rf venv
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
```

--------------------------------

### Python CLI Client: Audio Capture and Playback Setup

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/text-agent-to-strands-voice-agent/references/client-reference.md

Initializes PyAudio for microphone input and speaker output, setting up audio streams for real-time capture and playback. Requires `pyaudio` library.

```python
        pa = pyaudio.PyAudio()
        mic = pa.open(format=pyaudio.paInt16, channels=CHANNELS,
                      rate=RATE, input=True, frames_per_buffer=CHUNK)
        spk = pa.open(format=pyaudio.paInt16, channels=CHANNELS,
                      rate=RATE, output=True, frames_per_buffer=CHUNK)
```

--------------------------------

### Example Prompts for Gemini to Nova Migration

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/gemini-to-nova-migration/README.md

Use these example prompts to guide the migration of your Gemini code to Amazon Nova 2 Lite, covering standard code conversion, function-calling applications, and multimodal prompts.

```plaintext
Migrate this Gemini code to Amazon Nova 2 Lite:
<paste your google-genai code>
```

```plaintext
I have a Gemini 2.5 Flash function-calling app with thinking enabled.
Convert it to Nova 2 Lite.
```

```plaintext
Rewrite this Gemini multimodal prompt (image input) for Nova 2 Lite.
```

--------------------------------

### Run Setup Script

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/tutorials/path-to-production/00-setup/1_setup_aws_credentials.md

Command to navigate to the setup directory and execute the Python script for configuring AWS credentials.

```bash
cd /path/to/nova-act/tutorials/path-to-production/00-setup
python setup_aws_credentials.py
```

--------------------------------

### Required Python Packages

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/repeatable-patterns/session-continuation/console-python/README.md

List of Python packages required for the session continuation example. Ensure these are installed in your virtual environment.

```bash
pyaudio>=0.2.13
rx>=3.2.0
smithy-aws-core>=0.0.1
pytz
aws_sdk_bedrock_runtime>=0.1.0,<0.2.0
```

--------------------------------

### Navigate to Project Directory

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-embeddings/getting-started/README.md

Navigate to the root directory of the project. This is a prerequisite for subsequent setup steps.

```bash
cd path/to/amazon-nova-embeddings-beta
```

--------------------------------

### Run Streamlit Application

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/sample-apps/01-multimodal-with-helper-libraries/README.md

Execute the Streamlit application after installing all requirements. This command starts the interactive web interface for the multi-modal understanding demo.

```bash
streamlit run mm_understanding.py
```

--------------------------------

### Start Application Server

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/repeatable-patterns/bedrock-knowledge-base/README.md

Start the Node.js server to run the application.

```bash
npm start
```

--------------------------------

### Top-Down Approach Template

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/nova-prompter/powers/nova2-prompt/POWER.md

Employ this template for complex problems requiring a top-down breakdown. It guides the model to start with the big picture and progressively detail subproblems.

```plaintext
{User query}. Start with the big picture and break it down into progressively smaller, more detailed subproblems or steps.
```

--------------------------------

### Initialize AWS Session and Clients

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/repeatable-patterns/26-nova-citations/evaluate_with_create_evaluation_job.ipynb

Sets up the Boto3 session, SageMaker session, and Bedrock client. Ensure you have the correct AWS credentials and permissions configured.

```python
session = boto3.Session()
sagemaker_session = sagemaker.Session()
bedrock_client = session.client('bedrock')
```

--------------------------------

### Video Timestamps - Event Localization (Seconds, Explicit Format)

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/skills/nova-prompter/plugins/nova-prompting/skills/nova2-prompt/SKILL.md

Localizes an event in seconds with an explicit example format for start and end times, supporting multiple occurrences.

```text
Please localize the moment that the event "{event_description}" happens in the video. Answer with the starting and ending time of the event in seconds. e.g. [[72, 82]]. If the event happen multiple times, list all of them. e.g. [[40, 50], [72, 82]]
```

--------------------------------

### Setup API Keys and Environment

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/repeatable-patterns/27-multi-agent-orchastration-with-strands/multi_agent_strands.ipynb

Imports and initializes AWS SDK clients and Strands components. Configure AWS region and Bedrock model for LLM interactions.

```python
import os
import sys
import boto3
import sagemaker
import json
import requests
import time
import datetime
import uuid
from typing import List, Dict, Any, Optional
import matplotlib.pyplot as plt
import networkx as nx

from strands import Agent, tool
from strands.multiagent import GraphBuilder
from strands.models import BedrockModel
from strands_tools import file_read, file_write, editor, current_time

from IPython.display import display, Image
import logging
import re
import glob

sts_client = boto3.client('sts')
account_id = sts_client.get_caller_identity()["Account"]
session = sagemaker.Session()

aws_region = boto3.session.Session().region_name

bedrock_model = BedrockModel(
    model_id="us.amazon.nova-lite-v1:0",
    region_name=aws_region
)


# Function to extract filename from response
def clean_extract_filename(response_string):
    if not isinstance(response_string, str):
        response_string = str(response_string)
    pattern = r'<filename>\s*(.*?)\s*</filename>'
    matches = re.findall(pattern, response_string)
    if matches:
        # Clean the filename by removing any newlines or extra whitespace
        return matches[-1].strip()
    return None

def display_image(image_path):
    import matplotlib.pyplot as plt
    from PIL import Image
    
    img = Image.open(image_path)
    plt.imshow(img)
    plt.axis('off')
    plt.show()
    
def extract_filename(response_string):
    import re
    
    # Convert to string if not already a string
    if not isinstance(response_string, str):
        response_string = str(response_string)
        
    pattern = r'<filename>\s*(.*?)\s*</filename>'
    matches = re.findall(pattern, response_string)
    return matches[-1] if matches else None


logging.getLogger('botocore.credentials').setLevel(logging.ERROR)
logging.getLogger('botocore').setLevel(logging.ERROR)
logging.getLogger('boto3').setLevel(logging.ERROR)
```

--------------------------------

### Create and Configure .env File

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/usecases/qa-testing/tests/README.md

Copy the sample environment file and update it with your CloudFront Distribution URL and Nova Act API Key.

```bash
cp .env.sample .env
```

--------------------------------

### Get BDA Audio Task Invocation ARN

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/repeatable-patterns/16-multimodal-agentic-workflow/audio-video-rag/01_data_prep_using_bda.ipynb

Retrieves the invocation ARN for a started BDA audio task. This ARN is used to monitor the task's progress.

```python
invocation_audio_arn = response_aud.get("invocationArn")
print("BDA audio task started:", invocation_audio_arn)
```

--------------------------------

### Using System Prompts with Converse API

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/workshop/1-getting-started-with-nova-2/1-getting-started-with-api.ipynb

Pass system prompts via the `system` parameter to guide the model's behavior. This example shows how to act as a telecom marketer.

```python
response = client.converse(
    modelId=MODEL_ID,
    system=[
        {
            "text": "Act as a telecom industry marketer for AnyCompany Telecom. When the user provides you with a topic, write a LinkedIn Launch Post about that topic."
        }
    ],
    messages=[
        {
            "role": "user",
            "content": [{"text": "Meet our AI-powered customer support assistant"}]
        }
    ],
    inferenceConfig={"maxTokens": 1024, "temperature": 0.7}
)

output_text = response["output"]["message"]["content"][0]["text"]
display(Markdown(output_text))
```

--------------------------------

### Custom Retailer Sources Example

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/nova-act/usecases/price_comparison/README.md

This example demonstrates how to specify a custom set of retailers and their corresponding website URLs for price comparison.

```bash
python main.py --sources '[("Walmart", "https://www.walmart.com"), ("Newegg", "https://www.newegg.com")]'
```

--------------------------------

### Set up Python Virtual Environment

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/sample-apps/01-multimodal-with-helper-libraries/README.md

Create and activate a Python virtual environment for the project. Install project dependencies using pip.

```bash
python3 -m venv env
source env/bin/activate
pip install -r requirements.txt
```

--------------------------------

### Create a Simple Langchain Agent

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/getting-started/06_langchain_integration.ipynb

This code defines a basic Langchain agent that can invoke a simple function to get weather information. It's a straightforward example for understanding agent invocation with a single tool.

```python
from langchain.agents import create_agent
import json
def get_weather(city: str) -> str:
    """Get weather for a given city."""
    return f"Test Response: It's always sunny in {city}!"

