TLDR-474 remove insert_table parameter (#339)

* TLDR-474 remove insert_table parameter

* TLDR-474 remove is_inserted attribute

dedoc/api/api_args.py

@@ -15,7 +15,6 @@ class QueryParameters(BaseModel):
     return_base64: Optional[str]
     attachments_dir: Optional[str]
 
-    insert_table: Optional[str]
     need_pdf_table_analysis: Optional[str]
     table_type: Optional[str]
     orient_analysis_cells: Optional[str]
@@ -48,7 +47,6 @@ def __init__(self,
                  attachments_dir: Optional[str] = Body(description="path to the directory where to save files' attachments", default=None),  # noqa
 
                  # tables handling
-                 insert_table: Optional[str] = Body(description="Insert table into the result tree's content or not. Default: 'false'", default=None),  # noqa
                  need_pdf_table_analysis: Optional[str] = Body(description="include a table analysis into pdfs. Default: 'true'", default=None),  # noqa
                  table_type: Optional[str] = Body(description="a pipeline mode for a table recognition. Default: ''", default=None),  # noqa
                  orient_analysis_cells: Optional[str] = Body(description="a table recognition option enables analysis of rotated cells in table headers. Default: 'false'", default=None),  # noqa
@@ -83,7 +81,6 @@ def __init__(self,
         self.return_base64: str = return_base64 or "false"
         self.attachments_dir: str = attachments_dir
 
-        self.insert_table: str = insert_table or "false"
         self.need_pdf_table_analysis: str = need_pdf_table_analysis or "true"
         self.table_type: str = table_type or ""
         self.orient_analysis_cells: str = orient_analysis_cells or "false"

dedoc/api/static/html_eng/form_input.html

@@ -39,7 +39,6 @@ <h2>Structure Document Recognition</h2>
                         </select>
                     </label> return_format
                 </p>
-                <p><label><input type="checkbox" name="insert_table" value="true"> insert_table</label></p>
                 <p><label><input type="checkbox" name="handle_invisible_table" value="true"> handle_invisible_table</label></p>
                 <p>pages <input name="pages" type="text" size="8" value=":"> </p>
 

dedoc/api/static/html_eng/format_description.html

@@ -71,7 +71,6 @@ <h4 id="TableMetadata"> TableMetadata</h4>
 <ol>
     <li><strong>uid</strong>: <a> str (required field) </a> - unique identifier. </li>
     <li><strong>page_id</strong>: <a> integer </a> (optional field) - page number on which the table begins. Can be null.</li>
-    <li><strong>is_inserted</strong>: <a> bool </a> (optional field) - was table inserted into document.</li>
 </ol>
 
 <h3 id="TreeNode"> TreeNode</h3>

dedoc/api/static/html_eng/info.html

@@ -31,8 +31,6 @@ <h2>How to use</h2>
     <ol>
         <li><strong>language</strong>: string - document recognition language. The default value is "rus+eng". Available values: "rus+eng", "rus", "eng".</li>
         <li><strong>with_attachments</strong>: boolean - option including analysis of attached files. The option is False by default. Available values: True, False.</li>
-        <li><strong>insert_table</strong>: boolean - this option enables embedding the table in the document tree.
-             The option is False by default. Available values: True, False.</li>
         <li><strong>return_format</strong>: str - an option to return the response in pretty_json, html, json or tree form.
             The default value is json. Use the pretty_json, tree and html format for debug only.<br>
         Warning: html-format is used only for viewing the recognition result (in a readable form).

