### POST getRecommendations Endpoint Example

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/streaming_events/README.md

Example body and endpoint for retrieving recommendations. Ensure you include the API key in the request header.

```json
{
    "userId":"12345"
    
}
```

```plaintext
https://XXXXXX.execute-api.us-east-1.amazonaws.com/dev2/recommendations
```

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

### Build the SAM Project

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/streaming_events/README.md

Build your Serverless Application Model (SAM) project. Refer to the AWS SAM CLI installation guide for details.

```bash
sam build
```

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

### Get recommendations with promotions

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/filters/promotions/Promotions.ipynb

Retrieve recommendations for a user, incorporating promotions to boost specific items. This example uses a promotion to include a percentage of items from the 'halloween' category.

```python
get_recommendations_response = personalize_runtime.get_recommendations(
    recommenderArn = recommended_for_you_arn,
    userId = test_user_id,
    numResults = 20,
    promotions = [{
        "name" : "halloween_promotion",
        "percentPromotedItems" : 20,
        "filterArn": filter_arn,
        "filterValues": {
            "CATEGORY" : "\"halloween\""
        }
    }]
)


# Build a new dataframe for the recommendations
item_list = get_recommendations_response['itemList']
recommendation_id_list = []
recommendation_description_list = []
recommendation_category_list = []

for item in item_list:
    description = get_item_by_id(item['itemId'], items_df)
    recommendation_description_list.append(description)
    recommendation_id_list.append(item['itemId'])
    recommendation_category_list.append(get_category_by_id(item['itemId'], items_df))

user_recommendations_df = pd.DataFrame(recommendation_id_list, columns = ["ID"])
user_recommendations_df["description"] = recommendation_description_list
user_recommendations_df["category level 2"] = recommendation_category_list

pd.options.display.max_rows =20
display(user_recommendations_df)
```

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

### Deploy the SAM Project (Guided)

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/streaming_events/README.md

Deploy your project using SAM's guided deployment. You will be prompted to provide necessary parameters, including an email address for notifications.

```bash
sam deploy --guided
```

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

### Import Libraries and Check Versions

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_marketing_campaign/personalized_marketing_campaign_10032023_1600_github.ipynb

Imports necessary libraries such as boto3, botocore, transformers, and langchain. It also prints the installed versions of boto3, botocore, and transformers to verify the setup.

```python
import boto3
import botocore
import transformers

# Get the Boto3 version
boto3_version = boto3.__version__

# Get the Botocore version
botocore_version = botocore.__version__

# Get the Transformers version
transformers_version = transformers.__version__

# Print the Boto3 version
print("Current Boto3 Version:", boto3_version)

# Print the Botocore version
print("Current Botocore Version:", botocore_version)

# Print the Transformers version
print("Transformers Version:", transformers_version)
```

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

### Setup Personalize Clients

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/getting_started/notebooks/2.View_Campaign_And_Interactions.ipynb

Initializes boto3 clients for Amazon Personalize, Personalize Runtime, and Personalize Events.

```python
# Setup and Config
# Recommendations from Event data
personalize = boto3.client('personalize')
personalize_runtime = boto3.client('personalize-runtime')

# Establish a connection to Personalize's Event Streaming
personalize_events = boto3.client(service_name='personalize-events')
```

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

### Import Libraries and Install tqdm

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/data_science/offline_performance_evaluation/personalize_temporal_holdout.ipynb

Imports necessary libraries for data manipulation, AWS interaction, and progress tracking. Installs the 'tqdm' library for progress bars.

```python
import boto3, os
import json
import numpy as np
import pandas as pd
import time
from botocore.exceptions import ClientError
!pip install tqdm
from tqdm import tqdm_notebook
from metrics import mean_reciprocal_rank, ndcg_at_k, precision_at_k
```

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

### Test Get Recommended Products Tool

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/02_Recommender-Agent_Build-Agent-ConverseAPI.ipynb

Demonstrates how to call the `get_recommended_products` function with a sample user ID and shows the expected output format.

```python
get_recommended_products('1234')
```

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

### Include Filter Expression Example

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/filter_rotator/README.md

This example demonstrates how to include items based on interaction event types. It uses a SQL-like syntax to filter recommendations.

```sql
INCLUDE ItemID WHERE Interactions.event_type IN ('watched','favorited')
```

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

### Install Faker Library

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/automatic_context/notebooks/1.Building_Personalize_Campaign.ipynb

Install the Faker library, which is used for generating synthetic data for the campaign.

```python
!pip install Faker
```

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

### Test Get Product Details Tool

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/02_Recommender-Agent_Build-Agent-ConverseAPI.ipynb

Demonstrates how to call the `get_product_details` function with a sample product ID and shows the expected output format.

```python
get_product_details(['dff3f75f-838d-4b4d-9c7c-cee7ae2aac10'])
```

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

### POST Event Endpoint Example

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/streaming_events/README.md

Example body and endpoint for streaming interaction data. The API key must be included in the request header.

```json
{
    "Event":{
        "itemId": "ITEMID",
        "eventValue": EVENT-VALUE,
        "CONTEXT": "VALUE" //optional
    },
    "SessionId": "SESSION-ID-IDENTIFIER",
    "EventType": "YOUR-EVENT-TYPE",
    "UserId": "USERID"
}
```

```plaintext
https://XXXXXX.execute-api.us-east-1.amazonaws.com/dev2/history
```

