From a1b44161e206873be30c0640f5fab7a284023ba1 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Tue, 21 Jan 2020 10:59:27 -0800
Subject: persist: work around GROBID timeouts with no status_code

---
 python/sandcrawler/db.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'python/sandcrawler/db.py')

diff --git a/python/sandcrawler/db.py b/python/sandcrawler/db.py
index 3ec325e..e1414ba 100644
--- a/python/sandcrawler/db.py
+++ b/python/sandcrawler/db.py
@@ -161,7 +161,7 @@ class SandcrawlerPostgresClient:
                 r['metadata'] = json.dumps(r['metadata'], sort_keys=True)
         batch = [(d['key'],
                   d.get('grobid_version') or None,
-                  d['status_code'],
+                  d.get('status_code') or None,
                   d['status'],
                   d.get('fatcat_release') or None,
                   d.get('updated') or datetime.datetime.now(),
-- 
cgit v1.2.3