Error when trying to create TSV file

[Gurey]

Trying to create a TSV gives me an error.

endpoint: localhost:9292/ocr

with payload:

{
  "img_url": "http://bit.ly/ocrimage",
  "engine": "tesseract",
  "engine_args": {
    "config_vars": {
		"tessedit_create_tsv": "1",
		"tessedit_pageseg_mode": "1"
    }
  }
}

Gives me:
Error processing image url: . Error: Could not find outfile. Basename: /tmp/785ad93e-2721-4fd1-7892-b78ffc442ae0 Extensions: [txt hocr]

tesseract official TSV config:
https://github.com/tesseract-ocr/tesseract/blob/master/tessdata/configs/tsv

[tleyden]

Does it work if you remove the "config_vars"?

The relevant code is here:

open-ocr/tesseract_engine.go

Lines 233 to 248 in eb6a931

	func findOutfile(outfileBaseName string, fileExtensions []string) (string, error) {
	for _, fileExtension := range fileExtensions {
	outFile := fmt.Sprintf("%v.%v", outfileBaseName, fileExtension)
	logg.LogTo("OCR_TESSERACT", "checking if exists: %v", outFile)
	if _, err := os.Stat(outFile); err == nil {
	return outFile, nil
	}
	}
	return "", fmt.Errorf("Could not find outfile. Basename: %v Extensions: %v", outfileBaseName, fileExtensions)
	}

I wonder if tesseract is writing to a file extension that OpenOCR isn't expecting.. looks like OpenOCR looks for .txt or .hocr

[Gurey]

Hi!
thank you for getting back to me.

these 2 requests works fine:

{
  "img_url": "http://bit.ly/ocrimage",
  "engine": "tesseract",
  "engine_args": {
    "config_vars": {
    	"tessedit_create_hocr": "1",
	"tessedit_pageseg_mode": "1"
    }
  },
  "psm": 3 
}

{
  "img_url": "http://bit.ly/ocrimage",
  "engine": "tesseract",
  "engine_args": {
    "config_vars": {}
  },
  "psm": 3 
}

I just tried to run tesseract in my terminal with the command tesseract pic.jpg out tsv and that gives me a file named out.tsv

[tleyden]

Ok I think the fix is pretty easy, it just needs to look for files with .tsv extension. Are you able to submit a PR?

[luzanikita]

#126