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

### Get Initial User Recommendations

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/workshops/POC_in_a_box/05_Interacting_with_Campaigns_and_Filters.ipynb

Fetches the initial set of recommendations for a given user before any real-time events are applied. This serves as a baseline for comparison.

```python
# First pick a user
user_id = user

# Get recommendations for the user
get_recommendations_response = personalize_runtime.get_recommendations(
        campaignArn = userpersonalization_campaign_arn,
        userId = str(user_id),
    )

# Build a new dataframe for the recommendations
item_list = get_recommendations_response['itemList']
recommendation_list = []
for item in item_list:
    artist = get_movie_by_id(item['itemId'])
    recommendation_list.append(artist)
user_recommendations_df = pd.DataFrame(recommendation_list, columns = [user_id])
user_recommendations_df
```

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

### Create a Campaign

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/getting_started/notebooks/1.Building_Your_First_Campaign.ipynb

Deploy a trained solution version as a campaign to serve real-time recommendations. Configure minimum provisioned throughput and item exploration settings.

```python
create_campaign_response = personalize.create_campaign(
    name = "personalize-demo-camp",
    solutionVersionArn = solution_version_arn,
    minProvisionedTPS = 1,
    campaignConfig = {
        "itemExplorationConfig": {
            "explorationWeight": "0.5"
        }
    }
)

campaign_arn = create_campaign_response['campaignArn']
print(json.dumps(create_campaign_response, indent=2))
```

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

### Get Item Description by ID

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/getting_started/notebooks_managed_domains/Building_Your_First_Recommender_Ecommerce.ipynb

Retrieves the description for a specific item using its ID and the loaded items DataFrame. This is a direct usage example of the `get_item_by_id` helper function.

```python
# use a random valid id for a quick sanity check, modify the line of code bellow to a valid id in your dataset
get_item_by_id("c72257d4-430b-4eb7-9de3-28396e593381", items_df)
```

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

### Define CloudWatch Query Timeframe

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/evaluation/measuring_impact_of_recommendations/Measure_Impact_of_Recommendations.ipynb

Defines the start and end times for querying CloudWatch metrics in Unix timestamp format. This example queries data from the last 5 hours.

```python
# Define the start and end_times in Unix format this is the timeframe we will query
start_time = int(time.mktime((today + datetime.timedelta(hours = -5)).timetuple()))
end_time = int(time.mktime(today.timetuple()))


```

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

### AWS and Personalize Client Setup

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/lambda_examples/Sending_Events_to_S3.ipynb

Initializes AWS clients for Personalize, Personalize Runtime, and Personalize Events. Also configures S3 bucket name.

```python
# Setup and Config
# Recommendations from Event data
personalize = boto3.client('personalize')
personalize_runtime = boto3.client('personalize-runtime')

# Campaign ARN
Campaign_ARN = "arn:aws:personalize:us-east-1:059124553121:campaign/personalize-demo-camp"

# Dataset Group Arn:
datasetGroupArn = "arn:aws:personalize:us-east-1:059124553121:dataset-group/personalize-lambda-demo"

# Establish a connection to Personalize's Event Streaming
personalize_events = boto3.client(service_name='personalize-events')


# S3 Configuration
bucket = "personalizedemo9000chris"
```

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

### Create a Campaign

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/getting_started/notebooks/3.Best_Practices-Clientside.ipynb

Creates a campaign to deploy a solution version for real-time recommendations. Specify the solution version ARN and desired throughput.

```python
create_campaign_response = personalize.create_campaign(
    name = "DEMO-campaign",
    solutionVersionArn = solution_version_arn,
    minProvisionedTPS = 1
)

campaign_arn = create_campaign_response['campaignArn']
print(json.dumps(create_campaign_response, indent=2))
```

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

### Delete Filter Match Template Example

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/filter_rotator/README.md

This template matches filters with names starting with 'filter-include-recent-items-' and whose suffix (assumed to be a date) is older than one day. This allows for a transition period before deletion.

```text
starts_with(filter.name,'filter-include-recent-items-') and int(end(filter.name,8)) < int(datetime_format(now - timedelta_days(1),'%Y%m%d'))
```

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

### Initialize S3 and Personalize Clients

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/trending_now/trending_now_example.ipynb

Sets up the AWS SDK clients for S3, Personalize, and Personalize Events. Ensure your AWS credentials and region are configured.

```python
import boto3

region = "us-east-1" 

s3 = boto3.client('s3')
account_id = boto3.client('sts').get_caller_identity().get('Account')
bucket_name = account_id + "-" + region + "-" + "personalize-trending-now"
print('bucket_name:', bucket_name)

try:
    if region == "us-east-1":
        s3.create_bucket(Bucket=bucket_name)
    else:
        s3.create_bucket(
            Bucket = bucket_name,
            CreateBucketConfiguration={'LocationConstraint': region}
            )
except s3.exceptions.BucketAlreadyOwnedByYou:
    print("Bucket already exists. Using bucket", bucket_name)
```

```python
# Configure the SDK to Personalize:
personalize = boto3.Session(region_name=region).client('personalize')
personalize_events = boto3.Session(region_name=region).client('personalize-events')
personalize_runtime = boto3.Session(region_name=region).client('personalize-runtime')
```

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

### Install Python Packages for Generative AI

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_marketing_campaign/personalized_marketing_campaign_10032023_1600_github.ipynb

