From a53dac1a9787abc9b25008b96505b450eee010ac Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Tue, 30 Nov 2021 17:04:41 -0800
Subject: in fatcat exports, skip 'UNKNOWN_TITLE'

---
 chocula/database.py | 5 +++++
 1 file changed, 5 insertions(+)

(limited to 'chocula')

diff --git a/chocula/database.py b/chocula/database.py
index f3ae536..db14700 100644
--- a/chocula/database.py
+++ b/chocula/database.py
@@ -696,6 +696,10 @@ class ChoculaDatabase:
                 counts["empty-name"] += 1
                 continue
 
+            if "UNKNOWN_TITLE" in row["name"]:
+                counts["empty-name"] += 1
+                continue
+
             if len(name) <= 2:
                 counts["short-name"] += 1
                 continue
@@ -774,6 +778,7 @@ class ChoculaDatabase:
                     "nla.gov.au",
                     "purl.nla.gov.au",
                     "www.bibliothek.uni-regensburg.de",
+                    "zdb.uni-bielefeld.de",
                 ):
                     # individual books or google searches, not journal/conference homepages
                     # LOC scanned newspapers
-- 
cgit v1.2.3