Overview
The classification.yaml configuration file classifies the documents and then the classification-results.yaml determines which extraction model should be used for the system based on the document type.
Template
classification.yaml
kind: document metadata: name: extraction/v1/document-classification spec: classificationModel: commonclassification10withprebuiltandABC minimumClassificationConfidence: 0
| Parameter | Description |
|---|---|
| Classification model | The classification model name. |
| Minimum classification confidence | The confidence level based on which this classification model will be used. |
classification-results.yaml
kind: ruleSet metadata: name: extraction/v1/document-classification-result spec: nodes: - if: 'it["DocumentType"] == "prebuilt-invoice"' then: extractionModel: prebuilt-invoice extractionSuffix: currency,grossAmount,freightCharges - if: 'it["DocumentType"] == "ABCInvoice1"' then: extractionModel: ABCInvoice1
| Parameter | Description |
|---|---|
| If then | The condition to decide which extraction model should be used based on document type. |
For example, in the above template, if the document type is ABCInvoice1 then the extraction model to be used is ABCInvoice.
