home edit page issue tracker

This page pertains to UD version 2.

Cantonese HK

Back to index

Overview

   
type only spoken
available since 2.1
link https://github.com/UniversalDependencies/UD_Cantonese-HK
genre spoken
contact tswong-c@my.cityu.edu.hk; jsylee@cityu.edu.hk
sentences 1004
tokens 13918

Issue draft: UD_Cantonese-HK

Modality identification

Is spoken part clearly identifiable? N/A

Metadata review

corpus metadata

This is a parallel treebank with UD_Chinese-HK (same sentences, Cantonese vs. Mandarin), linked via parallel_id (e.g. hk/1). Per the README, the 1004 sentences come from 4 distinct sources, documented only in prose in the README, not in the .conllu:

sent_id range Proposed newdoc id Title Fluency
1-410 missing_days Missing days / 小時光 (film) mostly prepared dialogue
411-547 tempo_in_temple Tempo in Temple / 廟眾樂樂 (film) spontaneous interview, many disfluencies/reparandums
548-650 what_day_is_today What day is today / 今日星期幾 (film) mostly prepared, contains dead air
651-1004 legco_president_election_2016 Legislative Council presidential election meeting (2016-10-12) (Hansard, webcast) spontaneous discussion, many disfluencies/reparandums

doc (and paragraphs) metadata

(none found) - no newdoc id (or the _filename this page previously listed; it is not present in the current .conllu, that was stale). Add # newdoc id at sentences 1, 411, 548, and 651, using the proposed ids in the table above (please confirm the slugs - just short mnemonics derived from the titles).

modality metadata

(none found) - not required for a single-modality (only spoken) treebank, but # genre (film/interview/legislative-proceedings) and interaction parameters (degree_of_spontaneity = planned for the three films vs. unplanned for the legislative discussion) could be added per document using the table above.