Installs or upgrades essential Python packages including pip, boto3, botocore, langchain, and transformers. Restart the kernel if prompted after installation.

```python
%pip install --upgrade pip
%pip install boto3 --upgrade
%pip install botocore --upgrade
```

```python
%pip install langchain
%pip install transformers
```

```python
%pip install langchain --upgrade
%pip install transformers --upgrade
```

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

### Update pip and Install Botocore

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/updating_datasets/update-item-dataset-schema-example.ipynb

Ensures a current version of botocore is installed, which is necessary for accessing newer Amazon Personalize features. This command upgrades pip and then installs or reinstalls botocore without its dependencies.

```python
import sys
!{sys.executable} -m pip install --upgrade pip
!{sys.executable} -m pip install --upgrade --no-deps --force-reinstall botocore

```

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

### Install ipywidgets

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/02_Recommender-Agent_Build-Agent-ConverseAPI.ipynb

Installs the ipywidgets library, which is necessary for creating interactive elements in Jupyter notebooks.

```python
!pip install ipywidgets
```

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

### Initialize AWS Clients and SDK

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/ml_ops_ds_sdk/Personalize-Stepfunction-Workflow.ipynb

Initializes Boto3 clients for Amazon Personalize and Personalize Runtime, along with necessary libraries for Step Functions. Sets up logging for the Step Functions SDK.

```python
import boto3
import json
import numpy as np
import pandas as pd
import time

personalize = boto3.client('personalize')
personalize_runtime = boto3.client('personalize-runtime')


import stepfunctions
import logging

from stepfunctions.steps import *
from stepfunctions.workflow import Workflow

stepfunctions.set_stream_logger(level=logging.INFO)

workflow_execution_role = "<Workflow exection role name>" # paste the StepFunctionsWorkflowExecutionRole ARN from above
```

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

### Configure Boto3 SDK

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/workshops/POC_in_a_box/06_Clean_Up_Resources.ipynb

Initializes the Boto3 clients for Personalize and Personalize Runtime. Ensure the AWS SDK is configured before running.

```python
import boto3
import json
```

```python
# Configure the SDK to Personalize:
personalize = boto3.client('personalize')
personalize_runtime = boto3.client('personalize-runtime')
```

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

### Install Pillow Library

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_marketing_campaign/personalized_marketing_campaign_10032023_1600_github.ipynb

Installs the Pillow library, a fork of PIL, for image manipulation. Ensure version compatibility.

```python
%pip install --quiet "pillow>=9.5,<10"
```

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

### Exclude Filter Expression Example

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/filter_rotator/README.md

This example shows how to exclude items that are out of stock. It uses a SQL-like syntax to filter recommendations.

```sql
EXCLUDE ItemID WHERE Items.out_of_stock IN ('yes')
```

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

### Apply Filters to Recommendations

Source: https://context7.com/aws-samples/amazon-personalize-samples/llms.txt

Demonstrates how to apply static, dynamic, and promotional filters when retrieving recommendations. Ensure correct quoting for string values in filterValues.

```python
response = personalize_runtime.get_recommendations(
    campaignArn= campaign_arn,
    userId     = "user_42",
    filterArn  = filter_arn_static
)
```

```python
response_genre = personalize_runtime.get_recommendations(
    campaignArn  = campaign_arn,
    userId       = "user_42",
    filterArn    = filter_arn_dynamic,
    filterValues = {"GENRE": '"Action"'}   # double-quotes required for string values
)
```

```python
import time
thirty_days_ago = int(time.time()) - 30 * 24 * 3600
response_promo = personalize_runtime.get_recommendations(
    campaignArn  = campaign_arn,
    userId       = "user_42",
    filterArn    = filter_arn_promo,
    filterValues = {"MIN_TS": str(thirty_days_ago)},
    promotionList= [{"filterArn": filter_arn_promo, "percentPromotedItems": 30}]
)
```

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

### Download Retail Demo Store Datasets

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/getting_started/notebooks_managed_domains/Building_Your_First_Recommender_Ecommerce.ipynb

Downloads the 'interactions.csv' and 'items.csv' datasets from the Retail Demo Store project using wget.

```bash
!wget https://code.retaildemostore.retail.aws.dev/csvs/interactions.csv
!wget https://code.retaildemostore.retail.aws.dev/csvs/items.csv
```

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

### Upgrade Pip and Install Botocore

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/01_Recommender-Agent_Configure-Personalize-Resources.ipynb

Ensures the latest version of pip and botocore are installed for optimal SDK performance and feature availability.

```bash
import sys
!{sys.executable} -m pip install --upgrade pip
!{sys.executable} -m pip install --upgrade --no-deps --force-reinstall botocore
```

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

### Display Langchain Package Information

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_marketing_campaign/personalized_marketing_campaign_10032023_1600_github.ipynb

Shows the installed version and metadata for the langchain package. This is useful for confirming the installation and checking compatibility.

```python
%pip show langchain
```

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

### Create Campaign with Item Exploration Configuration

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/user_personalization/user-personalization-with-exploration.ipynb

