From e7ba648fce4b8359358c6661b6ecb34576efc70d Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Tue, 10 Mar 2020 22:39:03 -0700 Subject: ingest_file: --no-spn2 flag for single command --- python/ingest_file.py | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'python') diff --git a/python/ingest_file.py b/python/ingest_file.py index d4fdcac..f6f694e 100755 --- a/python/ingest_file.py +++ b/python/ingest_file.py @@ -17,7 +17,9 @@ def run_single_ingest(args): ) if args.force_recrawl: request['force_recrawl'] = True - ingester = IngestFileWorker() + ingester = IngestFileWorker( + try_spn2=not args.no_spn2, + ) result = ingester.process(request) print(json.dumps(result, sort_keys=True)) return result @@ -51,6 +53,9 @@ def main(): sub_single.add_argument('--force-recrawl', action='store_true', help="ignore GWB history and use SPNv2 to re-crawl") + sub_single.add_argument('--no-spn2', + action='store_true', + help="don't use live web (SPNv2)") sub_single.add_argument('--type', default="pdf", help="type of ingest (pdf, html, etc)") -- cgit v1.2.3