Fix issue where form fields can be skipped

formfyxer/lit_explorer.py

@@ -149,7 +149,9 @@ def _rewrite_pdf_fields_in_place(
                 "PDF field traversal count did not match parsed field-name count"
             )
 
-        for field_data, old_name, new_name in zip(flattened_fields, field_names, new_names):
+        for field_data, old_name, new_name in zip(
+            flattened_fields, field_names, new_names
+        ):
             cleaned_name = re.sub(r"^\*", "", new_name)
             if old_name == cleaned_name:
                 continue
@@ -159,7 +161,9 @@ def _rewrite_pdf_fields_in_place(
                 continue
 
             # Nested fields keep their parent hierarchy; flat fields accept full dotted names.
-            target_name = cleaned_name.split(".")[-1] if "." in old_name else cleaned_name
+            target_name = (
+                cleaned_name.split(".")[-1] if "." in old_name else cleaned_name
+            )
             target.T = target_name
 
         my_pdf.save(in_file)

formfyxer/pdf_wrangling.py

@@ -468,31 +468,37 @@ def _unnest_pdf_fields(
         parent_name = []
     if hasattr(field, "T"):
         parent_name.append(str(field.T))
-    if hasattr(field, "FT") and hasattr(field, "F"):
-        # PDF fields have bit flags for specific options. The 17th bit (or hex
-        # 10000) on Buttons mark a "push button", w/o a permanent value
-        # (e.g. "Print this PDF") They aren't really fields, just skip them.
-        if hasattr(field, "Ff") and field.FT == "/Btn" and bool(field.Ff & 0x10000):
-            return []
-        return [{"type": field.FT, "var_name": ".".join(parent_name), "all": field}]
-    elif parent_type and parent_flags and hasattr(field, "F"):
-        if parent_flags and parent_type == "/Btn" and bool(parent_flags & 0x10000):
-            return []
-        return [{"type": parent_type, "var_name": ".".join(parent_name), "all": field}]
-    elif hasattr(field, "Kids"):
+    effective_type = str(field.FT) if hasattr(field, "FT") else parent_type
+    effective_flags = int(field.Ff) if hasattr(field, "Ff") else parent_flags
+
+    if hasattr(field, "Kids"):
         return [
             y
             for x in field.Kids
             for y in _unnest_pdf_fields(
                 x,
                 copy(parent_name),
-                str(field.FT) if hasattr(field, "FT") else None,
-                int(field.Ff) if hasattr(field, "Ff") else None,
+                effective_type,
+                effective_flags,
             )
         ]
-    else:
+
+    if not effective_type:
+        return []
+
+    # PDF fields have bit flags for specific options. The 17th bit (or hex
+    # 10000) on Buttons mark a "push button", w/o a permanent value
+    # (e.g. "Print this PDF"). They aren't really fields, just skip them.
+    if effective_type == "/Btn" and bool((effective_flags or 0) & 0x10000):
         return []
 
+    if hasattr(field, "FT") or hasattr(field, "F"):
+        return [
+            {"type": effective_type, "var_name": ".".join(parent_name), "all": field}
+        ]
+
+    return []
+
 
 def has_fields(pdf_file: str) -> bool:
     """

formfyxer/tests/test_lit_explorer_pdf_labeling.py

@@ -99,7 +99,9 @@ def test_parse_form_accepts_model_and_openai_base_url(self):
     )
     @patch("formfyxer.lit_explorer.unlock_pdf_in_place")
     @patch("formfyxer.lit_explorer.pikepdf.open")
-    @patch("formfyxer.lit_explorer.get_openai_api_key_from_sources", return_value="sk-test")
+    @patch(
+        "formfyxer.lit_explorer.get_openai_api_key_from_sources", return_value="sk-test"
+    )
     @patch(
         "formfyxer.lit_explorer.rename_pdf_fields_with_context",
         return_value={
@@ -131,7 +133,9 @@ def test_parse_form_rewrite_uses_rename_helper_for_nested_fields(
         _mock_is_tagged,
         _mock_needs_calculations,
     ):
-        fake_pdf = SimpleNamespace(pages=[object()], docinfo=SimpleNamespace(Title="Form"))
+        fake_pdf = SimpleNamespace(
+            pages=[object()], docinfo=SimpleNamespace(Title="Form")
+        )
         mock_pikepdf_open.return_value = fake_pdf
 
