From 5bca4dfde78cc1ac30a249a128a87c3d0b1a263e Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Wed, 24 Nov 2021 15:54:45 -0800 Subject: codespell fixes --- notes/scaling_works.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'notes/scaling_works.md') diff --git a/notes/scaling_works.md b/notes/scaling_works.md index 60b4597..e9881e1 100644 --- a/notes/scaling_works.md +++ b/notes/scaling_works.md @@ -411,7 +411,7 @@ with highlighter. to allow query and highlight fields to be separate? or perhaps even a separate highlighter query. query "everything", highlight specific fields. -scoring/boosting large reponses (more than a few hundred thousand hits) seems +scoring/boosting large responses (more than a few hundred thousand hits) seems expensive. this include the trivial '*' query. some fulltext phrase queries seem to always be expensive. look in to phrase @@ -699,7 +699,7 @@ encoding: xml.etree.ElementTree.ParseError: not well-formed (invalid token): line 40, column 1122 Root cause was an issue in GROBID, which seems to have been fixed in more -recent versions of GROBID. Patched to continue, and separately commited patch +recent versions of GROBID. Patched to continue, and separately committed patch to fatcat-scholar code base. Ran several retries, manually. -- cgit v1.2.3