dedoc/api/static/html_rus/form_input.html

@@ -45,9 +45,6 @@ <h2>Распознавание структуры документа</h2>
                     </label> with_attachments</p>
                 <p>
                 <p>
-                    <label>
-                    <input type="checkbox" name="insert_table" value=true>
-                    </label> insert_table </p>
                 <p>
                     <label>
                         <select name="document_type">

dedoc/api/static/html_rus/format_description.html

@@ -72,7 +72,6 @@ <h4 id="TableMetadata"> TableMetadata. Метаинформация таблиц
 <ol>
     <li><strong>uid</strong>: <a> str (обязательное поле) </a> - уникальный идентификатор таблицы.</li>
     <li><strong>page_id</strong>: <a> int </a> (необязательное поле) - номер страницы на которой начинается таблица.</li>
-    <li><strong>is_inserted</strong>: <a> bool </a> (необязательное поле) - была ли таблица встроена в тело документа.</li>
 </ol>
 
 <h3 id="TreeNode"> TreeNode. Древовидная структура документа.</h3>

dedoc/api/static/html_rus/info.html

@@ -31,8 +31,6 @@ <h2>Как использовать</h2>
     <ol>
         <li><strong>language</strong>: str - язык рапознавания документа. По-умолчанию установлено значение "rus+eng".
         Доступные значения: "rus+eng", "rus", "eng".</li>
-        <li><strong>insert_table</strong>: boolean - опция включает встраивание таблицы в документное дерево.
-            По-умолчанию установлено значение False. Доступные значения True, False.</li>
         <li><strong>with_attachments</strong>: boolean - опция включающая анализ вложенных файлов. По-умолчанию установлено
         значение False. Доступные значения True, False.</li>
         <li><strong>return_format</strong>: str - опция для возврата ответа в html-виде, в виде дерева или в виде json.

dedoc/data_structures/table_metadata.py

@@ -11,23 +11,20 @@ class TableMetadata(Serializable):
     """
     This class holds the information about table unique identifier, rotation angle (if table has been rotated - for images) and so on.
     """
-    def __init__(self, page_id: Optional[int], uid: Optional[str] = None, is_inserted: bool = False, rotated_angle: float = 0.0) -> None:
+    def __init__(self, page_id: Optional[int], uid: Optional[str] = None, rotated_angle: float = 0.0) -> None:
         """
         :param page_id: number of the page where table starts
         :param uid: unique identifier of the table
-        :param is_inserted: indicator if table was already inserted into paragraphs list
         :param rotated_angle: value of the rotation angle by which the table was rotated during recognition
         """
         self.page_id = page_id
         self.uid = str(uuid.uuid1()) if not uid else uid
-        self.is_inserted = is_inserted
         self.rotated_angle = rotated_angle
 
     def to_dict(self) -> dict:
         res = OrderedDict()
         res["uid"] = self.uid
         res["page_id"] = self.page_id
-        res["is_inserted"] = self.is_inserted
         res["rotated_angle"] = self.rotated_angle
         return res
 
@@ -36,6 +33,5 @@ def get_api_dict(api: Api) -> Model:
         return api.model("TableMetadata", {
             "page_id": fields.Integer(readonly=False, description="table start page number"),
             "uid": fields.String(description="table unique id"),
-            "is_inserted": fields.Boolean(description="was the table inserted into document body"),
             "rotated_angle": fields.Float(readonly=False, description="At what angle should the table be rotated to use boxes")
         })

dedoc/readers/docx_reader/data_structures/table.py

@@ -90,7 +90,7 @@ def to_table(self) -> Table:
                 result_row.append(cell)
             result_cells_with_meta.append(result_row)
 
-        return Table(cells=result_cells_with_meta, metadata=TableMetadata(page_id=None, uid=self.uid, is_inserted=False))
+        return Table(cells=result_cells_with_meta, metadata=TableMetadata(page_id=None, uid=self.uid))
 
     def __get_cell_text(self, cell: Tag) -> str:
         cell_text = ""

dedoc/readers/pdf_reader/pdf_txtlayer_reader/pdf_tabby_reader.py

@@ -152,7 +152,7 @@ def __get_tables(self, page: dict, file_hash: str) -> Tuple[List[Table], List[Sc
 
