From 4dcbeb2d044041bd8cfb169374474e8f80ebf989 Mon Sep 17 00:00:00 2001
From: Bryan Newbold <bnewbold@archive.org>
Date: Sun, 8 Nov 2020 22:19:46 -0800
Subject: html: handle more traf error cases

---
 python/sandcrawler/html_ingest.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'python')

diff --git a/python/sandcrawler/html_ingest.py b/python/sandcrawler/html_ingest.py
index 50b193c..7594365 100644
--- a/python/sandcrawler/html_ingest.py
+++ b/python/sandcrawler/html_ingest.py
@@ -25,10 +25,10 @@ def html_extract_body_teixml(doc: bytes) -> dict:
             include_comments=False,
             include_formatting=True,
         )
-    except ValueError as ve:
+    except (ValueError, TypeError) as e:
         return dict(
             status="parse-error",
-            error_msg=str(ve)[:1000],
+            error_msg=str(e)[:1000],
         )
     if tei_xml:
         body_txt = teixml_body_text(tei_xml)
-- 
cgit v1.2.3