Properly check for CSV header, add tests

mhpob · Jun 14, 2024 · dc420e6 · dc420e6
1 parent a04434d
commit dc420e6
Show file tree

Hide file tree

Showing 2 changed files with 76 additions and 5 deletions.
diff --git a/R/utilities_make.R b/R/utilities_make.R
@@ -35,7 +35,8 @@ clean_otn_deployment <- function(deployment) {
     }
   } else if (grepl("^csv$", file_ext)) {
     # Check for OTN header
-    if (ncol(read.csv(deployment, nrows = 1)) == 0) {
+    check_head <- read.csv(deployment, nrows = 1, check.names = FALSE)
+    if (ncol(check_head) > length(unique(names(check_head)))) {
       deployment <- read.csv(deployment,
         skip = 3,
         na.strings = c("NA", "")

diff --git a/tests/testthat/test-clean_otn_deployment.R b/tests/testthat/test-clean_otn_deployment.R
@@ -23,6 +23,9 @@ test_that("expected classes", {
   )
 })
 
+
+
+
 test_that("date times are parsed", {
   deployment_xl <- clean_otn_deployment(pbsm$deployment)
 
@@ -36,13 +39,35 @@ test_that("date times are parsed", {
   )
 })
 
+
+
+
+test_that("Works when no header present", {
+  deployment_sheet1 <- file.path(
+    td,
+    "no_header.xlsx"
+  )
+
+  dep <- readxl::read_excel(pbsm$deployment, sheet = 2, skip = 3)
+
+  writexl::write_xlsx(list(readme = data.frame(),
+                           deployment = dep),
+                      deployment_sheet1)
+
+  expect_equal(
+    clean_otn_deployment(deployment_sheet1),
+    clean_otn_deployment(pbsm$deployment)
+  )
+})
+
+
 test_that("guesses sheet", {
   deployment_sheet1 <- file.path(
     td,
     "one_sheet.xlsx"
   )
 
-  readxl::read_excel(pbsm$deployment, sheet = 2) |>
+  readxl::read_excel(pbsm$deployment, sheet = 2, skip = 3) |>
     writexl::write_xlsx(deployment_sheet1)
 
   expect_equal(
@@ -53,13 +78,15 @@ test_that("guesses sheet", {
 
 
 
-test_that("accepts csv", {
+test_that("accepts csv with header", {
   deployment_csv <- file.path(
     td,
     "pbsm-instrument-deployment-short-form-2018.csv"
   )
-  readxl::read_excel(pbsm$deployment, sheet = 2) |>
-    write.csv(deployment_csv, row.names = FALSE)
+  readxl::read_excel(pbsm$deployment, sheet = 2, col_names = FALSE,
+                     .name_repair = "minimal") |>
+    write.table(deployment_csv, row.names = FALSE, col.names = FALSE,
+                sep = ",", na = "")
 
   expect_s3_class(
     deployment_csv <- clean_otn_deployment(deployment_csv),
@@ -94,6 +121,49 @@ test_that("accepts csv", {
 })
 
 
+
+
+test_that("accepts csv without header", {
+  deployment_csv <- file.path(
+    td,
+    "pbsm-instrument-deployment-short-form-2018.csv"
+  )
+  readxl::read_excel(pbsm$deployment, sheet = 2, skip = 3) |>
+    write.csv(deployment_csv, row.names = FALSE)
+
+  expect_s3_class(
+    deployment_csv <- clean_otn_deployment(deployment_csv),
+    "data.frame"
+  )
+  expect_named(
+    deployment_csv,
+    c(
+      "stationname", "receiver", "internal_transmitter", "deploy_date_time",
+      "deploy_lat", "deploy_long", "recover_date_time"
+    )
+  )
+  expect_type(deployment_csv$stationname, "character")
+  expect_type(deployment_csv$receiver, "character")
+  expect_s3_class(deployment_csv$deploy_date_time, c("POSIXct", "POSIXt"),
+                  exact = TRUE
+  )
+  expect_type(deployment_csv$deploy_lat, "double")
+  expect_type(deployment_csv$deploy_long, "double")
+  expect_s3_class(deployment_csv$recover_date_time, c("POSIXct", "POSIXt"),
+                  exact = TRUE
+  )
+
+  expect_equal(
+    attributes(deployment_csv$deploy_date_time)$tzone,
+    "UTC"
+  )
+  expect_equal(
+    attributes(deployment_csv$recover_date_time)$tzone,
+    "UTC"
+  )
+})
+
+
 test_that("errors if not a CSV or XLS(X)", {
   expect_error(
     clean_otn_deployment("abcs.doc"),