Update mdr_pdf_parser.py

mdr_pdf_parser.py

@@ -1855,7 +1855,7 @@ class MDROcrEngine:
                 det_model_dir=paths["det"],
                 cls_model_dir=paths["cls"],
                 rec_model_dir=paths["rec"],
-                rec_char_dict_path=None,
+                rec_char_dict_path=paths["keys"],
                 # much lower thresholds so we actually get some candidate masks:
                 det_db_thresh=0.1,
                 det_db_box_thresh=0.3,
@@ -3084,9 +3084,12 @@ class MDRDocumentIterator:
     def iterate_sections(self, params: MDRProcessingParams) -> Generator[
         tuple[int, MDRExtractionResult, list[MDRLayoutElement]], None, None]:
         """Yields page index, extraction result, and content layouts for each requested page."""
+        # In MDRDocumentIterator.iterate_sections
         for res, sec in self._process_and_link_sections(params):
-            framework = set(sec.find_framework_elements());
-            content = [l for l in res.layouts if l not in framework];
+            # Get the IDs of the framework elements
+            framework_element_ids = {id(fw_el) for fw_el in sec.find_framework_elements()}
+            # Filter content layouts by checking if their ID is not in the set of framework element IDs
+            content = [l for l in res.layouts if id(l) not in framework_element_ids]
             yield sec.page_index, res, content
 
     def _process_and_link_sections(self, params: MDRProcessingParams) -> Generator[