Filtering batch recommendations and user segments (custom resources)

Filtering batch recommendations and user segments works nearly the same as filtering real-time recommendations. It follows the same workflow described in Batch recommendations and user segments (custom resources). To filter batch recommendations or user segments, you do the following:

1. Create a filter just like you would for real-time recommendations. For more information see Filtering real-time recommendations.

2. Prepare your input data and upload it to Amazon S3 as described in Preparing input data for batch recommendations or Preparing input data for user segments. If your filter uses placeholder parameters, you must add an additional filterValues object. For more information, see Providing filter values in your input JSON. If your filter doesn't use placeholder parameters, your input data can follow the examples in Batch inference job input and output JSON examples Batch segment job input and output JSON examples

3. Create a separate location for your output data, either a folder or a different Amazon S3 bucket.

4. Create a batch inference job or a batch segment job. When you create the job, specify the Amazon Resource Name (ARN) of your filter.

5. When the batch inference or batch segment job is complete, retrieve the recommendations or user segments from your output location in Amazon S3.

Providing filter values in your input JSON

For filters with placeholder parameters, such as $GENRE, you must provide the values for the parameters in a filterValues object in your input JSON. For a filterValues object, each key is a parameter name. Each value is the criteria that you are passing as a parameter. Surround each value with escaped quotes: "filterValues":{"GENRES":"\"drama\""}. For multiple values, separate each value with a comma: "filterValues":{"GENRES":"\"horror\",\"comedy\",\"drama\""}

Batch inference job input JSON example

The following is an example of the first few lines of a JSON input file for a batch inference job. The example includes the filterValues object. The GENRES key corresponds to a $GENRES placeholder in the filter expression. The job in this example uses the User-Personalization recipe. For RELATED_ITEMS recipes, provide an itemId instead of the userId. For PERSONALIZED_RANKING recipes provide the userID and an itemList.

{"userId": "5","filterValues":{"GENRES":"\"horror\",\"comedy\",\"drama\""}}
{"userId": "3","filterValues":{"GENRES":"\"horror\",\"comedy\""}}
{"userId": "34","filterValues":{"GENRES":"\"drama\""}}
    

For more examples of batch inference job input data by recipe see Batch inference job input and output JSON examples. You can use these examples as a starting point and add the filterValues object from the above example.

Batch segment job input JSON example

The following is an example of the first few lines of a JSON input file with filter values for a batch segment job. The GENRES key corresponds to a $GENRES placeholder in the filter expression.

{"itemAttributes": "ITEMS.genres = \"Comedy\" AND ITEMS.genres = \"Action\"","filterValues":{"COUNTRY":"\"Japan\""}}
{"itemAttributes": "ITEMS.genres = \"Horror\"","filterValues":{"COUNTRY":"\"United States\"\""}}
{"itemAttributes": "ITEMS.genres = \"Action\" AND ITEMS.genres = \"Adventure\"","filterValues":{"COUNTRY":"\"England\""}}
    

For more examples of batch inference job input data by recipe see Batch segment job input and output JSON examples. You can use these examples as a starting point and add the filterValues object from the above example.

Filtering batch workflows (console)

To filter batch workflows with the Amazon Personalize console, you create a filter and then you create a batch inference job or batch segment job and choose the filter. For complete step by step instructions, see Creating a batch inference job (console) and Creating a batch segment job (console).

Filtering batch workflows (AWS SDKs)

To filter batch recommendations with the AWS SDKs, create a filter and include the FilterArn parameter in the CreateBatchInferenceJob or CreateBatchSegmentJob request.

The following code shows how to create a batch inference job with a filter using the AWS SDK for Python (Boto3). We recommend using a different location for your output data (either a folder or a different Amazon S3 bucket). For complete explanation of all fields, see see Creating a batch inference job (AWS SDKs).

import boto3
     
personalize = boto3.client("personalize")
 
personalize_rec.create_batch_inference_job (
    solutionVersionArn = "Solution version ARN",
    jobName = "Batch job name",
    roleArn = "IAM role ARN",
    filterArn = "Filter ARN",
    jobInput = 
        {"s3DataSource": {"path": "S3 input path"}},
    jobOutput = e
        {"S3DataDestination": {"path": "S3 output path"}}
)