OCR fails if metadata extractor is running

[gabriel-piles]

When Uwazi receives OCR results with metadata extraction enabled, the following error occurs:

TypeError: Cannot read properties of undefined (reading 'template')
at /uwazi/app/api/suggestions/eventListeners.ts:138:10
at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
at async Promise.all (index 0)
at async EventsBus.emit (uwazi/app/api/eventsbus/EventsBus.ts:19:7)
at async Object.save (uwazi/app/api/files/files.ts:49:7)
at async processPDF (uwazi/app/api/files/processDocument.ts:14:18)
at async processFiles (uwazi/app/api/services/ocr/OcrManager.ts:107:22)
at async uwazi/app/api/services/ocr/OcrManager.ts:138:7
at async Object.processResults (uwazi/app/api/services/ocr/OcrManager.ts:126:3)
at async TaskManager.checkForResults (uwazi/app/api/services/tasksmanager/TaskManager.ts:119:9)
at async Repeater.start (uwazi/app/api/utils/Repeater.js:24:7)

To Reproduce

1. Start Uwazi with

yarn blank state
export EXTERNAL_SERVICES=true
yarn hot

3. Activate metadata extractor and ocr flags

  "ocrServiceEnabled": true
  "features": {
    "ocr": {
      "url": "http://localhost:5051"
    },
    "metadataExtraction": {
      "url": "http://localhost:5056"
    },
    "metadata-extraction": true
  },

4. Start OCR service

git clone https://github.com/huridocs/pdf-document-layout-analysis-async.git
cd pdf-document-layout-analysis-async
make start

5. Send a PDF to OCR

[RafaPolit]

@gabriel-piles is this prior to the incorporation of OCR into the other services? Or is this still an issue?

[gabriel-piles]

@RafaPolit yes, it is prior to the OCR server change.

And, it is still an issue.