diff --git a/src/documents/tasks.py b/src/documents/tasks.py index 8f346e36c..5b00db75f 100644 --- a/src/documents/tasks.py +++ b/src/documents/tasks.py @@ -312,6 +312,7 @@ def update_document_content_maybe_archive_file(document_id) -> None: mime_type, produce_archive=produce_archive, ) + parsed_content = parser.get_text() or "" thumbnail = parser.get_thumbnail(document.source_path, mime_type) @@ -329,7 +330,7 @@ def update_document_content_maybe_archive_file(document_id) -> None: ) Document.objects.filter(pk=document.pk).update( archive_checksum=checksum, - content=parser.get_text(), + content=parsed_content, archive_filename=document.archive_filename, ) newDocument = Document.objects.get(pk=document.pk) @@ -354,14 +355,14 @@ def update_document_content_maybe_archive_file(document_id) -> None: ) else: Document.objects.filter(pk=document.pk).update( - content=parser.get_text(), + content=parsed_content, ) if settings.AUDIT_LOG_ENABLED: LogEntry.objects.log_create( instance=oldDocument, changes={ - "content": [oldDocument.content, parser.get_text()], + "content": [oldDocument.content, parsed_content], }, additional_data={ "reason": "Update document content",