AWS Kendra

learn Amazon Kendra, intelligent search service powered by machine learning

Concepts

• Index - two types. document (unstructured) and FAQ (structured)

  • You can add documents directly to an index using the BatchPutDocument operation
  • You can add questions and answers (FAQs) directly to your index using the console or the CreateFaq operation

• Data Source - s3, salesforce documents, sharepoint, custom

• Documents - html, ppt, doc, pdf, txt

  • custom fields - can define custom fields of types: Date, Number, String, String List

• Queries - natural language (NLP), keyword queries, Factoid questions — Simple who, what, when, or where questions.

  • Facets are scoped views of a set of search results
  • By default, Query returns all search results. To filter responses, you can perform logical operations on the document attributes.

• Tags - can assign tags to indexes, data sources, and FAQs

• default encryption at rest with customer of AWS KMS keys

General Steps

1. create index
2. add data sources for index
3. synchronize the data source
4. query/search the index

Editions

• Developer
• Enterprise

You specify Edition: DEVELOPER_EDITION | ENTERPRISE_EDITION when you create an index

Creating Example Index

see main.py

sam deploy --guided
# copy outputs into `.env`
python main.py

Resources

• Amazon Kendra Docs
• Getting started (AWS SDK for Python (Boto3)) - Amazon Kendra - shows create index, create data source for index, sync data source, all while waiting for each step to complete in between
• Deploying Amazon Kendra - pre-built react frontend for kendra. provides <search /> react component
  • https://kendrasamples.s3.amazonaws.com/kendrasamples.zip
• aws-samples/enterprise-search-with-amazon-kendra-workshop
• Kendra | Using a custom data source