From 4294c40a378c386b8158f563168a29e65553395c Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Thu, 3 Oct 2019 18:30:32 -0700
Subject: handle GROBID fetch empty blob condition

---
 python/sandcrawler/grobid.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/python/sandcrawler/grobid.py b/python/sandcrawler/grobid.py
index cdcb339..4c4112f 100644
--- a/python/sandcrawler/grobid.py
+++ b/python/sandcrawler/grobid.py
@@ -85,7 +85,8 @@ class GrobidWorker(SandcrawlerWorker):
             blob = resp.body
         else:
             raise ValueError("not a CDX (wayback) or petabox (archive.org) dict; not sure how to proceed")
-        assert blob
+        if not blob:
+            return dict(status="error", error_msg="empty blob", source=record)
         result = self.grobid_client.process_fulltext(blob, consolidate_mode=self.consolidate_mode)
         result['file_meta'] = gen_file_metadata(blob)
         result['source'] = record
-- 
cgit v1.2.3