                 result_cells.append(result_row)
             table_bbox = BBox.from_two_points((x_top_left, y_top_left), (x_bottom_right, y_bottom_right))  # noqa TODO add table location into TableMetadata
-            tables.append(Table(cells=result_cells, metadata=TableMetadata(page_id=page_number, is_inserted=False)))
+            tables.append(Table(cells=result_cells, metadata=TableMetadata(page_id=page_number)))
             table_name = file_hash + str(page_number) + str(table_num)
             tables_on_image.append(ScanTable(page_number=page_number, matrix_cells=None, bbox=table_bbox, name=table_name, order=order))
 

dedoc/structure_constructors/structure_constructor_composition.py

@@ -4,7 +4,6 @@
 from dedoc.data_structures.parsed_document import ParsedDocument
 from dedoc.data_structures.unstructured_document import UnstructuredDocument
 from dedoc.structure_constructors.abstract_structure_constructor import AbstractStructureConstructor
-from dedoc.structure_constructors.table_patcher import TablePatcher
 
 
 class StructureConstructorComposition(AbstractStructureConstructor):
@@ -20,19 +19,13 @@ def __init__(self, constructors: Dict[str, AbstractStructureConstructor], defaul
         """
         self.constructors = constructors
         self.default_constructor = default_constructor
-        self.table_patcher = TablePatcher()
 
     def structure_document(self, document: UnstructuredDocument, structure_type: Optional[str] = None, parameters: Optional[dict] = None) -> ParsedDocument:
         """
         Construct the result document structure according to the `structure_type` parameter.
         If `structure_type` is empty string or None the default constructor will be used.
         To get the information about the parameters look at the documentation of :class:`~dedoc.structure_constructors.AbstractStructureConstructor`.
         """
-        parameters = {} if parameters is None else parameters
-
-        if parameters.get("insert_table", "False").lower() == "true":
-            document = self.table_patcher.insert_table(document=document)
-
         if structure_type in self.constructors:
             return self.constructors[structure_type].structure_document(document)
 

docs/source/_static/code_examples/dedoc_return_format.py

@@ -55,16 +55,6 @@ def with_parsed_attachments_example() -> dict:
     return json.loads(result)
 
 
-def with_inserted_table_example() -> dict:
-    with open(filename, "rb") as file:
-        files = {"file": (filename, file)}
-        r = requests.post("http://localhost:1231/upload", files=files, data=dict(insert_table="true"))
-        result = r.content.decode("utf-8")
-
-    assert r.status_code == 200
-    return json.loads(result)
-
-
 if __name__ == "__main__":
     with open("../json_format_examples/basic_example.json", "w") as f:
         json.dump(basic_example(), f, indent=2, ensure_ascii=False)
@@ -80,6 +70,3 @@ def with_inserted_table_example() -> dict:
 
     with open("../json_format_examples/with_parsed_attachments.json", "w") as f:
         json.dump(with_parsed_attachments_example(), f, indent=2, ensure_ascii=False)
-
-    with open("../json_format_examples/with_inserted_table.json", "w") as f:
-        json.dump(with_inserted_table_example(), f, indent=2, ensure_ascii=False)

docs/source/_static/json_format_examples/basic_example.json

@@ -298,7 +298,7 @@
                           "start": 0,
                           "end": 14,
                           "name": "attachment",
-                          "value": "attach_fa1143ae-5d3c-11ee-b518-0242ac120002"
+                          "value": "attach_75af2486-5df1-11ee-bfc1-0242ac120002"
                         }
                       ],
                       "metadata": {
@@ -420,20 +420,19 @@
         "metadata": {
           "uid": "3a327789721e09b3fa6fd9560f3ee263",
           "page_id": null,
-          "is_inserted": false,
           "rotated_angle": 0.0
         }
       }
     ]
   },
   "metadata": {
-    "uid": "doc_uid_auto_fa1f6786-5d3c-11ee-b518-0242ac120002",
+    "uid": "doc_uid_auto_75c93394-5df1-11ee-bfc1-0242ac120002",
     "file_name": "example_return_format.docx",
-    "temporary_file_name": "1695822696_268.docx",
+    "temporary_file_name": "1695900213_314.docx",
     "size": 21270,
-    "modified_time": 1695822696,
-    "created_time": 1695822696,
-    "access_time": 1695822696,
+    "modified_time": 1695900213,
+    "created_time": 1695900213,
+    "access_time": 1695900213,
     "file_type": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
     "document_subject": "",
     "keywords": "",

