From e59d1b617d4abd5f002d9e59b6bbaebc9ff30993 Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Mon, 23 Dec 2019 17:59:10 -0800 Subject: basic shadow importer --- python/fatcat_import.py | 15 +++++++++++++++ 1 file changed, 15 insertions(+) (limited to 'python/fatcat_import.py') diff --git a/python/fatcat_import.py b/python/fatcat_import.py index ad4de0e2..1f026edc 100755 --- a/python/fatcat_import.py +++ b/python/fatcat_import.py @@ -166,6 +166,11 @@ def run_grobid_metadata(args): bezerk_mode=args.bezerk_mode) LinePusher(fmi, args.tsv_file).run() +def run_shadow_lib(args): + fmi = ShadowLibraryImporter(args.api, + edit_batch_size=100) + JsonLinePusher(fmi, args.json_file).run() + def run_wayback_static(args): api = args.api @@ -473,6 +478,16 @@ def main(): action='store_true', help="don't lookup existing files, just insert (clobbers; only for fast bootstrap)") + sub_shadow_lib = subparsers.add_parser('shadow-lib', + help="create release and file entities based on GROBID PDF metadata extraction") + sub_shadow_lib.set_defaults( + func=run_shadow_lib, + auth_var="FATCAT_API_AUTH_TOKEN", + ) + sub_shadow_lib.add_argument('json_file', + help="JSON file to import from (or stdin)", + default=sys.stdin, type=argparse.FileType('r')) + sub_wayback_static = subparsers.add_parser('wayback-static', help="crude crawl+ingest tool for single-page HTML docs from wayback") sub_wayback_static.set_defaults( -- cgit v1.2.3