diff options
Diffstat (limited to 'skate/verify.go')
| -rw-r--r-- | skate/verify.go | 9 |
1 files changed, 5 insertions, 4 deletions
diff --git a/skate/verify.go b/skate/verify.go index 18b2f4e..842bedc 100644 --- a/skate/verify.go +++ b/skate/verify.go @@ -636,10 +636,11 @@ func doiPrefix(s string) string { // unifyDigits replaces all digit groups with a hopefully rare placeholder, // e.g. "<NUM>"; This is for discovering very similar, yet different -// publications, where e.g. titles differ only by a single char representing a -// year. Examples are yearly publications, e.g. "World Health Report 2020", -// where any plain similarity score would yield a high number, yet publications -// are obviously not the same. +// publications, where e.g. titles differ only by a single or few chars +// representing a year. Examples are yearly publications, e.g. "World Health +// Report 2020", "World Health Report 2021", ... where any plain similarity +// score would yield a high number, yet publications are obviously not the +// same. func unifyDigits(s string) string { return PatDigits.ReplaceAllString(s, "<NUM>") } |