docs/source/_static/json_format_examples/linear_structure_type.json

@@ -388,7 +388,7 @@
               "start": 0,
               "end": 14,
               "name": "attachment",
-              "value": "attach_fa23fd78-5d3c-11ee-b518-0242ac120002"
+              "value": "attach_75d13b70-5df1-11ee-bfc1-0242ac120002"
             }
           ],
           "metadata": {
@@ -504,20 +504,19 @@
         "metadata": {
           "uid": "3a327789721e09b3fa6fd9560f3ee263",
           "page_id": null,
-          "is_inserted": false,
           "rotated_angle": 0.0
         }
       }
     ]
   },
   "metadata": {
-    "uid": "doc_uid_auto_fa309d08-5d3c-11ee-b518-0242ac120002",
+    "uid": "doc_uid_auto_75e45e94-5df1-11ee-bfc1-0242ac120002",
     "file_name": "example_return_format.docx",
-    "temporary_file_name": "1695822697_827.docx",
+    "temporary_file_name": "1695900214_259.docx",
     "size": 21270,
-    "modified_time": 1695822697,
-    "created_time": 1695822697,
-    "access_time": 1695822697,
+    "modified_time": 1695900213,
+    "created_time": 1695900213,
+    "access_time": 1695900214,
     "file_type": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
     "document_subject": "",
     "keywords": "",

docs/source/_static/json_format_examples/with_attachments.json

@@ -298,7 +298,7 @@
                           "start": 0,
                           "end": 14,
                           "name": "attachment",
-                          "value": "attach_fa355abe-5d3c-11ee-b518-0242ac120002"
+                          "value": "attach_75ea598e-5df1-11ee-bfc1-0242ac120002"
                         }
                       ],
                       "metadata": {
@@ -420,20 +420,19 @@
         "metadata": {
           "uid": "3a327789721e09b3fa6fd9560f3ee263",
           "page_id": null,
-          "is_inserted": false,
           "rotated_angle": 0.0
         }
       }
     ]
   },
   "metadata": {
-    "uid": "doc_uid_auto_fa4285e0-5d3c-11ee-b518-0242ac120002",
+    "uid": "doc_uid_auto_75fac01c-5df1-11ee-bfc1-0242ac120002",
     "file_name": "example_return_format.docx",
-    "temporary_file_name": "1695822697_953.docx",
+    "temporary_file_name": "1695900214_51.docx",
     "size": 21270,
-    "modified_time": 1695822697,
-    "created_time": 1695822697,
-    "access_time": 1695822697,
+    "modified_time": 1695900214,
+    "created_time": 1695900214,
+    "access_time": 1695900214,
     "file_type": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
     "document_subject": "",
     "keywords": "",
@@ -476,13 +475,13 @@
         "tables": []
       },
       "metadata": {
-        "uid": "attach_fa355abe-5d3c-11ee-b518-0242ac120002",
+        "uid": "attach_75ea598e-5df1-11ee-bfc1-0242ac120002",
         "file_name": "image1.png",
-        "temporary_file_name": "1695822697_181.png",
+        "temporary_file_name": "1695900214_864.png",
         "size": 14874,
-        "modified_time": 1695822697,
-        "created_time": 1695822697,
-        "access_time": 1695822697,
+        "modified_time": 1695900214,
+        "created_time": 1695900214,
+        "access_time": 1695900214,
         "file_type": "image/png",
         "other_fields": {}
       },