MODEL_ID = "amazon.nova-2-lite-v1:0"
agent = create_agent(
    model=MODEL_ID,
    tools=[get_weather],
    system_prompt="You are a helpful assistant",
)

# Run the agent
response = agent.invoke(
    {"messages": [{"role": "user", "content": "what is the weather in sf"}]}
)

response['messages'][3].content
```

--------------------------------

### Few-Shot Prompting for Classification

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/multimodal-understanding/workshop/1-getting-started-with-nova-2/2-general-best-practices.ipynb

Use few-shot prompting to guide the model's reasoning and output format for specific tasks like classification. Provide clear examples of input, desired output, and reasoning.

```python
few_shot_prompt = ""
##Task##
Classify the following customer support ticket into one of these categories:
- Billing
- Technical
- Account
- General Inquiry

##Examples##
Example 1:
Ticket: "My internet speed is much slower than what I'm paying for."
Category: Technical
Reason: Customer reports a service performance issue related to network speed.
---
Example 2:
Ticket: "I want to change the primary name on my account to my spouse."
Category: Account
Reason: Customer requests a modification to account holder information.
---
Example 3:
Ticket: "Can you explain what the 'network access fee' on my bill means?"
Category: Billing
Reason: Customer is asking about a specific charge on their invoice.

DO NOT mention examples in your response.

##Input##
Ticket: "I was charged twice for my November bill and I need a refund for the duplicate payment."

##Output Requirements##
You MUST respond with only the category and a one-sentence explanation in the following format:

Format: [Category]: [One-sentence explanation]

DO NOT include any additional text, commentary, or formatting beyond what is specified above.
"

response = client.converse(
    modelId=MODEL_ID,
    messages=[{
        "role": "user",
        "content": [{"text": few_shot_prompt}]
    }],
    inferenceConfig={"maxTokens": 128, "temperature": 0}
)

output_text = response["output"]["message"]["content"][0]["text"]
display(Markdown(output_text))
```

--------------------------------

### Full Example of Launcher Execution

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/customization/forge-assets/launcher/LAUNCHER_README.md

A comprehensive example demonstrating the execution of the hyperpod_launcher.py script with various parameters, including data sources, output locations, MLflow integration, container image, instance type, and namespace.

```bash
python3 hyperpod_launcher.py \
    --recipe micro_pt_recipe.yaml \
    --name dewanup-cpt-nodm-ce-chkpt \
    --data s3://618100645563-nova-customization-beta/dewanup/data/cpt/train.jsonl \
    --val s3://618100645563-nova-customization-beta/dewanup/data/cpt/val.jsonl \
    --output s3://618100645563-nova-customization-beta/dewanup/output/ \
    --mlflow arn:aws:sagemaker:us-east-1:618100645563:mlflow-app/app-QYBQZINLZDLY \
    --container 708977205387.dkr.ecr.us-east-1.amazonaws.com/nova-fine-tune-repo:SM-HP-CPT-V2-BETA-latest \
    --instance-type ml.p5.48xlarge \
    --namespace kubeflow
```

--------------------------------

### Install AWS CLI V2

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/customization/hyperpod-rig-cluster-setup/Hyperpod_Nova_Cluster_and_Dependencies_setup.ipynb

Installs AWS CLI V2, removing any existing v1 installations. The installation process downloads the zip archive, extracts it, and installs the CLI to /usr/local/bin and /usr/local/aws-cli.

```bash
set -euo pipefail

# Remove v1
sudo rm -rf /opt/conda/bin/aws
rm -rf /home/sagemaker-user/nova_sdk/bin/aws

curl -fsSL "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o /tmp/awscliv2.zip
unzip -q /tmp/awscliv2.zip -d /tmp
sudo /tmp/aws/install --bin-dir /usr/local/bin --install-dir /usr/local/aws-cli --update

rm -rf /tmp/awscliv2.zip /tmp/aws
```

--------------------------------

### Clone Repository and Setup Workshop Directory

Source: https://github.com/aws-samples/amazon-nova-samples/blob/main/speech-to-speech/amazon-nova-2-sonic/workshops/agent-core/README.md

Clone the Amazon Nova Samples repository and set up the workshop directory. This is a prerequisite for deploying the agents.

```bash
git clone https://github.com/aws-samples/amazon-nova-samples
mv amazon-nova-samples/speech-to-speech/workshops nova-s2s-workshop
rm -rf amazon-nova-samples
```