Improved hybrid test case

.github/workflows/r-conda.yml

@@ -56,6 +56,8 @@ jobs:
         run: wget -P tests/testdata/filtered -i tests/remote-files/filtered.txt
       - name: Fetch features test data
         run: wget -P tests/testdata/features -i tests/remote-files/features.txt
+      - name: Fetch features test data
+        run: wget -P tests/testdata/hybrid -i tests/remote-files/hybrid.txt
       - name: Run devtools::check()
         run: |
           conda activate recetox-aplcms

CHANGELOG.md

@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - added tests with realistic testdata for `extract_features.R` [#42](https://github.com/RECETOX/recetox-aplcms/pull/42), [#54](https://github.com/RECETOX/recetox-aplcms/pull/54)
 - added tests for `feature.align.R` ([#40](https://github.com/RECETOX/recetox-aplcms/pull/40)), and `adjust.time.R` ([#39](https://github.com/RECETOX/recetox-aplcms/pull/40))
 - added CI to repository's GitHub Actions [#45](https://github.com/RECETOX/recetox-aplcms/pull/45),[#49](https://github.com/RECETOX/recetox-aplcms/pull/49)
+- added additional test cases for hybrid [#133](https://github.com/RECETOX/recetox-aplcms/pull/133)
 ### Changed
 - refactored `feature.align.R` [#63](https://github.com/RECETOX/recetox-aplcms/pull/63)[#88](https://github.com/RECETOX/recetox-aplcms/pull/88)[#102](https://github.com/RECETOX/recetox-aplcms/pull/102)
 - refactored `adjust.time.R` [#64](https://github.com/RECETOX/recetox-aplcms/pull/64)[#102](https://github.com/RECETOX/recetox-aplcms/pull/102)

NAMESPACE

@@ -40,6 +40,7 @@ export(find.turn.point)
 export(find_local_maxima)
 export(get_custom_rt_tol)
 export(get_mzrange_bound_indices)
+export(get_num_workers)
 export(get_rt_region_indices)
 export(get_times_to_use)
 export(hybrid)

R/utils.R

@@ -110,4 +110,20 @@ compute_mz_sd <- function(feature_groups) {
     }
   }
   return(mz_sd)
+}
+
+#' @export
+get_num_workers <- function() {
+  # CRAN limits the number of cores available to packages to 2
+  # source https://stackoverflow.com/questions/50571325/r-cran-check-fail-when-using-parallel-functions#50571533
+  chk <- Sys.getenv("_R_CHECK_LIMIT_CORES_", "")
+
+  if (nzchar(chk) && chk == "TRUE") {
+    # use 2 cores in CRAN/Travis/AppVeyor
+    num_workers <- 2L
+  } else {
+    # use all cores in devtools::test()
+    num_workers <- parallel::detectCores()
+  }
+  return(num_workers)
 }

README.md

@@ -22,6 +22,8 @@ wget -P tests/testdata/recovered/recovered-extracted -i tests/remote-files/recov
 wget -P tests/testdata/recovered/recovered-corrected -i tests/remote-files/recovered-corrected.txt
 wget -P tests/testdata/filtered -i tests/remote-files/filtered.txt
 wget -P tests/testdata/features -i tests/remote-files/features.txt
+wget -P tests/testdata/hybrid -i tests/remote-files/hybrid.txt
+
 ```
 
 The `hybrid` and `unsupervised` tests of recetox-aplcms are [reported](https://github.com/RECETOX/recetox-aplcms/issues/24) to be OS specific and may fail depending on the platrform they are run on. To ensure reproducibility during development process you can run the tests in a designated Docker container as follows:

tests/remote-files/hybrid.txt

@@ -0,0 +1,3 @@
+https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/hybrid/RCX_shortened_recovered_feature_sample_table.parquet
+https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/hybrid/known_table.parquet
+https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/hybrid/mbr_recovered_feature_sample_table.parquet

tests/remote-files/input.txt

@@ -1,4 +1,6 @@
 https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/input/mbr_test0.mzml
+https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/input/mbr_test1.mzml
+https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/input/mbr_test2.mzml
 https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/input/RCX_06_shortened.mzML
 https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/input/RCX_07_shortened.mzML
 https://gitlab.ics.muni.cz/umsa/umsa-files/-/raw/master/testdata/recetox-aplcms/input/RCX_08_shortened.mzML

tests/testthat/test-hybrid.R

@@ -1,30 +1,30 @@
-get_num_workers <- function() {
-  # CRAN limits the number of cores available to packages to 2
-  # source https://stackoverflow.com/questions/50571325/r-cran-check-fail-when-using-parallel-functions#50571533
-  chk <- Sys.getenv("_R_CHECK_LIMIT_CORES_", "")
-
-  if (nzchar(chk) && chk == "TRUE") {
-    # use 2 cores in CRAN/Travis/AppVeyor
-    num_workers <- 2L
-  } else {
-    # use all cores in devtools::test()
-    num_workers <- parallel::detectCores()
-  }
-  return(num_workers)
-}
+patrick::with_parameters_test_that("basic hybrid test", {
+  if(ci_skip == TRUE) skip_on_ci()
 
-test_that("basic hybrid test", {
-  test_files <- c('../testdata/mbr_test0.mzml',
-                  '../testdata/mbr_test1.mzml',
-                  '../testdata/mbr_test2.mzml')
-
-  expected <- arrow::read_parquet('../testdata/hybrid_recovered_feature_sample_table.parquet')
-  known_table <- arrow::read_parquet('../testdata/known_table.parquet')
+  testdata <- file.path("..", "testdata")
+
+  test_files <- sapply(files, function(x) {
+    file.path(testdata, "input", x)
+  })
+
+  known_table <- arrow::read_parquet(
+    file.path(testdata, "hybrid", "known_table.parquet")
+  )
 
   actual <- hybrid(test_files, known_table, cluster = get_num_workers())
 
-  actual$recovered_feature_sample_table <- actual$recovered_feature_sample_table |> dplyr::arrange_all()
-  expected <- expected |> dplyr::arrange_all()
+  expected <- arrow::read_parquet(
+    file.path(testdata, "hybrid", paste0(.test_name, "_recovered_feature_sample_table.parquet"))
+  )
 
   expect_equal(actual$recovered_feature_sample_table, expected)
-})
+}, patrick::cases(
+  mbr = list(
+    files = c("mbr_test0.mzml", "mbr_test1.mzml", "mbr_test2.mzml"),
+    ci_skip = TRUE
+  ),
+  RCX_shortened = list(
+    files = c("RCX_06_shortened.mzML", "RCX_07_shortened.mzML", "RCX_08_shortened.mzML"),
+    ci_skip = FALSE
+  )
+))

tests/testthat/test-unsupervised.R

@@ -1,7 +1,7 @@
 test_that("basic unsupervised test", {
-  test_files <- c('../testdata/mbr_test0.mzml',
-                  '../testdata/mbr_test1.mzml',
-                  '../testdata/mbr_test2.mzml')
+  test_files <- c('../testdata/input/mbr_test0.mzml',
+                  '../testdata/input/mbr_test1.mzml',
+                  '../testdata/input/mbr_test2.mzml')
 
   expected <- arrow::read_parquet('../testdata/unsupervised_recovered_feature_sample_table.parquet')