fixed actor metadata generation as well, because the same actorId might occur multiple times in a sentence, if that actor has multiple functions during the same period.

3 years ago · 8875630235
parent 9419d6dc08
commit 8875630235
1 changed files with 1 additions and 1 deletions
--- a/R/sent_merger.R
+++ b/R/sent_merger.R
@ -141,7 +141,7 @@ sent_merger <- function(df, actors_meta = NULL, actor_groups = NULL, pos_cutoff


    ## Create aggregate measures for individual actors
-    actors <- df[str_starts(ids, 'A_'),
+    actors <- df[str_starts(ids, 'A_')] %>% .[!duplicated(.,by = c('id','ids','sentence_id')),
                 .(actor.sent = sum(sent_binary_weighted)/sum(words),
                   actor.sent_words = sum(sent_words),
                   actor.words = sum(words),