Handle things like URLs in Package ID when exporting ezEML data package.

webapp/home/utils/file_utils.py

@@ -0,0 +1,30 @@
+import re
+import unicodedata
+
+def sanitize_filename(value):
+    """
+    Sanitize a string to make it safe for use as a filename.
+
+    Parameters:
+    value (str): The original string.
+
+    Returns:
+    str: A sanitized string safe for filenames.
+    """
+    # Normalize the string to remove accents and special characters
+    value = unicodedata.normalize('NFKD', value)
+    value = value.encode('ascii', 'ignore').decode('ascii')
+
+    # Replace invalid characters with an underscore
+    value = re.sub(r'[<>:"/\\|?*]', '_', value)
+
+    # Remove any remaining non-word characters (except spaces and hyphens)
+    value = re.sub(r'[^\w\s-]', '', value)
+
+    # Replace spaces and hyphens with a single underscore
+    value = re.sub(r'[\s-]+', '_', value)
+
+    # Trim leading and trailing underscores
+    value = value.strip('_')
+
+    return value

webapp/home/views.py

@@ -37,6 +37,7 @@
 
 from markupsafe import Markup
 
+from webapp.home.utils.file_utils import sanitize_filename
 import webapp.home.utils.node_utils
 import webapp.mimemail as mimemail
 
@@ -2732,7 +2733,7 @@ def create_ezeml_package_manifest(user_folder, manifest_files):
     package_id = eml_node.attribute_value("packageId")
     if package_id and package_id != current_document:
         # copy the EML file using the package_id as name
-        arcname = f'{package_id}.xml'
+        arcname = f'{sanitize_filename(package_id)}.xml'
         copyfile(f'{user_folder}/{current_document}.xml', f'{user_folder}/{arcname}')
     else:
         arcname = f'{current_document}.xml'