Fix pdfoxide parser TypeError on None return from extract_text

pdfparser/pdfoxide_parser.py

@@ -69,7 +69,7 @@ def parse_pdf_pdfoxide(path: str) -> Dict[str, Any]:
         # Extract transactions from all pages
         all_text = ""
         for page_num in range(page_count):
-            page_text = doc.extract_text(page_num)  # type: ignore[attr-defined] or ""
+            page_text = doc.extract_text(page_num) or ""  # type: ignore[attr-defined]
             all_text += page_text + "\n"
         transactions = extract_transactions(all_text)
 

requirements.txt

@@ -11,3 +11,4 @@ tabulate>=0.9.0          # For formatted benchmark tables
 
 # Configuration management
 python-dotenv>=1.0.0     # Environment variable management, Python 3.9 compatible
+pdf-oxide>=0.2.2

tests/test_pdfoxide_edge_cases.py

@@ -0,0 +1,42 @@
+import pytest
+from unittest.mock import MagicMock, patch
+from pdfparser.pdfoxide_parser import parse_pdf_pdfoxide
+
+class TestPdfOxideEdgeCases:
+    @patch('pdfparser.pdfoxide_parser.PdfDocument')
+    def test_extract_text_returns_none(self, mock_pdf_document_cls, tmp_path):
+        """
+        Regression test for bug where extract_text returning None caused TypeError.
+        """
+        # Create a dummy file so validation passes
+        pdf_path = tmp_path / "dummy.pdf"
+        pdf_path.touch()
+
+        # Setup mock instance
+        mock_doc = MagicMock()
+        mock_pdf_document_cls.return_value = mock_doc
+
+        # Mock page_count
+        mock_doc.page_count.return_value = 2
+
+        # Mock extract_text to return None for the second page (index 1)
+        # First page (index 0) returns "metadata"
+        # The parser calls extract_text(0) for metadata, and then loops all pages for transactions
+        def side_effect(page_num):
+            if page_num == 0:
+                return "No. Rekening\nAccount No\n:\n123456\nStatement Date\n:\n01/01/2023"
+            return None
+
+        mock_doc.extract_text.side_effect = side_effect
+
+        # Call parser
+        result = parse_pdf_pdfoxide(str(pdf_path))
+
+        # Verify result structure
+        assert result['metadata']['account_no'] == '123456'
+        # Transactions might be empty as we didn't provide transaction text
+        assert isinstance(result['transactions'], list)
+
+        # Verify full_text handling
+        # It should contain the text from page 0 and a newline for page 1
+        assert "123456" in result['full_text']