Creates a campaign for the solution version, enabling item exploration with a specified exploration weight and item age cut-off. This encourages exploration of newer items.

```python
create_campaign_response = personalize.create_campaign(
    name = prefix + suffix,
    solutionVersionArn = solution_version_arn,
    minProvisionedTPS = 1,
    campaignConfig = {
        "itemExplorationConfig": {
            "explorationWeight": "0.9",
            "explorationItemAgeCutOff": "7"
        }
    }
)

campaign_arn = create_campaign_response['campaignArn']
print('campaign_arn:', campaign_arn)
```

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

### Initialize S3 Client

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/user_personalization/user-personalization-with-exploration.ipynb

Sets up an S3 client using provided AWS credentials and region. This client is used for interacting with S3 buckets.

```python
# Public s3 bucket owned by Personalize service which used to store the example dataset.

personalize_s3_bucket = "personalize-cli-json-models"
s3_client = boto3.Session(aws_access_key_id=accessKeyId,
                           aws_secret_access_key=secretAccessKey, region_name=region_name).client('s3')
```

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

### Create Campaign for Recommendations

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/trending_now/trending_now_example.ipynb

Deploys the trained solution version as a campaign to serve real-time recommendations. It sets the minimum provisioned transactions per second (TPS).

```python
campaign_response = personalize.create_campaign(
    name="trending-now-campaign",
    solutionVersionArn=solution_version_arn,
    minProvisionedTPS=1,
)
campaign_arn = campaign_response['campaignArn']
```

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

### Get Recommendations from Campaign

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/trending_now/trending_now_example.ipynb

Retrieves a list of recommended items from a trained Personalize campaign. Use this to get initial recommendations after training.

```python
item_ids = personalize_runtime.get_recommendations(campaignArn=campaign_arn, 
                                                   numResults=25)
```

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

### Initialize Bedrock LLM with Anthropic Claude Instant

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_marketing_campaign/personalized_marketing_campaign_10032023_1600_github.ipynb

Configure and initialize the Bedrock LLM using the 'anthropic.claude-instant-v1' model. This model is chosen for its speed, cost-effectiveness, and capability in various text-based tasks. Ensure 'boto3_bedrock' client is available.

```python
from langchain.llms.bedrock import Bedrock

inference_modifier = {"max_tokens_to_sample":4096, 
                      "temperature":0.7,
                      "top_k":250,
                      "top_p":1,
                      "stop_sequences": ["\n\nHuman"]
                     }

textgen_llm = Bedrock(model_id = "anthropic.claude-instant-v1",
                    client = boto3_bedrock, 
                    model_kwargs = inference_modifier 
                    )

```

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

### Print 'aws-sims' Solution Version Details

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/related_items/personalize_aws_similar_items_example.ipynb

Outputs the details of the created 'aws-sims' solution version, including its ARN. This confirms the successful creation of the solution version.

```python
sims_solution_version_arn = sims_solution_version_response['solutionVersionArn']
print(json.dumps(sims_solution_version_response, indent=2))
```

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

### Install Step Functions SDK

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/ml_ops_ds_sdk/Personalize-Stepfunction-Workflow.ipynb

Installs or upgrades the AWS Step Functions Data Science SDK. Ensure you are using the correct Python executable for your environment.

```python
#import sys
#!{sys.executable} -m pip install --upgrade stepfunctions
```

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

### Download Product Details YAML

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/02_Recommender-Agent_Build-Agent-ConverseAPI.ipynb

Downloads the products.yaml file from a GitHub repository using wget. This file contains product details for the retail demo store.

```bash
!wget https://raw.githubusercontent.com/aws-samples/retail-demo-store/master/src/products/data/products.yaml -O ./products.yaml
```

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

### Print Solution Version Details

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/metadata/item-text-unstructured-metadata.ipynb

Prints the full response from creating a solution version, including the solution version ARN and metadata. Use `json.dumps` for pretty-printing.

```python
user_personalization_with_solution_version_arn = user_personalization_solution_version_response['solutionVersionArn']
print(json.dumps(user_personalization_solution_version_response, indent=2))
```

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

### Install and Upgrade Pip and Botocore

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/filters/promotions/Promotions.ipynb

Ensures the latest versions of pip and botocore are installed for optimal SDK functionality. This is a prerequisite for using the AWS SDK with Amazon Personalize.

```python
# Get the latest version of botocore to ensure we have the latest features in the SDK
import sys
!{sys.executable} -m pip install --upgrade pip
!{sys.executable} -m pip install --upgrade --no-deps --force-reinstall botocore
```

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

### Install Boto3 for Amazon Bedrock Converse API

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/02_Recommender-Agent_Build-Agent-ConverseAPI.ipynb

Installs or upgrades the Boto3 library to a version compatible with the Amazon Bedrock Converse API. Ensure version is greater than 1.34.123.

```python
!pip install boto3==1.34.123
```

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

### Build and Deploy the Application with SAM CLI

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/filter_rotator/README.md

Use these commands to build the application's source code and then deploy it to your AWS account. The deploy command will prompt for configuration details.

```bash
sam build --use-container --cached
sam deploy --guided
```

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

### Import Libraries and Install Unzip

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/getting_started/notebooks/1.Building_Your_First_Campaign.ipynb

Imports necessary Python libraries for AWS SDK, data manipulation, and time operations. Also installs the 'unzip' utility using conda.

```python
# Imports
import boto3
import json
import numpy as np
import pandas as pd
import time
import os

!conda install -y -c conda-forge unzip
```

