Detecting Custom Entities with an Asynchronous Batch Job with Amazon Comprehend

Once you have a trained model, use the StartEntitiesDetectionJob operation to detect custom entities in your documents.

Before you begin

Before you can detect custom entities, you must have a custom entity recognition model. For more information about these models, see Training Custom Entity Recognizers. For the steps to train a model, see Creating a Custom Entity Recognizer Using the Console.

Using this operation, you provide the same information as you would when detecting preset entities. However, in addition to the input and output locations (S3 buckets), you also provide the EntityRecognizerArn, which is the Amazon Resource Name (ARN) of the trained model. This ARN is supplied by the response to the CreateEntityRecognizer operation.

You can examine one document or many, and each model can be trained on up to 12 custom entities at a time. You can search for up to 12 entities per StartEntitiesDetectionJob operation..

To detect custom entities in a document set, you use the following request syntax:

The examples are formatted for Unix, Linux, and macOS. For Windows, replace the backslash (\) Unix continuation character at the end of each line with a caret (^).

aws comprehend start-entities-detection-job \
     --entity-recognizer-arn "entity recognizer arn" \
     --job-name job name \
     --data-access-role-arn "data access role arn" \
     --language-code en \
     --input-data-config "S3Uri=s3://Bucket Name/Bucket Path" \
     --output-data-config "S3Uri=s3://Bucket Name/Bucket Path/" \
     --region region

Amazon Comprehend will respond with the JobID and JobStatus and will return the output from the job in the S3 bucket that you specified in your request. This output will be similar to the following:

{"File": "50_docs", "Line": 0, "Entities": [{"BeginOffset": 0, "EndOffset": 22, "Score": 0.9763959646224976, "Text": "John Johnson", "Type": "JUDGE"}"]}
{"File": "50_docs", "Line": 1, "Entities": [{"BeginOffset": 11, "EndOffset": 15, "Score": 0.9615424871444702, "Text": "Thomas Kincaid", "Type": "JUDGE"}}]}