✨ dinglehopper: Add OCR-D parameter to choose TextEquiv level

qurator/dinglehopper/extracted_text.py

@@ -148,10 +148,7 @@ class ExtractedText:
     @property
     def text(self):
         if self._text is not None:
-            if self._text == '':
-                return None
-            else:
-                return self._text
+            return self._text
         else:
             return self.joiner.join(s.text for s in self.segments)
 

qurator/dinglehopper/ocr_files.py

@@ -82,7 +82,7 @@ def page_extract(tree, *, textequiv_level='region'):
             regions.append(ExtractedText.from_text_segment(region, nsmap, textequiv_level=textequiv_level))
 
     # Filter empty region texts
-    regions = [r for r in regions if r.text is not None]
+    regions = [r for r in regions if r.text != '']
 
     return ExtractedText(None, regions, '\n', None)
 

qurator/dinglehopper/ocrd-tool.json

@@ -22,6 +22,12 @@
           "type": "boolean",
           "default": true,
           "description": "Enable/disable metrics and green/red"
+        },
+        "textequiv_level": {
+          "type": "string",
+          "enum": ["region", "line"],
+          "default": "region",
+          "description": "PAGE XML hierarchy level to extract the text from"
         }
       }
     }

qurator/dinglehopper/ocrd_cli.py

@@ -32,6 +32,7 @@ class OcrdDinglehopperEvaluate(Processor):
         log = getLogger('processor.OcrdDinglehopperEvaluate')
 
         metrics = self.parameter['metrics']
+        textequiv_level = self.parameter['textequiv_level']
         gt_grp, ocr_grp = self.input_file_grp.split(',')
         for n, page_id in enumerate(self.workspace.mets.physical_pages):
             gt_file = next(self.workspace.mets.find_files(fileGrp=gt_grp, pageId=page_id))
@@ -52,7 +53,8 @@ class OcrdDinglehopperEvaluate(Processor):
                     gt_file.local_filename,
                     ocr_file.local_filename,
                     report_prefix,
-                    metrics=metrics
+                    metrics=metrics,
+                    textequiv_level=textequiv_level
             )
 
             # Add reports to the workspace