From 44b91bed5a0fb3dab46fd0641691bb9e6b5b6078 Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Wed, 30 Dec 2020 11:40:40 -0800 Subject: web ingest: terminal URL mismatch as skip, not assert --- python/fatcat_tools/importers/ingest.py | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/python/fatcat_tools/importers/ingest.py b/python/fatcat_tools/importers/ingest.py index 7a98775d..04ff8db6 100644 --- a/python/fatcat_tools/importers/ingest.py +++ b/python/fatcat_tools/importers/ingest.py @@ -437,7 +437,9 @@ class IngestWebResultImporter(IngestFileResultImporter): if 'revisit_cdx' in row: terminal_cdx = row['revisit_cdx'] assert terminal_cdx['surt'] - assert terminal_cdx['url'] == terminal['terminal_url'] + if terminal_cdx['url'] != terminal['terminal_url']: + self.counts['skip-terminal-url-mismatch'] += 1 + return None wc_cdx = [] # primary resource first -- cgit v1.2.3