R example to get purity from the GDC HTTP request API

gistfile1.txt

library("tidyverse")
library("httr")

try_fields <- function(case_hash) {
  print(str_c("Now on case with hash:", case_hash, sep=" ") )
  # all fields that can contain a tumor purity value
  fields = c(
    "files.cases.samples.portions.slides.percent_tumor_cells",
    "files.cases.samples.portions.slides.percent_tumor_nuclei",
    "samples.portions.slides.percent_tumor_cells",
    "samples.portions.slides.percent_tumor_nuclei"
  )
  for (field in fields) {
    url = str_c(
      "https://api.gdc.cancer.gov/cases/",
      case_hash,
      # here, "?expand=" is a good alternative to examine all entries at a
      # particular level, e.g. "samples.portions.slides" gives all info for
      # all slides
      "?fields=",
      field,
      # using "$format=JSON" probably allows for working with the results
      # more systematically, as TSV simply flattens all info into separate columns
      "&format=TSV",
      collapse = ""     
    )
    print(str_c("Now trying url:", url, sep=" ") )
    response <- 
      content(
        GET(
          url = url,
          timeout(7)
        )
      )
    if (length(response) == 1) {
      break
    }
  }
  if (length(response) == 1) {
    print(str_c("found purity:", response[[1]], sep=" ") )
    return(response[[1]])
  } else {
    return(NA)
  } 
}

# using the sapply on a vector of entries, allows to use
# this function in a dplyr::mutate() statement
get_purity <- function(case_hashes) {
  return(sapply(case_hashes, try_fields))
}