TLDR-861 remove orient cell params

ispras · Dec 13, 2024 · 2a3d0e6 · 2a3d0e6
1 parent ccf6d15
commit 2a3d0e6
Show file tree

Hide file tree

Showing 6 changed files with 4 additions and 37 deletions.
diff --git a/dedoc/api/api_args.py b/dedoc/api/api_args.py
@@ -22,9 +22,6 @@ class QueryParameters:
     # tables handling
     need_pdf_table_analysis: str = Form("true", enum=["true", "false"], description="Enable table recognition for pdf")
     table_type: str = Form("", description="Pipeline mode for table recognition")
-    orient_analysis_cells: str = Form("false", enum=["true", "false"], description="Enable analysis of rotated cells in table headers")
-    orient_cell_angle: str = Form("90", enum=["90", "270"],
-                                  description='Set cells orientation in table headers, "90" means 90 degrees counterclockwise cells rotation')
 
     # pdf handling
     pdf_with_text_layer: str = Form("auto_tabby", enum=["true", "false", "auto", "auto_tabby", "tabby"],

diff --git a/dedoc/api/web/index.html b/dedoc/api/web/index.html
@@ -101,21 +101,13 @@ <h4>Attachments handling</h4>
 
         <div class="parameters">
             <h4>Tables handling </h4>
-            <details><summary>need_pdf_table_analysis, orient_analysis_cells, orient_cell_angle</summary>
+            <details><summary>need_pdf_table_analysis</summary>
                 <br>
                 <p>
                     <label>
                         <input type="hidden" name="need_pdf_table_analysis" value="false">
                         <input type="checkbox" name="need_pdf_table_analysis" value="true" checked> need_pdf_table_analysis</label>
                 </p>
-
-                <p>
-                    <label><input name="orient_analysis_cells" type="checkbox" value="true"> orient_analysis_cells</label>
-                </p>
-
-                <p>
-                    <label>orient_cell_angle <input name="orient_cell_angle" type="number" size="5" value="90"></label>
-                </p>
             </details>
         </div>
 

diff --git a/dedoc/readers/pdf_reader/data_classes/tables/scantable.py b/dedoc/readers/pdf_reader/data_classes/tables/scantable.py
@@ -33,9 +33,6 @@ def check_on_cell_instance(self) -> bool:
             return False
         return True
 
-    def to_table(self) -> Table:
-        return super()
-
     @staticmethod
     def get_cells_text(cells: List[List[CellWithMeta]]) -> List[List[str]]:
         return [[cell.get_text() for cell in row] for row in cells]

diff --git a/dedoc/readers/pdf_reader/pdf_base_reader.py b/dedoc/readers/pdf_reader/pdf_base_reader.py
@@ -87,12 +87,12 @@ def read(self, file_path: str, parameters: Optional[dict] = None) -> Unstructure
         )
 
         lines, scan_tables, attachments, warnings, metadata = self._parse_document(file_path, params_for_parse)
-        tables = [scan_table.to_table() for scan_table in scan_tables]
+        # tables = [scan_table.to_table() for scan_table in scan_tables]
 
         if params_for_parse.with_attachments and self.attachment_extractor.can_extract(file_path):
             attachments += self.attachment_extractor.extract(file_path=file_path, parameters=parameters)
 
-        result = UnstructuredDocument(lines=lines, tables=tables, attachments=attachments, warnings=warnings, metadata=metadata)
+        result = UnstructuredDocument(lines=lines, tables=scan_tables, attachments=attachments, warnings=warnings, metadata=metadata)
         return self._postprocess(result)
 
     def _parse_document(self, path: str, parameters: ParametersForParseDoc) -> (

diff --git a/dedoc/readers/pdf_reader/pdf_txtlayer_reader/pdf_tabby_reader.py b/dedoc/readers/pdf_reader/pdf_txtlayer_reader/pdf_tabby_reader.py
@@ -136,9 +136,7 @@ def __extract(self, path: str, parameters: dict, warnings: List[str], tmp_dir: s
         mp_tables = self.table_recognizer.convert_to_multipages_tables(all_scan_tables, lines_with_meta=all_lines)
         all_lines = self.linker.link_objects(lines=all_lines, tables=mp_tables, images=all_attached_images)
 
-        tables = [scan_table.to_table() for scan_table in mp_tables]
-
-        return all_lines, tables, all_attached_images, document_metadata
+        return all_lines, mp_tables, all_attached_images, document_metadata
 
     def __save_gost_frame_boxes_to_json(self, first_page: Optional[int], last_page: Optional[int], page_count: int, path: str, tmp_dir: str) -> str:
         from joblib import Parallel, delayed

diff --git a/dedoc/utils/parameter_utils.py b/dedoc/utils/parameter_utils.py
@@ -33,13 +33,6 @@ def get_param_document_type(parameters: Optional[dict]) -> str:
     return document_type
 
 
-def get_param_orient_analysis_cells(parameters: Optional[dict]) -> bool:
-    if parameters is None:
-        return False
-    orient_analysis_cells = str(parameters.get("orient_analysis_cells", "False")).lower() == "true"
-    return orient_analysis_cells
-
-
 def get_param_with_attachments(parameters: Optional[dict]) -> bool:
     if parameters is None:
         return False
@@ -80,16 +73,6 @@ def get_param_need_binarization(parameters: Optional[dict]) -> bool:
     return need_binarization
 
 
-def get_param_orient_cell_angle(parameters: Optional[dict]) -> int:
-    if parameters is None:
-        return 90
-
-    orient_cell_angle = str(parameters.get("orient_cell_angle", "90"))
-    if orient_cell_angle == "":
-        orient_cell_angle = "90"
-    return int(orient_cell_angle)
-
-
 def get_param_is_one_column_document(parameters: Optional[dict]) -> Optional[bool]:
     if parameters is None:
         return None