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

### Dynamic Filter Expression Example

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/filter_rotator/README.md

This example illustrates a dynamic filter that includes items based on a genre specified at runtime. The filter expression uses a placeholder ($GENRES) for the dynamic value.

```sql
INCLUDE ItemID WHERE Items.genre IN ($GENRES)
```

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

### Set Data Directory and Download MovieLens Dataset

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/workshops/POC_in_a_box/completed/01_Validating_and_Importing_User_Item_Interaction_Data.ipynb

Configures the data directory and conditionally downloads either the small or full MovieLens dataset, then unzips it. Use `USE_FULL_MOVIELENS = True` for the larger 25M dataset.

```python
USE_FULL_MOVIELENS = False
```

```python
data_dir = "poc_data"
!mkdir $data_dir

if not USE_FULL_MOVIELENS:
    !cd $data_dir && wget http://files.grouplens.org/datasets/movielens/ml-latest-small.zip
    !cd $data_dir && unzip ml-latest-small.zip
    dataset_dir = data_dir + "/ml-latest-small/"
else:
    !cd $data_dir && wget http://files.grouplens.org/datasets/movielens/ml-25m.zip
    !cd $data_dir && unzip ml-25m.zip
    dataset_dir = data_dir + "/ml-25m/"
```

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

### Get Personalized Ranking with a Rerank Campaign

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/workshops/POC_in_a_box/completed/05_Interacting_with_Campaigns_and_Filters.ipynb

Use the personalize_runtime client to get a reranked list of items for a given user and input list. This is useful for reordering a set of items based on personalization.

```python
get_recommendations_response_rerank = personalize_runtime.get_personalized_ranking(
        campaignArn = rerank_campaign_arn,
        userId = user_id,
        inputList = rerank_item_list
)
```

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

### Import Libraries and Install Packages

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/objective_optimization/objective-optimization.ipynb

Imports necessary Python libraries (boto3, pandas, numpy, etc.) and installs the 'unzip' package using conda. Sets pandas display options for chained assignments.

```python
# Imports
import boto3
import json
import numpy as np
import pandas as pd
import time
from botocore.exceptions import ClientError

!conda install -y -c conda-forge unzip

pd.options.mode.chained_assignment = None
```

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

### Initialize Bedrock Client and List Models

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_marketing_campaign/personalized_marketing_campaign_10032023_1600_github.ipynb

Initializes the AWS Bedrock client for the 'us-east-1' region and lists available foundation models. Requires AWS credentials and configuration.

```python
import boto3
from PIL import Image
bedrock_client = boto3.client('bedrock' , 'us-east-1', endpoint_url='https://bedrock.us-east-1.amazonaws.com')
bedrock_client.list_foundation_models()
```

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

### Inference Job Output Example

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/updating_datasets/update-datasets-user-personalization-example.ipynb

This is an example of the output from a batch inference job. It shows the job status progression and the final JSON output containing recommended items and their scores for specific user IDs. Errors, if any, are also detailed.

```json
Inference Job Started on:  04:02:59 PM
DatasetInferenceJob: CREATE PENDING
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: CREATE IN_PROGRESS
DatasetInferenceJob: ACTIVE
Inference Job Completed on:  04:16:01 PM
{"input":{"userId":"4016"},"output":{"recommendedItems":["ccdf737c-c4fd-4c78-abd2-d5ef0428ef20","425cc876-3935-4e87-ad8d-77f42b0b6a75","6be08307-1ec0-44dc-b436-5d489a8010e8","2c1b34d6-0f3d-463d-be76-226cb87bdc6d","89c4eeb4-c146-4434-a9f1-6943b4b552dc","5a94b7d5-b210-44b3-9287-c8b0b5488a15","61b1ad14-4e70-4029-ba55-d17bbf4ab62b","8f8f015a-4166-4e9e-ac0b-6d980614ca5d","eecbee28-73a3-425d-84e8-516c326e399c","1daacea7-7d46-464a-8326-ed81951fecab"],"scores":[0.8449224,0.0186943,0.0148416,0.0100115,0.008885,0.0055925,0.0055068,0.0053057,0.0040184,0.0038556]},"error":null}
{"input":{"userId":"2755"},"output":{"recommendedItems":["3630053e-3962-4549-bcce-402c3a980557","6d488475-1d67-4076-96b1-8e706709a847","b947ee58-a7e7-40bf-9926-42a445f3480f","90ccfbb9-4538-4951-af8d-4f728578b237","d537d92a-23fe-4673-a697-795652ff10c8","78080d05-b078-441f-b245-54b2a2dec872","61840d6a-6ba2-4ece-a644-6db6a3377b1c","2e95f6fc-6be7-46cf-9e50-8c35313c2768","b630250c-41f3-4f14-865c-c1dc12e448ac","d2d8147f-0f24-42c3-bcbe-a232bab7e94d"],"scores":[0.5338279,0.0922309,0.0577862,0.0209987,0.0206773,0.0148014,0.0141705,0.0129197,0.0120157,0.0077951]},"error":null}
{"input":{"userId":"5866"},"output":{"recommendedItems":["5afced84-ed2d-4520-a06d-dcfeab382e52","6cc0deb8-4a56-4148-a2ab-677277522c80","e1146e90-3274-4ad6-a6a2-0170f0f8d597","575c0ac0-5494-4c64-a886-a9c0cf8b779a","24c62ad2-6977-4f69-be75-e37d897c1434","4496471c-b098-4915-9a1a-8b9e60043737","9c1a2048-7aac-4565-b836-d8d4f726322c","ccb407b1-7620-4303-8521-fea86c51f503","8cd7ffe0-a8a6-45b1-8d1f-bf731c9cd17b","aa564ee3-67ef-4428-8ad9-fe785a0fff63"],"scores":[0.5773515,0.2774695,0.1156833,0.0155479,0.0066222,0.0059881,3.773E-4,7.17E-5,6.78E-5,3.1E-5]},"error":null}
{"input":{"userId":"2896"},"output":{"recommendedItems":["1dd4c2da-d174-43b1-8d40-fadc666c26c9","e99c24df-ebe9-429e-8c69-cd80132b87b3","e06f53cd-7776-41ce-9f7a-a88986192e24","75cb828e-ccc5-41ff-9bdd-9ac3dc7740aa","153b2374-36e3-466c-b08c-1078b839cd9b","2f2995da-4768-478a-a4ae-906b76d8c6fe","2a0a5c7b-ca68-4abf-9798-18ffb706832b","e8e48eb7-0b66-4087-b280-1c3a62804a5c","e5816ea5-3ce2-4b86-9530-9b221b357b43","17ab5081-8414-4cab-9003-033ec02b44da"],"scores":[0.6146742,0.0403633,0.0256617,0.0107649,0.0107413,0.0106864,0.0100017,0.007999,0.0077459,0.0065279]},"error":null}
{"input":{"userId":"5023"},"output":{"recommendedItems":["c6dd0909-46f3-4cf9-a059-fbdff93198dd","441c2a65-4b68-4864-b014-04a9bd9fe08a","84d6c26d-9760-49d8-854b-0a22becd8241","9257351d-59f7-481a-86c4-30dea451afa2","635be5a7-3345-46f9-aa0d-419a6652b0f2","72ae72f3-e7f0-4f03-b8eb-12e78c77741d","d4cf35dd-b543-4b4f-9efb-c2de473c3fed","4994caee-f0b7-4ce8-a4df-d542ce1d9bda","0e3eb8f1-8f23-41fd-9f45-8e7747a5eb37","fe96a096-a0b6-4b20-a332-e11db6c0c7b0"],"scores":[0.3460568,0.2430023,0.1881302,0.025838,0.023981,0.0178926,0.0130894,0.0092487,0.0080664,0.0077484]},"error":null}
CPU times: user 366 ms, sys: 32.5 ms, total: 399 ms
Wall time: 13min 1s
```

