Working through some elision-related fixes recently, I noticed the corpus uses several different Unicode codepoints for the elision apostrophe:
- U+2019 RIGHT SINGLE QUOTATION MARK
- U+02BC MODIFIER LETTER APOSTROPHE
- U+1FBD GREEK KORONIS
- U+1FBF GREEK PSILI
- U+0027 APOSTROPHE (ASCII)
In First1KGreek the dominant choice is U+2019 (~318k occurrences vs ~70k each for U+02BC/U+1FBF). In Perseus canonical-greekLit it's reversed — U+02BC dominates (~261k vs ~43k for U+2019).
Would there be any interest in standardizing an option across repos? I could help create corpus-wide normalization PR's if a direction is chosen.
Cross-posted to PerseusDL/canonical-greekLit#1879.
Working through some elision-related fixes recently, I noticed the corpus uses several different Unicode codepoints for the elision apostrophe:
In First1KGreek the dominant choice is U+2019 (~318k occurrences vs ~70k each for U+02BC/U+1FBF). In Perseus canonical-greekLit it's reversed — U+02BC dominates (~261k vs ~43k for U+2019).
Would there be any interest in standardizing an option across repos? I could help create corpus-wide normalization PR's if a direction is chosen.
Cross-posted to PerseusDL/canonical-greekLit#1879.