Keyphrase Configuration

V1.1

Overview

In the key-phrase.yaml configuration file you can define the key words/phrases based on which a document is searched and categorized. If a key phrase is associated with more than one category, only the first category it appears in will be considered for extraction.

Keyphrase configuration supports different languages.

Template

kind: document
metadata:
  name: extraction/v1/documents/key-phrase
spec:
  categories:
    - name: Credit Invoice
      keyPhrases:
        - "Credit Note"
        - "Credit"
        - "Credit Memo"
        - "クレジット"
        - "Crédito"
    - name: Non AP Document
      keyPhrases:
        - "Purchase Order"
        - "Proforma Invoice"
        - "Statement"
        - "Receipts"
        - "Service Quote"
        - "Order Confirmation"
        - "Sales Order"
        - "声明"
        - "Declaración"
        - "Proforma"
    - name: Test-Invoice
      keyPhrases:
        - "invoice"

For example, in the above template:

    - name: Credit Invoice
      keyPhrases:
        - "Credit Note"
        - "Credit"
        - "Credit Memo"
        - "クレジット"
        - "Crédito"

If the above mentioned key phrases are found for an invoice, then the invoice is categorized as a Credit Invoice.