diff options
| author | Bryan Newbold <bnewbold@archive.org> | 2020-02-22 19:32:45 -0800 |
|---|---|---|
| committer | Bryan Newbold <bnewbold@archive.org> | 2020-02-22 19:32:45 -0800 |
| commit | 43d9cc3d87654e6e3bd199fbb01972cc47df863e (patch) | |
| tree | ec503da6a4c9f40dfdb40db3a4c58c488abc179c | |
| parent | e608c22854c8796619e8d6cac1264a3e936eb9e9 (diff) | |
| download | sandcrawler-43d9cc3d87654e6e3bd199fbb01972cc47df863e.tar.gz sandcrawler-43d9cc3d87654e6e3bd199fbb01972cc47df863e.zip | |
ingest: more revisit fixes
| -rw-r--r-- | python/sandcrawler/ia.py | 8 |
1 files changed, 4 insertions, 4 deletions
diff --git a/python/sandcrawler/ia.py b/python/sandcrawler/ia.py index 053fbbf..945c136 100644 --- a/python/sandcrawler/ia.py +++ b/python/sandcrawler/ia.py @@ -560,12 +560,12 @@ class WaybackClient: start_url=start_url, hit=True, status="success", - terminal_url=resource.revisit_cdx.url, - terminal_dt=resource.revisit_cdx.datetime, - terminal_status_code=resource.revisit_cdx.status_code, + terminal_url=cdx_row.url, + terminal_dt=cdx_row.datetime, + terminal_status_code=resource.revisit_cdx.status_code, # ? body=resource.body, cdx=cdx_row, - revisit_cdx=revisit_cdx, + revisit_cdx=resource.revisit_cdx, ) if cdx_row.status_code in (200, 226): |