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

### Build and Deploy ML Ops Pipeline

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/ml_ops/README.md

Commands to clone the repository, build the SAM project, and deploy the Step Functions pipeline. Follow the guided deployment for configuration.

```bash
git clone https://github.com/aws-samples/amazon-personalize-samples.git
```

```bash
cd next_steps/operations/ml_ops/personalize-step-functions
```

```bash
sam build
```

```bash
sam deploy --guided
```

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

### Execute Recommendation Workflow

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/ml_ops_ds_sdk/Personalize-Stepfunction-Workflow.ipynb

Starts an execution of the 'Recommendation_Workflow4' using the `execute()` method.

```python
recommendation_workflow_execution = recommendation_workflow.execute()
```

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

### Create and Train Solution Version

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/trending_now/trending_now_example.ipynb

Creates a new version of the solution and initiates the training process. The `trainingMode` can be set to 'FULL' for complete retraining.

```python
solution_version_response = personalize.create_solution_version(solutionArn=solution_arn, 
                                                         trainingMode="FULL")
solution_version_arn = solution_version_response["solutionVersionArn"]
print(json.dumps(solution_version_response, indent=4, default=str))
```

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

### Create Solution Version

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_marketing_campaign/airline_ticket_user_segmentation_09212023_github.ipynb

Initiates the creation of a new version for a Personalize solution.

```python
create_solution_version_response = personalize.create_solution_version(
    solutionArn = solution_arn
)
solution_version_arn = create_solution_version_response['solutionVersionArn']
print(solution_version_arn)
```

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

### Display Boto3 Version

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/02_Recommender-Agent_Build-Agent-ConverseAPI.ipynb

Prints the installed version of the Boto3 library to verify compatibility.

```python
print(boto3.__version__)
```

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

### Create 'Top Picks for You' Recommender

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/user_personalized_marketing_messaging_with_amazon_personalize_and_gen_ai/03_Train_Personalize_Model_02_Training.ipynb

Creates a 'Top picks for you' recommender for the VIDEO_ON_DEMAND domain. Includes configuration to enable metadata with recommendations. Handles cases where the recommender already exists.

```python
recommenderConfig = {"enableMetadataWithRecommendations": True}
try:
    create_recommender_response = personalize.create_recommender(
        name = recommender_top_picks_for_you_name,
        recipeArn = 'arn:aws:personalize:::recipe/aws-vod-top-picks',
        datasetGroupArn = workshop_dataset_group_arn,
        recommenderConfig = {"enableMetadataWithRecommendations": True}
    )
    workshop_recommender_top_picks_arn = create_recommender_response["recommenderArn"]
    
    print (json.dumps(create_recommender_response))
    print ('\nCreating the Top Picks For You recommender with workshop_recommender_top_picks_arn = {}'.format(workshop_recommender_top_picks_arn))
    
except personalize.exceptions.ResourceAlreadyExistsException as e:
    workshop_recommender_top_picks_arn =  'arn:aws:personalize:'+region+':'+account_id+':recommender/'+recommender_top_picks_for_you_name
    print('The Top Picks For You recommender {} already exists.'.format(workshop_recommender_top_picks_arn))
    print ('\nWe will be using the existing Top Picks For You recommender with workshop_recommender_top_picks_arn = {}'.format(workshop_recommender_top_picks_arn))
```

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

### Get Item Description

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/related_items/personalize_aws_similar_items_example.ipynb

Retrieves the description of an item given its ID from the items DataFrame.

```python
def get_item_description(item_id):
    """
    Takes in an ID, returns its brand
    """

    return items_df.query('ITEM_ID=="{}"'.format(item_id))['DESCRIPTION'].item()
```

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

### Get Item Price

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/related_items/personalize_aws_similar_items_example.ipynb

Retrieves the price of an item given its ID from the items DataFrame.

```python
def get_item_price(item_id):
    """
    Takes in an ID, returns its brand
    """

    return items_df.query('ITEM_ID=="{}"'.format(item_id))['PRICE'].item()
```

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

### Create Campaign with Automatic Exploration

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/automatic_context/notebooks/1.Building_Personalize_Campaign.ipynb

Create a campaign (a hosted solution version endpoint) and configure automatic item exploration. Set the minimum provisioned TPS and specify exploration parameters like explorationWeight.

```python
create_campaign_response = personalize.create_campaign(
    name = "personalize-auto-context-demo-campaign",
    solutionVersionArn = solution_version_arn,
    minProvisionedTPS = 1,
    campaignConfig = {
        "itemExplorationConfig": {
            "explorationWeight": "0.5"
        }
    }
)

campaign_arn = create_campaign_response['campaignArn']
print(json.dumps(create_campaign_response, indent=2))
```

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

### Download Sample Datasets

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/user_personalization/user-personalization-with-exploration.ipynb

Downloads sample interaction and item metadata datasets from a public S3 bucket to the local environment for use in the demo.

```python
interaction_dataset_key = "sample-dataset/interactions-sample.csv"
items_dataset_key = "sample-dataset/items-with-creation-timestamp-sample.csv"
interactions_file = os.getcwd() + "/interaction_raw.csv"
items_metadata_file = os.getcwd() + "/items_raw.csv"
s3_client.download_file(personalize_s3_bucket, interaction_dataset_key, interactions_file)
s3_client.download_file(personalize_s3_bucket, items_dataset_key, items_metadata_file)
```

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

### Get Item Brand

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/related_items/personalize_aws_similar_items_example.ipynb

Retrieves the brand of an item given its ID from the items DataFrame.

```python
def get_item_brand(item_id):
    """
    Takes in an ID, returns its brand
    """

    return items_df.query('ITEM_ID=="{}"'.format(item_id))['BRAND'].item()
```

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

### Render Campaign Workflow Graph (Commented Out)

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/ml_ops_ds_sdk/Personalize-Stepfunction-Workflow.ipynb

This is a commented-out example for rendering the graph of a campaign workflow.

```python
#Main_workflow_execution.render_graph()
```

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

### Create Similar Items Solution and Handle Existing Resources

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/01_Recommender-Agent_Configure-Personalize-Resources.ipynb

Creates a Personalize solution for the Similar-Items recipe. If a solution with the same name already exists, it retrieves the existing solution ARN and its most recent solution version ARN.

```python
similar_items_solution_version_arn = None

try:
    create_solution_response = personalize.create_solution(
        name = "related-items-demo",
        datasetGroupArn = dataset_group_arn,
        recipeArn = similar_items_recipe_arn
    )

    similar_items_solution_arn = create_solution_response['solutionArn']
    print(json.dumps(create_solution_response, indent=2))
except personalize.exceptions.ResourceAlreadyExistsException:
    print('You aready created this solution, seemingly')
    paginator = personalize.get_paginator('list_solutions')
    for paginate_result in paginator.paginate(datasetGroupArn = dataset_group_arn):
        for solution in paginate_result['solutions']:
            if solution['name'] == 'related-items-demo':
                similar_items_solution_arn = solution['solutionArn']
                print(f'Similar Items solution ARN = {similar_items_solution_arn}')
                
                response = personalize.list_solution_versions(
                    solutionArn = similar_items_solution_arn,
                    maxResults = 100
                )
                if len(response['solutionVersions']) > 0:
                    similar_items_solution_version_arn = response['solutionVersions'][-1]['solutionVersionArn']
                    print(f'Will use most recent solution version for this solution: {similar_items_solution_version_arn}')
                    
                break
```

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

### Get Movie Title by ID

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/workshops/POC_in_a_box/05_Interacting_with_Campaigns_and_Filters.ipynb

Retrieves the title of a movie from the loaded DataFrame using its ID.

```python
movie_id_example = 589
title = items_df.loc[movie_id_example]['title']
print(title)
```

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

### Get Recommendation Output

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/ml_ops_ds_sdk/Personalize-Stepfunction-Workflow.ipynb

Retrieves the output payload from the recommendation workflow execution, specifically the 'item_list'.

```python
item_list = recommendation_workflow_execution.get_output()['Payload']['item_list']
```

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

### Load Item Data and Sample User

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/operations/ml_ops_ds_sdk/Personalize-Stepfunction-Workflow.ipynb

