From db5df5c7de7e691c5aab01b3a12bd7308df23b8d Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Tue, 15 Dec 2020 19:08:44 -0800 Subject: persist: don't expect HTML TEI-XML in result object --- python/sandcrawler/persist.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'python') diff --git a/python/sandcrawler/persist.py b/python/sandcrawler/persist.py index f13b1f3..d840da2 100644 --- a/python/sandcrawler/persist.py +++ b/python/sandcrawler/persist.py @@ -168,7 +168,7 @@ class PersistIngestFileResultWorker(SandcrawlerWorker): sha1hex=file_meta["sha1hex"], status=record.get('status'), scope=record.get('scope'), - has_teixml=bool(html_body and html_body['status'] == 'success' and html_body.get('tei_xml')), + has_teixml=bool(html_body and html_body['status'] == 'success'), has_thumbnail=False, # TODO word_count=(html_body and html_body.get('word_count')) or None, biblio=record.get('html_biblio'), -- cgit v1.2.3