christian/hateimparlament
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

change spelling of greens fraction, fix parsing issue in comments table

Browse Source
This commit is contained in:
christian
2021-08-10 15:40:21 +02:00
parent c80ec6259e
commit 77f79f0d86
4 changed files with 32 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
R/repair.R
+23 -19
View File
@@ -1,7 +1,7 @@
fractions <- c("AFD" = "AfD",
"AFD&FRAKTIONSLOS" = "AfD&Fraktionslos",
"BÜNDNIS90/" = "BÜNDNIS 90 / DIE GRÜNEN",
"BÜNDNIS90/DIEGRÜNEN" = "BÜNDNIS 90 / DIE GRÜNEN",
"BÜNDNIS90/" = "BÜNDNIS 90/DIE GRÜNEN",
"BÜNDNIS90/DIEGRÜNEN" = "BÜNDNIS 90/DIE GRÜNEN",
"FRAKTIONSLOS" = "Fraktionslos",
"DIELINKE" = "DIE LINKE",
"SPD" = "SPD",
@@ -81,35 +81,39 @@ lookup_speaker <- function(tb, speaker, name_variable) {
mutate(speaker = Vectorize(find_match)(str_replace_all({{name_variable}}, tobereplaced, "")))
}
repair_comments <- function(comments, speaker) {
cat(paste0("Looking up speaker id's for names in comments. This may take a while ...\n",
"Use repair(, repair_commments = FALSE) to skip this.\n"))
# try to find a speaker id for each actual comment
repair_comments <- function(comments, speaker, lookup_speaker=F) {
comments %>%
filter(!is.na(commenter)) %>%
lookup_speaker(speaker, commenter) %>%
left_join(comments, ., by="commenter") %>%
select(-commenter)
filter(!is.na(commenter) | !is.na(content) | !is.na(fraction)) ->
tb
if (lookup_speaker) {
cat(paste0("Looking up speaker id's for names in comments. This may take a while ...\n",
"Use repair(, lookup_speaker = FALSE) to skip this.\n"))
# try to find a speaker id for each actual comment
tb %>%
filter(!is.na(commenter)) %>%
lookup_speaker(speaker, commenter) %>%
left_join(tb, ., by="commenter")
} else tb
}
#' Repair parsed tables
#'
#' @param parse_output tibble
#' @param repair_comments bool
#' @param lookup_speaker bool
#'
#' If repair_comments is TRUE, members of the parliament mentioned in comments are looked up in speaker table.
#' If lookup_speaker is TRUE, members of the parliament mentioned in comments are looked up in speaker table.
#'
#' Possible test: check identical(repair(res), repair(repair(res))) == TRUE
#' Since repaired tables should be a fixpoint of repair.
#' @export
repair <- function(parse_output, repair_comments = FALSE) {
repair <- function(parse_output, lookup_speaker = FALSE) {
is_valid_res(parse_output)
stopifnot("lookup_speaker must be of type logical" = is.logical(lookup_speaker))
list(speaker = repair_speaker(parse_output$speaker),
speeches = repair_speeches(parse_output$speeches),
talks = repair_talks(parse_output$talks),
comments = if(repair_comments) repair_comments(parse_output$comments,
parse_output$speaker)
else parse_output$comments,
applause = parse_output$applause
)
comments = repair_comments(parse_output$comments,
parse_output$speaker,
lookup_speaker),
applause = parse_output$applause)
}