[formrecognizer] Add proper defaults on DocumentTableCell (Azure#21213)

* add proper defaults on document table cell

* default kind field

* set default in _from_generated and improve init

* revert changes to init, fix test comment

* set content default for kind

* update testcase to account for default scenarios

Author: catalinaperalta

sdk/formrecognizer/azure-ai-formrecognizer/azure/ai/formrecognizer/_models.py

@@ -3120,11 +3120,11 @@ def __init__(self, **kwargs):
     @classmethod
     def _from_generated(cls, cell):
         return cls(
-            kind=cell.kind,
+            kind=cell.kind if cell.kind else "content",
             row_index=cell.row_index,
             column_index=cell.column_index,
-            row_span=cell.row_span,
-            column_span=cell.column_span,
+            row_span=cell.row_span if cell.row_span else 1,
+            column_span=cell.column_span if cell.column_span else 1,
             content=cell.content,
             bounding_regions=[
                 BoundingRegion._from_generated(region)
@@ -3184,11 +3184,11 @@ def from_dict(cls, data):
         :rtype: DocumentTableCell
         """
         return cls(
-            kind=data.get("kind", None),
+            kind=data.get("kind", "content"),
             row_index=data.get("row_index", None),
             column_index=data.get("column_index", None),
-            row_span=data.get("row_span", None),
-            column_span=data.get("column_span", None),
+            row_span=data.get("row_span", 1),
+            column_span=data.get("column_span", 1),
             content=data.get("content", None),
             bounding_regions=[BoundingRegion.from_dict(v) for v in data.get("bounding_regions")]  # type: ignore
             if len(data.get("bounding_regions", [])) > 0

sdk/formrecognizer/azure-ai-formrecognizer/tests/test_to_dict_v3.py

@@ -890,6 +890,57 @@ def test_document_table_cell_to_dict(self):
 
         assert d == final
 
+    def test_document_table_cell_to_dict_use_defaults(self):
+        # NOTE: kind, column_span, and row_span are not included on purpose to test that the proper defaults are set.
+        model = _models.DocumentTableCell(
+            row_index=2,
+            column_index=3,
+            content="cell content",
+            bounding_regions=[
+                _models.BoundingRegion(
+                    bounding_box=[_models.Point(1, 2), _models.Point(3, 4)],
+                    page_number=1,
+                ),
+            ],
+            spans=[
+                _models.DocumentSpan(
+                    offset=5,
+                    length=2,
+                ),
+            ],
+        )
+
+        d = model.to_dict()
+
+        final = {
+            "kind": "content",
+            "row_index": 2,
+            "column_index": 3,
+            "row_span": 1,
+            "column_span": 1,
+            "content": "cell content",
+            "bounding_regions": [
+                {
+                    "page_number": 1,
+                    "bounding_box": [
+                        {"x": 1, "y": 2},
+                        {
+                            "x": 3,
+                            "y": 4,
+                        },
+                    ],
+                },
+            ],
+            "spans": [
+                {
+                    "offset": 5,
+                    "length": 2,
+                },
+            ],
+        }
+
+        assert d == final
+
     def test_model_operation_info_to_dict(self):
         model = _models.ModelOperationInfo(
             operation_id="id123",

sdk/formrecognizer/azure-ai-formrecognizer/tests/testcase.py

@@ -698,11 +698,20 @@ def assertDocumentTablesTransformCorrect(self, transformed_tables, raw_tables, *
             self.assertBoundingRegionsTransformCorrect(table.bounding_regions, expected.bounding_regions)
 
     def assertDocumentTableCellTransformCorrect(self, transformed_cell, raw_cell, **kwargs):
-        assert transformed_cell.kind == raw_cell.kind
+        if raw_cell.kind:
+            assert transformed_cell.kind == raw_cell.kind
+        else:
+            assert transformed_cell.kind == "content"
         assert transformed_cell.row_index == raw_cell.row_index
         assert transformed_cell.column_index == raw_cell.column_index
-        assert transformed_cell.row_span == raw_cell.row_span
-        assert transformed_cell.column_span == raw_cell.column_span
+        if raw_cell.row_span:
+            assert transformed_cell.row_span == raw_cell.row_span
+        else:
+            assert transformed_cell.row_span == 1
+        if raw_cell.column_span:
+            assert transformed_cell.column_span == raw_cell.column_span
+        else:
+            assert transformed_cell.column_span == 1
         assert transformed_cell.content == raw_cell.content
 
         for span, expected_span in zip(transformed_cell.spans or [], raw_cell.spans or []):