@@ -26,16 +26,16 @@ is_nested <- function(a, b) {
2626
2727# Recompute the Mode-1 per-cluster candidates exactly as create_cluster_labels does.
2828cluster_candidates <- function (bundle ) {
29- md <- add_heuristic_keyword_fields(bundle $ metadata , bundle $ stops )
29+ md <- add_heuristic_keyword_fields(backfill_subject_is_heuristic( bundle $ metadata ) , bundle $ stops )
3030 md $ keywords_rank_cleaned <- md $ subject
31- co <- get_cluster_corpus(bundle $ clusters , md , bundle $ stops , bundle $ taxonomy_separator , heuristic_col = HEUR_MIN2 )
31+ co <- get_cluster_corpus(bundle $ clusters , md , bundle $ stops , bundle $ taxonomy_separator , percluster_filter = TRUE )
3232 tdm <- TermDocumentMatrix(co $ corpus , control = list (
3333 tokenize = SplitTokenizer , weighting = function (x ) weightSMART(x , spec = " ntn" ),
3434 bounds = list (local = c(1 , Inf )), tolower = TRUE ))
3535 tt <- apply(tdm , 2 , function (x ) { x2 <- sort(x , TRUE ); x2 [x2 > 0 ] })
3636 empty <- which(apply(tdm , 2 , sum ) == 0 )
3737 if (length(empty )) {
38- fb <- get_cluster_corpus(bundle $ clusters , md , bundle $ stops , bundle $ taxonomy_separator , heuristic_col = HEUR_MIN1 )$ corpus
38+ fb <- get_cluster_corpus(bundle $ clusters , md , bundle $ stops , bundle $ taxonomy_separator , percluster_filter = FALSE )$ corpus
3939 tt [empty ] <- fill_empty_clusters(fb )[empty ]
4040 }
4141
@@ -78,6 +78,10 @@ mine <- function(path) {
7878 }
7979}
8080
81- args <- commandArgs(trailingOnly = TRUE )
82- files <- if (length(args )) args else fixture_files()
83- for (f in files ) mine(f )
81+ # Only mine when run as a script (Rscript test/mine_cases.R ...), not when another
82+ # script sources this file for its helpers (e.g. test/orcid_review_list.R).
83+ if (sys.nframe() == 0 ) {
84+ args <- commandArgs(trailingOnly = TRUE )
85+ files <- if (length(args )) args else fixture_files()
86+ for (f in files ) mine(f )
87+ }
0 commit comments