         with patch(
@@ -174,7 +178,9 @@ def test_rewrite_pdf_fields_in_place_preserves_order_and_full_flat_names(
         ]
         fake_pdf = SimpleNamespace(
             Root=SimpleNamespace(
-                AcroForm=SimpleNamespace(Fields=[object(), object(), object(), object()])
+                AcroForm=SimpleNamespace(
+                    Fields=[object(), object(), object(), object()]
+                )
             ),
             save=Mock(),
             close=Mock(),

formfyxer/tests/test_pdf_labeling_rules.py

@@ -117,9 +117,7 @@ def test_copy_pdf_fields_anchor_adjusts_rectangles(self, mock_get_transforms):
             c.showPage()
             c.save()
 
-            mock_get_transforms.return_value = [
-                None
-            ]
+            mock_get_transforms.return_value = [None]
 
             with pikepdf.Pdf.open(str(source_path)) as source_pdf, pikepdf.Pdf.open(
                 str(destination_path), allow_overwriting_input=True
@@ -308,6 +306,50 @@ def test_get_existing_pdf_fields_uses_overlap_avoidance_for_unmapped_fields(self
             base_path.unlink(missing_ok=True)
             patched_path.unlink(missing_ok=True)
 
+    def test_get_existing_pdf_fields_keeps_named_parent_widget_fields(self):
+        with NamedTemporaryFile(suffix=".pdf", delete=False) as base_tmp:
+            base_path = Path(base_tmp.name)
+        with NamedTemporaryFile(suffix=".pdf", delete=False) as patched_tmp:
+            patched_path = Path(patched_tmp.name)
+
+        try:
+            c = canvas.Canvas(str(base_path))
+            c.drawString(72, 720, "Page 1")
+            c.save()
+
+            with pikepdf.Pdf.open(str(base_path), allow_overwriting_input=True) as pdf:
+                page_obj = pdf.pages[0].obj
+                widget = pikepdf.Dictionary(
+                    Type=pikepdf.Name("/Annot"),
+                    Subtype=pikepdf.Name("/Widget"),
+                    Rect=pikepdf.Array([72, 650, 212, 670]),
+                    F=4,
+                    P=page_obj,
+                )
+                widget_ref = pdf.make_indirect(widget)
+                parent = pikepdf.Dictionary(
+                    FT=pikepdf.Name("/Tx"),
+                    T=pikepdf.String("recipient_org"),
+                    Kids=pikepdf.Array([widget_ref]),
+                )
+                parent_ref = pdf.make_indirect(parent)
+                widget.Parent = parent_ref
+                page_obj.Annots = pikepdf.Array([widget_ref])
+                pdf.Root.AcroForm = pikepdf.Dictionary(
+                    Fields=pikepdf.Array([parent_ref])
+                )
+                pdf.save(str(patched_path))
+
+            loaded_fields = get_existing_pdf_fields(str(patched_path))
+            self.assertEqual(len(loaded_fields), 1)
+            self.assertEqual(len(loaded_fields[0]), 1)
+            self.assertEqual(loaded_fields[0][0].name, "recipient_org")
+            self.assertEqual(loaded_fields[0][0].x, 72)
+            self.assertEqual(loaded_fields[0][0].configs["width"], 140.0)
+        finally:
+            base_path.unlink(missing_ok=True)
+            patched_path.unlink(missing_ok=True)
+
 
 if __name__ == "__main__":
     unittest.main()