From 501360998dd27484b3731e003bd25137e5710f66 Mon Sep 17 00:00:00 2001 From: Martin Czygan Date: Thu, 19 Nov 2020 11:53:51 +0100 Subject: verify: allow a larger gap --- fuzzycat/verify.py | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/fuzzycat/verify.py b/fuzzycat/verify.py index a644262..7c3cf0d 100644 --- a/fuzzycat/verify.py +++ b/fuzzycat/verify.py @@ -224,7 +224,11 @@ def compare(a, b): if a.get("title", "").lower() == b.get("title", "").lower(): if a_authors and (a_authors == b_authors): - if a_release_year and b_release_year and a_release_year != b_release_year: + # TODO: https://fatcat.wiki/release/utx5r5e6azbvljipznv7ejqzvq, + # https://fatcat.wiki/release/oceozrqtcbc4tloizhddxaj2ti + # preprint and published work may not be published in the same + # year; compromise allow a small gap + if a_release_year and b_release_year and abs(int(a_release_year) - int(b_release_year)) > 1: return (Status.DIFFERENT, Miss.YEAR) return (Status.EXACT, OK.TITLE_AUTHOR_MATCH) @@ -305,6 +309,7 @@ TITLE_FRAGMENT_BLACKLIST = set([ "student government minutes:", "ieee membership application", "nouvelles du corps médical", + "les conventions de genève", ]) # There titles appear too often, so ignore them for now. -- cgit v1.2.3