Merge pull request #4331 from PascalEgn/require-document-pdf

documents: check for pdf in key
inspirehep · Nov 20, 2023 · 7801a15 · 7801a15
2 parents b91d41a + c25d5aa
commit 7801a15
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: