Failing isCentroidedFromFile unit test

[lgatto]

The last unit test in test_centroided.R currently fails with

* checking tests ...
  Running ‘testthat.R’
 ERROR
Running the tests in ‘tests/testthat.R’ failed.
Last 13 lines of output:
    my description
   10 features of interest:
     P20353, P53501  ...  Q9VCK0, Q9VIU7
  A collection of 1 features of interest.
  A collection of 10 features of interest.
  reading dummyiTRAQ.mzid... DONE!
  Iterations of EM: 
  1...2...3...4...5...6...7...8...9...10...11...
  [1] 0.07947339
  ....══ testthat results  ═══════════════════════════════════════════════════════════
  OK: 2010 SKIPPED: 8 FAILED: 1
  1. Failure: isCentroidedFromFile (@test_centroided.R#33) 
  
  Error: testthat unit tests failed
  Execution halted

See also https://master.bioconductor.org/checkResults/3.7/bioc-LATEST/MSnbase/malbec2-checksrc.html

I have issues reproducing this interactively.

@jotsetung - could you have a go, please.

[jorainer]

I can also not reproduce it. I tested with MSnbase 2.5.14 (current master). For version 2.5.13 (checked out from Bioc git) I get:

> baseUrl <- "https://raw.githubusercontent.com/HUPO-PSI/mzTab/master/examples/"
> 
> ## a list of mzTab objects
> mzt <- sapply(file.path(baseUrl, fls), MzTab)
Warning in file(con, "r") :
  cannot open URL 'https://raw.githubusercontent.com/HUPO-PSI/mzTab/master/examples//Cytidine.mzTab': HTTP status was '404 Not Found'
Error in file(con, "r") : 
  cannot open the connection to 'https://raw.githubusercontent.com/HUPO-PSI/mzTab/master/examples//Cytidine.mzTab'
Calls: sapply -> sapply -> lapply -> FUN -> readLines -> file
Execution halted
* checking for unstated dependencies in ‘tests’ ... OK
* checking tests ...
  Running ‘testthat.R’
 ERROR
Running the tests in ‘tests/testthat.R’ failed.
Last 13 lines of output:
   - 131.1382 +/- 0.002 (#BC80BD)
  Object of class "Spectrum1"
   Retention time: 0:0 
   MSn level: 1 
   Total ion count: 684 
   Polarity: 1 
  ══ testthat results  ═══════════════════════════════════════════════════════════
  OK: 1974 SKIPPED: 8 FAILED: 4
  1. Error: fileNames accessor MzTab (@test_fileNames.R#10) 
  2. Error: read MzTab data v 0.9 and 1.0 (@test_MzTab_09.R#8) 
  3. Error: MzTab creation and accessors (@test_MzTab.R#8) 
  4. Error: Conversion to MSnSetList (@test_MzTab.R#34) 
  
  Error: testthat unit tests failed
  Execution halted

but no error in the test_centroided.R

[lgatto]

These errors you observe come from remote files that have unexpectedly changed location - this has been fixed on github. Thank you for testing.

[jorainer]

the only thing I've seen is that lines 27 and 33 produce warnings - eventually there is a limit in the number of allowed warnings in unit tests on the BioC build machines?
we could actually wrap these calls into expect_warning (because in fact we expect that a warning is thrown, since one of the two files does not provide any centroiding information):

expect_warning(cnt <- isCentroidedFromFile(microtofq_on_disk))
expect_identical(names(cnt), featureNames(microtofq_on_disk))
##
## subsetting
k <- sort(sample(length(microtofq_on_disk), 10))
xx <- microtofq_on_disk[k]
expect_warning(expect_identical(isCentroidedFromFile(xx), cnt[k]))

[jorainer]

I think I found the reason for the error:

library("testthat")
library("MSnbase")
f <- c(system.file("microtofq/MM14.mzML", package = "msdata"),
       system.file("microtofq/MM8.mzML", package = "msdata"))
microtofq_on_disk <- readMSData(f, mode = "onDisk")

cnt <- isCentroidedFromFile(microtofq_on_disk)

## For the first file centroiding information can be determined:
head(cnt)
F1.S001 F1.S002 F1.S003 F1.S004 F1.S005 F1.S006 
   TRUE    TRUE    TRUE    TRUE    TRUE    TRUE 

## But for the second not:
tail(cnt)
F2.S193 F2.S194 F2.S195 F2.S196 F2.S197 F2.S198 
     NA      NA      NA      NA      NA      NA

## If k represents indices of spectra from the first and the second file it works:
k <- c(1:5, 300:305)
xx <- microtofq_on_disk[k]
expect_identical(isCentroidedFromFile(xx), cnt[k])

## But if k represents indices of spectra only of the second file:
k <- 300:310
xx <- microtofq_on_disk[k]
expect_identical(isCentroidedFromFile(xx), cnt[k])
Error: isCentroidedFromFile(xx) not identical to cnt[k].
Names: 'is.NA' value mismatch: 0 in current 11 in target
In addition: Warning message:
In .isCentroidedFromFile(f) : NAs introduced by coercion

Summarizing, the solution would be to add set.seed(123) before k <- sort(sample(length(microtofq_on_disk), 10)), or to use:

expect_identical(unname(isCentroidedFromFile(xx)), unname(cnt[k]))

[lgatto]

Thank you!