Remove tesseract engine mode selection (#46)

Bug: T285262
wikimedia · Jun 22, 2021 · adb07c3 · adb07c3
1 parent 1d8d923
commit adb07c3
Show file tree

Hide file tree

Showing 5 changed files with 4 additions and 44 deletions.
diff --git a/i18n/en.json b/i18n/en.json
@@ -37,11 +37,6 @@
     "tesseract-psm-11": "Sparse text. Find as much text as possible in no particular order.",
     "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-oem-label": "OCR Engine modes",
-    "tesseract-oem-0": "Legacy engine only.",
-    "tesseract-oem-1": "Neural nets LSTM engine only.",
-    "tesseract-oem-2": "Legacy + LSTM engines.",
-    "tesseract-oem-3": "Default, based on what is available.",
     "tesseract-param-error": "The '$1' option with a value of $2 is not supported by Tesseract. Maximum value: $3",
     "tesseract-internal-error": "The tesseract engine returned an internal error."
 }
diff --git a/i18n/qqq.json b/i18n/qqq.json
@@ -41,11 +41,6 @@
 	"tesseract-psm-11": "Form option for Tesseract page segmentation mode.",
 	"tesseract-psm-12": "Form option for Tesseract page segmentation mode.",
 	"tesseract-psm-13": "Form option for Tesseract page segmentation mode.",
-	"tesseract-oem-label": "Form label for the Tesseract OCR engine mode.",
-	"tesseract-oem-0": "Form option for Tesseract OCR engine mode.",
-	"tesseract-oem-1": "Form option for Tesseract OCR engine mode.",
-	"tesseract-oem-2": "Form option for Tesseract OCR engine mode.",
-	"tesseract-oem-3": "Form option for Tesseract OCR engine 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. This will either be {{msg-wm|tesseract-psm-label}} or {{msg-wm|tesseract-oem-label}}.\n* $2 – The value that was given.\n* $3 – the maximum value for the option (this will be an integer).",
 	"tesseract-internal-error": "Generic error message displayed when the tesseract command fails."
 }
diff --git a/src/Controller/OcrController.php b/src/Controller/OcrController.php
@@ -47,7 +47,6 @@ class OcrController extends AbstractController
         'engine' => self::DEFAULT_ENGINE,
         'langs' => [],
         'psm' => TesseractEngine::DEFAULT_PSM,
-        'oem' => TesseractEngine::DEFAULT_OEM,
     ];
 
     /** @var string */
@@ -102,16 +101,14 @@ private function setup(): void
      */
     private function setEngineOptions(): void
     {
-        // These are always set, even if Tesseract isn't initially chosen as the engine
-        // because we want these defaults set if the user changes the engine to Tesseract.
+        // This is always set, even if Tesseract isn't initially chosen as the engine
+        // because we want the default set if the user changes the engine to Tesseract.
         static::$params['psm'] = (int)$this->request->query->get('psm', (string)static::$params['psm']);
-        static::$params['oem'] = (int)$this->request->query->get('oem', (string)static::$params['oem']);
 
-        // Apply the settings to the Engine itself. This is only done when Tesseract is chosen
-        // because these setters don't exist for the GoogleCloudVisionEngine.
+        // Apply the tesseract-specific settings
+        // NOTE: Intentionally excluding `oem`, see T285262
         if (TesseractEngine::getId() === static::$params['engine']) {
             $this->engine->setPsm(static::$params['psm']);
-            $this->engine->setOem(static::$params['oem']);
         }
     }
 
@@ -216,7 +213,6 @@ private function getText(): string
                 static::$params['engine'],
                 implode('|', static::$params['langs']),
                 static::$params['psm'],
-                static::$params['oem'],
             ]
         ));
 

diff --git a/src/Engine/TesseractEngine.php b/src/Engine/TesseractEngine.php
@@ -23,12 +23,6 @@ class TesseractEngine extends EngineBase
     /** @var int Default value for page segmentation mode. */
     public const DEFAULT_PSM = 3;
 
-    /** @var int Maximum value for OCR engine mde. */
-    public const MAX_OEM = 3;
-
-    /** @var int Default value for OCR engine mode. */
-    public const DEFAULT_OEM = 3;
-
     /**
      * TesseractEngine constructor.
      * @param HttpClientInterface $httpClient
@@ -98,16 +92,6 @@ public function setPsm(int $psm): void
         $this->ocr->psm($psm);
     }
 
-    /**
-     * Set the OCR engine mode.
-     * @param int $oem
-     */
-    public function setOem(int $oem): void
-    {
-        $this->validateOption('oem', $oem, self::MAX_OEM);
-        $this->ocr->oem($oem);
-    }
-
     /**
      * Validates the given option.
      * @param string $option

diff --git a/templates/_tesseract_options.html.twig b/templates/_tesseract_options.html.twig
@@ -12,14 +12,4 @@
             {% endfor %}
         </select>
     </div>
-    <div class="form-group">
-        <label for="oem">{{ msg('tesseract-oem-label') }}</label>
-        <select name="oem" id="oem" class="form-control">
-            {# The following messages are used:
-             #   'tesseract-oem-0', 'tesseract-oem-1', 'tesseract-oem-2', 'tesseract-oem-3' #}
-            {% for i in 0..3 %}
-                <option value="{{ i }}" {% if i == oem %}selected="selected"{% endif %}>{{ msg('tesseract-oem-' ~ i) }}</option>
-            {% endfor %}
-        </select>
-    </div>
 </fieldset>