mindee/src/imageOperations/invoiceSplitterExtractor/invoiceSplitterExtractor.js

Mindee Client Library for Node.js

"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
exports.extractInvoices = extractInvoices;
const pdf_lib_1 = require("@cantoo/pdf-lib");
const errors_1 = require("../../errors");
const product_1 = require("../../product");
const extractedInvoiceSplitterImage_1 = require("./extractedInvoiceSplitterImage");
async function splitPdf(pdfDoc, invoicePageGroups) {
    if (invoicePageGroups.length === 0) {
        return [];
    }
    const generatedPdfs = [];
    for (let i = 0; i < invoicePageGroups.length; i++) {
        const subdocument = await pdf_lib_1.PDFDocument.create();
        const fullIndexes = [];
        for (let j = invoicePageGroups[i][0]; j <= invoicePageGroups[i][invoicePageGroups[i].length - 1]; j++) {
            fullIndexes.push(j);
        }
        const copiedPages = await subdocument.copyPages(pdfDoc, fullIndexes);
        copiedPages.map((page) => {
            subdocument.addPage(page);
        });
        const subdocumentBytes = await subdocument.save();
        generatedPdfs.push(new extractedInvoiceSplitterImage_1.ExtractedInvoiceSplitterImage(subdocumentBytes, [invoicePageGroups[i][0], invoicePageGroups[i][invoicePageGroups[i].length - 1]]));
    }
    return generatedPdfs;
}
async function getPdfDoc(inputFile) {
    await inputFile.init();
    if (!inputFile.isPdf()) {
        throw new errors_1.MindeeMimeTypeError("Invoice Splitter is only compatible with pdf documents.");
    }
    const pdfDoc = await pdf_lib_1.PDFDocument.load(inputFile.fileObject, {
        ignoreEncryption: true,
        password: ""
    });
    if (pdfDoc.getPageCount() < 2) {
        throw new errors_1.MindeeError("Invoice Splitter is only compatible with multi-page-pdf documents.");
    }
    return pdfDoc;
}
/**
 * Extracts & cuts the pages of a main document invoice according to the provided indexes.
 *
 * @param inputFile File to extract sub-invoices from.
 * @param indexes List of indexes to cut the document according to.
 *  Can be provided either as a InvoiceSplitterV1 inference, or a direct list of splits.
 * @param strict If set to true, doesn't cut pages where the API isn't 100% confident.
 * @returns A promise of extracted images, as an array of ExtractedInvoiceSplitterImage.
 */
async function extractInvoices(inputFile, indexes, strict = false) {
    if (!indexes) {
        throw new errors_1.MindeeError("No possible receipts candidates found for MultiReceipts extraction.");
    }
    let customIndexes = [];
    if (indexes instanceof product_1.InvoiceSplitterV1) {
        indexes.prediction.invoicePageGroups.map((invoicePageGroup) => {
            if (!strict || invoicePageGroup.confidence === 1) {
                customIndexes.push(invoicePageGroup.pageIndexes ?? []);
            }
        });
    }
    else {
        customIndexes = indexes;
    }
    const pdfDoc = await getPdfDoc(inputFile);
    const pageCount = pdfDoc.getPageCount();
    customIndexes.forEach((pageGroup) => {
        pageGroup.forEach((index) => {
            if (index >= pageCount) {
                throw new errors_1.MindeeError(`Given index ${index} doesn't exist in page range (0-${pageCount - 1})`);
            }
        });
    });
    return await splitPdf(pdfDoc, customIndexes);
}