documents: check for pdf in key

inspirehep · Nov 20, 2023 · c25d5aa · c25d5aa
1 parent 6f7a368
commit c25d5aa
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/inspirehep/modules/workflows/utils/__init__.py b/inspirehep/modules/workflows/utils/__init__.py
@@ -259,7 +259,7 @@ def wrapper(*args, **kwargs):
 @with_debug_logging
 def get_document_url_for_reference_extraction(obj):
     documents = obj.data.get("documents", [])
-    fulltexts = [document for document in documents if document.get("fulltext")]
+    fulltexts = [document for document in documents if (document.get("fulltext") and document.get("key", "").endswith(".pdf"))]
     documents = fulltexts or documents
 
     if not documents: