workflow_run.outputs is dynamic based on the type of document processor.
The output shapes for each document processor are described below.
Extract
Example Extract Output
{
"object": "document_processor_run",
"id": "dpr_1234",
"edited": true,
"type": "EXTRACT",
// see ExtractConfig type below
"config": {
"fieldsToExtract": [
{
"id": "field1",
"type": "boolean",
"fieldName": "is_invoice",
"description": "Is this an invoice?"
}
]
},
// see ExtractOutput type below
"initialOutput": {
"is_invoice": {
"id": "field1",
"type": "boolean",
"value": true,
"confidence": 1,
"references": []
}
},
// see ExtractOutput type below
"reviewedOutput": {
"is_invoice": {
"id": "field1",
"type": "boolean",
"value": false,
"confidence": 1,
"references": []
}
},
// see ExtractOutput type below
"output": {
"is_invoice": {
"id": "field1",
"type": "boolean",
"value": false,
"confidence": 1,
"references": []
}
}
}
Extract output and config subfield types
/**
* The output subfield type when type=EXTRACT
* Applies to "initialOutput", "reviewedOutput", and "output" subfields of the document_processor_run object.
*/
type ExtractOutput = {
[key: string]: ExtractFieldResult;
}
type ExtractFieldResult = {
id: string;
value: string | number | Currency | boolean | ExtractValueArray | ExtractValueObject | null;
confidence: number;
page?: number;
type: "string" | "number" | "currency" | "boolean" | "date" | "array" | "object" | "null";
/**
* Only set when type=array or type=object to provide structure of sub fields.
* Useful for dynamically parsing nested data.
*/
schema?: ExtractFieldSchemaValue[];
references?: ExtractFieldResultReference[];
}
type Currency = {
amount: number;
iso_4217_currency_code: string;
};
type ExtractValueArray = Array<ExtractValueObject>;
type ExtractValueObject = {
[key: string]: string | number | Currency | boolean | null;
};
type ExtractFieldResultReference = {
/* The field id. When nested for arrays, this is the index of the row number */
id: string;
/* The field name */
fieldName: string;
/* The page number (starting at 1) that this bounding box is from */
page: number;
/**
* Array of bounding box references for this field.
* There can be multiple is the extraction result was drawn from multiple distinct sources on the page.
*/
boundingBoxes: BoundingBox[];
};
/* See the Bounding boxes guide for information on how to use/interpret this data */
type BoundingBox = {
/* The left most position of the bounding box */
left: number;
/* The top most position of the bounding box */
top: number;
/* The right most position of the bounding box */
right: number;
/* The bottom most position of the bounding box */
bottom: number;
};
/**
* The config subfield type when type=EXTRACT
* Applies to "config" subfield of the document_processor_run object.
*/
type ExtractConfig = {
fieldsToExtract: Array<ExtractFieldConfig>;
};
type ExtractFieldConfig = {
id: string;
fieldName: string;
description: string;
type: "string" | "number" | "currency" | "boolean" | "date" | "any[]" | "object" | "null";
schema?: ExtractFieldSchemaValue[];
}
type ExtractFieldSchemaValue = {
id: string;
fieldName: string;
description: string;
type: "string" | "number" | "currency" | "boolean" | "date" | "null";
}
Classify
Example Classify Output
{
"object": "document_processor_run",
"id": "dpr_1234",
"edited": true,
"type": "EXTRACT",
// see ClassifyConfig type below
"config": {
"classifications": [
{
"id": "classification1",
"type": "other",
"description": "Use the `other` document type when the provided document can not clearly be classified into one of the described classifications."
},
{
"id": "classification2",
"type": "invoice",
"description": "an invoice"
}
]
},
// see ClassifyOutput type below
"initialOutput": {
"id": "classification2",
"type": "invoice",
"confidence": 0.95,
"observation": "The document contains information about invoice number, order number, invoice date, due date, total due, and payment details. It matches the description of an invoice."
},
// see ClassifyOutput type below
"output": {
"id": "classification2",
"type": "invoice",
"confidence": 0.95,
"observation": "The document contains information about invoice number, order number, invoice date, due date, total due, and payment details. It matches the description of an invoice."
}
}
CLASSIFY output and config subfield types
/**
* The output subfield type when type=CLASSIFY
* Applies to "initialOutput" and "output" subfields of the document_processor_run object.
*/
type ClassifyOutput = {
id: string;
type: string;
confidence: number;
observation: string;
};
/**
* The config subfield type when type=CLASSIFY
* Applies to "config" subfield of the document_processor_run object.
*/
type ClassifyConfig = {
classifications: Classification[];
}
type Classification = {
id: string;
type: string;
description: string;
};
Instruct QA
Example Instruct QA Output
{
"object": "document_processor_run",
"id": "dpr_1234",
"edited": true,
"type": "INSTRUCT_QA",
// see InstructQAConfig type below
"config": {
"id": "instructqa1",
"query": "buyer",
"context": "- The buyer name is the name of the company or individual that is purchasing the goods or services.\n- The buyer name is often found at the top of the invoice, near the seller's name.",
"question": "Does this invoice contain a name?",
"documentType": "invoice"
},
// see InstructQAOutput type below
"initialOutput": {
"answer": "yes",
"confidence": 1,
"notes": [
"The invoice contains information about the sender (DEMO - Sliced Invoices) and the recipient (Test Business).",
"The invoice contains names of the sender and recipient."
],
"references": [
{
"page": 1
}
]
},
// see InstructQAOutput type below
"output": {
"answer": "yes",
"confidence": 1,
"notes": [
"The invoice contains information about the sender (DEMO - Sliced Invoices) and the recipient (Test Business).",
"The invoice contains names of the sender and recipient."
],
"references": [
{
"page": 1
}
]
}
}
INSTRUCT_QA output and config subfield types
/**
* The output subfield type when type=INSTRUCT_QA
* Applies to "initialOutput" and "output" subfields of the document_processor_run object.
*/
export type InstructQAOutput = {
answer: InstructQAAnswer;
confidence: number;
notes: string[];
references: PageReference[];
};
/**
* Note: Bounding box is not yet supported for INSTRUCT_QA, but will be in the future.
*/
type PageReference = {
page: number;
};
/**
* The config subfield type when type=INSTRUCT_QA
* Applies to "config" subfield of the document_processor_run object.
*/
type InstructQAConfig = {
id: string;
query: string;
context: string;
question: string;
documentType: string;
}