From 155a4c9918cf052bed7b73165a7a9ed965e69c6e Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Thu, 25 Jun 2020 16:02:30 -0700 Subject: pdfextract_tool fixes from prod usage --- python/pdfextract_tool.py | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'python/pdfextract_tool.py') diff --git a/python/pdfextract_tool.py b/python/pdfextract_tool.py index e05d48d..0d33ec9 100755 --- a/python/pdfextract_tool.py +++ b/python/pdfextract_tool.py @@ -61,8 +61,11 @@ def run_single(args): worker = PdfExtractBlobWorker(sink=args.sink, thumbnail_sink=args.thumbnail_sink) with open(args.pdf_file, 'rb') as pdf_file: pdf_bytes = pdf_file.read() - result = worker.process(pdf_bytes) - print(json.dumps(result.to_pdftext_dict(), sort_keys=True)) + worker.push_record(pdf_bytes) + worker.finish() + if args.thumbnail_sink: + args.thumbnail_sink.finish() + def main(): -- cgit v1.2.3