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