fixed xml when finding bounding boxes (#429)

* replace dagger html entity

* strip extra spaces at beginning of xml document

sciencebeam_gym/tools/image_annotation/find_bounding_boxes_utils.py

@@ -441,6 +441,14 @@ def get_cache(temp_dir: str, memory_cache_size: int):
     ])
 
 
+def parse_and_fix_xml(xml_path: str) -> etree.ElementBase:
+    LOGGER.info('parsing XML file(%r)', os.path.basename(xml_path))
+    xml_data = read_bytes(xml_path)
+    xml_data = xml_data.lstrip()
+    xml_data = xml_data.replace(b'†', b'†')
+    return etree.fromstring(xml_data)
+
+
 def process_single_document(
     pdf_path: str,
     image_paths: Optional[List[str]],
@@ -461,8 +469,7 @@ def process_single_document(
     pdf_images = get_images_from_pdf(pdf_path, pdf_scale_to=pdf_scale_to)
     xml_root: Optional[etree.ElementBase] = None
     if xml_path:
-        LOGGER.info('parsing XML file(%r)', os.path.basename(xml_path))
-        xml_root = etree.fromstring(read_bytes(xml_path))
+        xml_root = parse_and_fix_xml(xml_path)
         image_descriptors = get_graphic_element_descriptors_from_xml_node(
             xml_root,
             parent_dirname=os.path.dirname(xml_path)

tests/tools/image_annotation/find_bounding_boxes_utils_test.py

@@ -31,6 +31,7 @@ from sciencebeam_gym.tools.image_annotation.find_bounding_boxes_utils import (
     GraphicImageNotFoundError,
     format_coords_attribute_value,
     main,
+    parse_and_fix_xml,
     parse_args,
     save_annotated_images
 )
@@ -104,6 +105,29 @@ def save_images_as_pdf(path_or_io: Union[str, Path, IO], images: List[PIL.Image.
     )
 
 
+class TestParseAndFixXml:
+    def test_should_parse_valid_xml(self, tmp_path: Path):
+        xml_file = tmp_path / 'test.xml'
+        xml_file.write_text('<?xml version="1.0" encoding="UTF-8"?><xml>text</xml>')
+        xml_root = parse_and_fix_xml(str(xml_file))
+        assert xml_root.tag == 'xml'
+        assert xml_root.text == 'text'
+
+    def test_should_parse_xml_with_missing_dagger_entity(self, tmp_path: Path):
+        xml_file = tmp_path / 'test.xml'
+        xml_file.write_text('<?xml version="1.0" encoding="UTF-8"?><xml>text&dagger;</xml>')
+        xml_root = parse_and_fix_xml(str(xml_file))
+        assert xml_root.tag == 'xml'
+        assert xml_root.text == 'text\u2020'
+
+    def test_should_parse_xml_with_extra_spaces_in_the_beginning(self, tmp_path: Path):
+        xml_file = tmp_path / 'test.xml'
+        xml_file.write_text(' \n <?xml version="1.0" encoding="UTF-8"?>\n<xml>text</xml>')
+        xml_root = parse_and_fix_xml(str(xml_file))
+        assert xml_root.tag == 'xml'
+        assert xml_root.text == 'text'
+
+
 class TestSaveAnnotatedImages:
     def test_should_not_fail(
         self,