Tesseract: Stop throwing an exception for images with no text

The Tesseract wrapper library we use throws an exception when the
command doesn't produce any output (i.e. an output file of size
0). We can't just catch that exception because it's also used for
other things, so this adds a check for the contents of the message
and returns a warning instead.

Bug: T373161

i18n/en.json

@@ -51,6 +51,7 @@
     "tesseract-psm-12": "Sparse text with OSD.",
     "tesseract-psm-13": "Raw line. Treat the image as a single text line, bypassing hacks that are Tesseract-specific.",
     "tesseract-param-error": "The '$1' option with a value of $2 is not supported by Tesseract. Maximum value: $3",
+    "tesseract-no-text-error": "The Tesseract engine did not return any text for this image.",
     "tesseract-internal-error": "The tesseract engine returned an internal error.",
     "transkribus-language-code": "Language Model",
     "transkribus-unauthorized-error": "Error Code '$1' :: The request is not authorized",

i18n/qqq.json

@@ -58,6 +58,7 @@
 	"tesseract-psm-12": "Form option for Tesseract page segmentation mode.",
 	"tesseract-psm-13": "Form option for Tesseract page segmentation mode.",
 	"tesseract-param-error": "Error message displayed when invalid values for Tesseract options are submitted.\n\nParameters:\n* $1 – the form label for the option. Currently, the only possible value here is {{msg-wm|wikimedia-ocr-tesseract-psm-label}}.\n* $2 – The value that was given.\n* $3 – the maximum value for the option (this will be an integer).",
+	"tesseract-no-text-error": "Error message displayed when Tesseract returned no text (e.g. for an image that has no text in it).",
 	"tesseract-internal-error": "Generic error message displayed when the tesseract command fails.",
 	"transkribus-language-code": "Form label for the 'Language Model' field for the Transkribus engine",
 	"transkribus-unauthorized-error": "Error message displayed when Transkribus access token has expired or login credentials are invalid.",

src/Engine/TesseractEngine.php

@@ -7,6 +7,7 @@
 use Krinkle\Intuition\Intuition;
 use Symfony\Contracts\HttpClient\HttpClientInterface;
 use thiagoalessio\TesseractOCR\TesseractOCR;
+use thiagoalessio\TesseractOCR\UnsuccessfulCommandException;
 
 class TesseractEngine extends EngineBase {
 
@@ -69,7 +70,17 @@ public function getResult(
 		// so we have to load this one manually. We only process one image at a time, so don't benefit from
 		// multiple threads. See https://github.com/tesseract-ocr/tesseract/issues/898 for some more info.
 		putenv( 'OMP_THREAD_LIMIT=1' );
-		$text = $this->ocr->run();
+		try {
+			$text = $this->ocr->run();
+		} catch ( UnsuccessfulCommandException $e ) {
+			// An UnsuccessfulCommandException is thrown when there's no output, but that's not an
+			// actual error so we check for it here and just show a warning. The same exception class
+			// is also used for other things, hence the message check here.
+			if ( strpos( $e->getMessage(), 'The command did not produce any output' ) !== false ) {
+				return new EngineResult( '', [ $this->intuition->msg( 'tesseract-no-text-error' ) ] );
+			}
+			throw $e;
+		}
 
 		$warnings = $invalidLangs ? [ $this->getInvalidLangsWarning( $invalidLangs ) ] : [];
 		return new EngineResult( $text, $warnings );