fixed _gen_document_dict() to fallover gracefully if a doc is missing…

… DOI
TheDataGuild · Oct 9, 2023 · d034dde · d034dde
1 parent f8cc555
commit d034dde
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/mind_palace/extract.py b/mind_palace/extract.py
@@ -7,7 +7,9 @@
 def _gen_document_dict(file_path) -> dict[str, TextNode]:
     xml = docs.load_tei_xml(file_path)
     doi = xml.header.doi
-    assert doi is not None
+    if doi is None:
+        print(f"DOI is None for {file_path}. Replacing with title instead.")
+        doi = xml.header.title
 
     try:
         title_node = docs.title(xml, doi)