From f751f37a1328069b77ea8cf30c5de6ab00aca808 Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Thu, 25 Jun 2020 15:00:44 -0700 Subject: pdfextract: fix pdf_extra key names --- python/sandcrawler/pdfextract.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/python/sandcrawler/pdfextract.py b/python/sandcrawler/pdfextract.py index 301754a..efb6cca 100644 --- a/python/sandcrawler/pdfextract.py +++ b/python/sandcrawler/pdfextract.py @@ -199,8 +199,8 @@ def process_pdf(blob: bytes, thumb_size=(180,300), thumb_type="JPEG") -> PdfExtr meta_xml=pdf.metadata or None, pdf_info=pdf_info, pdf_extra=dict( - height=page0rect.height, - width=page0rect.width, + page0_height=page0rect.height, + page0_width=page0rect.width, page_count=pdf.pages, permanent_id=permanent_id, update_id=update_id, -- cgit v1.2.3