Loads item metadata from a CSV file and samples a user, item, and rating from the dataset for recommendation generation.

```python
items = pd.read_csv('./ml-100k/u.item', sep='|', usecols=[0,1], encoding='latin-1')
items.columns = ['ITEM_ID', 'TITLE']


user_id, item_id, rating, timestamp = data.sample().values[0]

user_id = int(user_id)
item_id = int(item_id)

print("user_id",user_id)
print("items",items)


item_title = items.loc[items['ITEM_ID'] == item_id].values[0][-1]
print("USER: {}".format(user_id))
print("ITEM: {}".format(item_title))
print("ITEM ID: {}".format(item_id))
```

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

### Get Random User ID

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/getting_started/notebooks/1.Building_Your_First_Campaign.ipynb

Selects a random user ID from the dataset for generating recommendations.

```python
# Getting a random user:
user_id, item_id, _ = data.sample().values[0]
print("USER: {}".format(user_id))
```

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

### Create SQLite Database for Products

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/02_Recommender-Agent_Build-Agent-ConverseAPI.ipynb

Creates a local SQLite database named 'products.db' and populates it with product data from a YAML file. It defines a 'products' table and inserts data, handling potential missing fields.

```python
import sqlite3
import yaml

# Read the YAML file and parse the data
with open('products.yaml', 'r') as file:
    products = yaml.safe_load(file)

# Create an SQLite database and table for storing the products
conn = sqlite3.connect('products.db')
cursor = conn.cursor()

# Create the products table
cursor.execute('''
    CREATE TABLE IF NOT EXISTS products (
        id TEXT PRIMARY KEY,
        category TEXT,
        current_stock INTEGER,
        description TEXT,
        gender_affinity TEXT,
        image TEXT,
        image_license TEXT,
        link TEXT,
        name TEXT,
        price REAL,
        style TEXT,
        where_visible TEXT,
        promoted BOOLEAN
    )
''')

# Insert the products into the SQLite table
for product in products:
    # Prepare the product data with null values for missing fields
    product_data = (
        product.get('id', None),
        product.get('category', None),
        product.get('current_stock', None),
        product.get('description', None),
        product.get('gender_affinity', None),
        product.get('image', None),
        product.get('image_license', None),
        product.get('link', None),
        product.get('name', None),
        product.get('price', None),
        product.get('style', None),
        product.get('where_visible', None),
        product.get('promoted', False)
    )

    # Insert the product into the table
    cursor.execute('''
        INSERT OR IGNORE INTO products (
            id, category, current_stock, description, gender_affinity, image,
            image_license, link, name, price, style, where_visible, promoted
        ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
    ''', product_data)

# Commit the changes and close the connection
conn.commit()
conn.close()
```

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

### Get Random Item ID

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/related_items/personalize_aws_similar_items_example.ipynb

Selects a random item ID from the asin_interaction_count_df DataFrame for testing recommendations.

```python
asin_interaction_count_df.sample()['asin'].item()
```

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

### Setup AWS Credentials

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/user_personalization/user-personalization-with-exploration.ipynb

Configure your AWS access key ID, secret access key, and region for API access. Ensure these credentials have the necessary permissions.

```python
accessKeyId = ""
secretAccessKey = ""
region_name = ""
```

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

### Get Item Interaction Count

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/core_use_cases/related_items/personalize_aws_similar_items_example.ipynb

Retrieves the interaction count for a given item ID from the asin_interaction_count_df DataFrame.

```python
def get_item_count(item_id):
    return asin_interaction_count_df.query('asin=="{}"'.format(item_id))['count'].item()
```

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

### Create Solution Version for Similar Items

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/generative_ai/personalized_recommender_agent/01_Recommender-Agent_Configure-Personalize-Resources.ipynb

Creates a solution version for the similar items solution if one does not already exist. Prints the ARN upon successful creation.

```python
if not similar_items_solution_version_arn:
    create_solution_version_response = personalize.create_solution_version(
        solutionArn = similar_items_solution_arn
    )

    similar_items_solution_version_arn = create_solution_version_response['solutionVersionArn']
    print(json.dumps(create_solution_version_response, indent=2))
else:
    print(f'Solution version {similar_items_solution_version_arn} already exists; not creating')
```

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

### Get Recommendations

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/getting_started/notebooks/1.Building_Your_First_Campaign.ipynb

Retrieves personalized movie recommendations for a given user ID using the trained campaign.

```python
get_recommendations_response = personalize_runtime.get_recommendations(
    campaignArn = campaign_arn,
    userId = str(user_id),
)
# Update DF rendering
pd.set_option('display.max_rows', 30)

print("Recommendations for user: ", user_id)

item_list = get_recommendations_response['itemList']

recommendation_list = []

for item in item_list:
    title = get_movie_title(item['itemId'])
    recommendation_list.append(title)
    
recommendations_df = pd.DataFrame(recommendation_list, columns = ['OriginalRecs'])
recommendations_df
```

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

### Initialize Personalize Clients

Source: https://github.com/aws-samples/amazon-personalize-samples/blob/master/next_steps/data_science/offline_performance_evaluation/personalize_temporal_holdout.ipynb

Initializes boto3 clients for Amazon Personalize and Personalize Runtime services.

```python
personalize = boto3.client('personalize')
personalize_runtime = boto3.client('personalize-runtime')
```