Chinese OCR failed - Node.js

Report · Dec 21, 2020

I tried to use OCR of PDF Tools API to identify a chinese pdf, but something went wrong.

I modified OCRLocale to 'ZH_CN' according to document 'OCR with explicit language', like this:

const options = new PDFToolsSdk.OCR.options.OCROptions.Builder()
        .withOcrType(PDFToolsSdk.OCR.options.OCRSupportedType.SEARCHABLE_IMAGE_EXACT)
        .withOcrLang(PDFToolsSdk.OCR.options.OCRSupportedLocale.ZH_CN)
        .build();

I run the code but an error occurs, the error info:

2020-12-22T10:40:11.252:[DEBUG]: Resolved request uri : https://senseicore-ue1.adobe.io/services/v2/status/owFg2AYgawPrJCHU54YF7e4lWUj2teJVException encountered while executing operation ServiceApiError: vector::_M_range_check: __n (which is 0) >= this->size() (which is 0)
    at E:\project\test\adobe-dc-pdf-tools-sdk-node-samples\node_modules\@adobe\documentservices-pdftools-node-sdk\src\internal\api\cpf-api.js:170:20
    at IncomingForm.<anonymous> (E:\project\test\adobe-dc-pdf-tools-sdk-node-samples\node_modules\formidable\lib\incoming_form.js:107:9)   
    at IncomingForm.emit (events.js:310:20)
    at IncomingForm._maybeEnd (E:\project\test\adobe-dc-pdf-tools-sdk-node-samples\node_modules\formidable\lib\incoming_form.js:557:8)     
    at JSONParser.parser.onEnd (E:\project\test\adobe-dc-pdf-tools-sdk-node-samples\node_modules\formidable\lib\incoming_form.js:532:10)   
    at JSONParser.end (E:\project\test\adobe-dc-pdf-tools-sdk-node-samples\node_modules\formidable\lib\json_parser.js:29:8)
    at IncomingMessage.<anonymous> (E:\project\test\adobe-dc-pdf-tools-sdk-node-samples\node_modules\formidable\lib\incoming_form.js:132:30)
    at IncomingMessage.emit (events.js:322:22)
    at endReadableNT (_stream_readable.js:1187:12)
    at processTicksAndRejections (internal/process/task_queues.js:84:21) {
  requestTrackingId: 'owFg2AYgawPrJCHU54YF7e4lWUj2teJV',
  statusCode: 500
}

The same problem also appears on the OCRSupportedLocale.ZH_HK.

Please help to solve this problem, thank you.

Report · Dec 22, 2020

Thanks for sharing the details, it is a known issue and it is due to OCR language mapping in SDK. We will update here once the issue is fixed, appreciate your patience in the meantime !!!