AMR/data-raw/extractATCs.R

106 lines
3.1 KiB
R
Raw Normal View History

2022-10-30 14:31:45 +01:00
library(rvest)
library(dplyr)
library(tidyr)
library(AMR)
# we need J01, J02 and J04 (J03 does not exist)
url <- "https://www.whocc.no/atc_ddd_index/?code={code}&showdescription=no"
complete_vector <- character(0)
for (Jxx in c("J01", "J02", "J04")) {
site <- gsub("{code}", Jxx, url, fixed = TRUE)
tbl <- site |>
read_html() |>
html_element("#content") |>
html_text() |>
strsplit(Jxx, fixed = TRUE)
out <- paste0(Jxx, tbl[[1]])
out <- out[out %like% paste0(Jxx, "[A-Z0-9]")]
out <- gsub("(.*?)\n.*", "\\1", out)
complete_vector <- c(complete_vector, out)
}
next_codes <- gsub("(.*?) .*", "\\1", complete_vector)
for (Jxxx in next_codes) {
site <- gsub("{code}", Jxxx, url, fixed = TRUE)
tbl <- site |>
read_html() |>
html_element("#content") |>
html_text() |>
strsplit(Jxxx, fixed = TRUE)
out <- paste0(Jxxx, tbl[[1]])
out <- out[out %like% paste0(Jxxx, "[A-Z0-9]")]
out <- gsub("(.*?)\n.*", "\\1", out)
complete_vector <- c(complete_vector, out)
}
next_codes <- gsub("(.*?) .*", "\\1", complete_vector[complete_vector %like% "[A-Z][0-9][0-9][A-Z][A-Z]"])
complete_tbl <- NULL
for (Jxxxxx in next_codes) {
message(Jxxxxx)
site <- gsub("{code}", Jxxxxx, url, fixed = TRUE)
tbl <- site |>
read_html() |>
html_element("table") |>
html_table(header = TRUE) |>
mutate_if(is.character, function(x) ifelse(x == "", NA_character_, x)) |>
fill(`ATC code`, Name, .direction = "down")
if (is.null(complete_tbl)) {
complete_tbl <- tbl
} else {
complete_tbl <- bind_rows(complete_tbl, tbl)
}
}
new_ab <- complete_tbl |>
select(atc = `ATC code`, name = Name, ddd = DDD, ddd_units = U, route = Adm.R) |>
filter(route %in% c("O", "P", NA)) |>
mutate(route = case_when(
route == "O" ~ "oral",
route == "P" ~ "iv",
TRUE ~ NA_character_
)) |>
pivot_wider(names_from = route, values_from = c(ddd, ddd_units), values_fn = first) |>
select(atc, name,
oral_ddd = ddd_oral, oral_units = ddd_units_oral,
iv_ddd = ddd_iv, iv_units = ddd_units_iv
) |>
mutate(name = paste0(substr(toupper(name), 1, 1), substr(name, 2, 999))) |>
mutate(name = gsub(" and ", "/", name)) |>
filter(name %unlike% "^Combinations") |>
arrange(name)
# check these - any new?
new_ab |>
filter(!name %in% antibiotics$name | !atc %in% unlist(antibiotics$atc)) |>
mutate(
name_old = ab_name(atc, language = NULL),
new = !atc %in% unlist(antibiotics$atc)
) |>
View()
atc_ref <- antibiotics |>
select(ab, group, atc_group1, atc_group2) |>
mutate(atc = ab_atc(ab, only_first = TRUE)) |>
filter(!is.na(atc)) |>
mutate(atc = substr(atc, 1, 5)) |>
select(atc, everything(), -ab) |>
distinct() |>
arrange(atc)
atc_ref <- atc_ref |>
arrange(atc, group, atc_group1, atc_group2) |>
distinct(atc, .keep_all = TRUE)
antibiotics |>
bind_rows(new |>
select(-group, -atc_group1, -atc_group2) |>
left_join(atc_ref, by = c("atc_part" = "atc")) %>%
mutate(
atc = as.list(atc),
loinc = as.list(character(nrow(.))),
abbreviations = loinc,
synonyms = loinc
) |>
select(-atc_part)) |>
View()