enhancement: better leaf element string check in XML parsing (#734)

* Enhance leaf element string check in XML parsing

* fix is_string check

* changelog and version

---------

Co-authored-by: Matt Robinson <[email protected]>
Co-authored-by: Matt Robinson <[email protected]>

Author: MalteHB

CHANGELOG.md

@@ -1,9 +1,9 @@
-## 0.7.9-dev0
+## 0.7.9-dev1
 
 ### Enhancements
 
-* Adds --partition-ocr-languages to unstructured-ingest
-
+* Improvements to string check for leafs in `partition_xml`.
+* Adds --partition-ocr-languages to unstructured-ingest.
 
 ### Features
 

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.7.9-dev0"  # pragma: no cover
+__version__ = "0.7.9-dev1"  # pragma: no cover

unstructured/partition/xml.py

@@ -13,6 +13,10 @@ def is_leaf(elem):
     return not bool(elem)
 
 
+def is_string(elem):
+    return isinstance(elem, str) or (hasattr(elem, "text") and isinstance(elem.text, str))
+
+
 def get_leaf_elements(
     filename: Optional[str] = None,
     file: Optional[Union[IO, SpooledTemporaryFile]] = None,
@@ -33,7 +37,7 @@ def get_leaf_elements(
 
     for elem in root.findall(xml_path):
         for subelem in elem.iter():
-            if is_leaf(subelem):
+            if is_leaf(subelem) and is_string(subelem.text):
                 leaf_elements.append(subelem.text)
 
     return "\n".join(leaf_elements)  # type: ignore