From f4cce5a765a9f80f9c5e9c907689c06dc9ebf102 Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Mon, 2 Mar 2020 16:41:02 -0800 Subject: add --force-crawl flag to ingest tool --- python/fatcat_ingest.py | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/python/fatcat_ingest.py b/python/fatcat_ingest.py index 6fda74c5..9ba95015 100755 --- a/python/fatcat_ingest.py +++ b/python/fatcat_ingest.py @@ -90,6 +90,8 @@ def _run_search_dump(args, search): ) if not ingest_request: continue + if args.force_recrawl: + ingest_request['force_recrawl'] = True counts['ingest_request'] += 1 if args.dry_run: continue @@ -206,6 +208,9 @@ def main(): parser.add_argument('--allow-non-oa', action='store_true', help="By default, we limit to OA releases. This removes that filter") + parser.add_argument('--force-recrawl', + action='store_true', + help="Tell ingest worker to skip GWB history lookup and do SPNv2 crawl") subparsers = parser.add_subparsers() sub_container = subparsers.add_parser('container', -- cgit v1.2.3