`tidybulk` Remove hard dependency for `DESeq2`

[stemangiola]

If I install tidybulk without DESeq2 it throws an error, even if DESeq2 is just in the suggests, in the DESCRIPTION file.

[chilampoon]

i also got this error in tests before I installed DESeq2

── Failed tests ─────────────────────────────────────
Error (test-bulk_methods.R:658:2): DESeq2 differential trancript abundance - no object
<packageNotFoundError/error/condition>
Error in `loadNamespace(x)`: there is no package called ‘DESeq2’
Backtrace:
    ▆
 1. └─base::loadNamespace(x) at test-bulk_methods.R:658:8
 2.   └─base::withRestarts(stop(cond), retry_loadNamespace = function() NULL)
 3.     └─base (local) withOneRestart(expr, restarts[[1L]])
 4.       └─base (local) doWithOneRestart(return(expr), restart)

Error (test-bulk_methods.R:854:2): test prefix
<packageNotFoundError/error/condition>
Error in `library(DESeq2)`: there is no package called ‘DESeq2’
Backtrace:
    ▆
 1. └─base::library(DESeq2) at test-bulk_methods.R:854:8

[ FAIL 2 | WARN 42 | SKIP 2 | PASS 203 ]

[stemangiola]

Why could it be. Maybe there is some object in /data directory which is carrying the DESeq2 class.

[chilampoon]

that error is just because i haven't installed DESeq2 and Suggests in DESCRIPTION won't install the package automatically during the installation of tidybulk.

usually I'd add a few lines to detect uninstalled CRAN packages in tests/vignette, e.g.

list.pkg <- c("stringr", "DESeq2")
new.pkg <- list.pkg[!(list.pkg %in% installed.packages()[ ,"Package"])]
if (length(new.pkg)) {
  install.packages(new.pkg, repos = "https://cloud.r-project.org/")
}

[stemangiola]

The problem is that it should not care about DESeq2 until you call test_differential_abundance, in the same way, it does not care about edgeR until you call test_differential_abundance.

There must be some trace of DESeq2 somewhere. I think it is in the data/ directory. If some file in data/ directory includes DESeq2 class we should remove it.

By the way, I don't remember this error before, I m curious if for a tidybulk version, 1-month-old, we have the same problem.

[chilampoon]

hmm seems like in tests/testthat/test-bulk_methods.R

test_that("test prefix",{
  library(DESeq2)
  library(stringr)

here directly uses DESeq2.

In get_differential_transcript_abundance_deseq2_SE() in R/functions_SE.R there is a install step already 👍

    # Check if package is installed, otherwise install
    if (find.package("DESeq2", quiet = TRUE) %>% length %>% equals(0)) {
        message("Installing DESeq2 needed for differential transcript abundance analyses")
        if (!requireNamespace("BiocManager", quietly = TRUE))
            install.packages("BiocManager", repos = "https://cloud.r-project.org")
        BiocManager::install("DESeq2", ask = FALSE)
    }

[stemangiola]

I was literally writing the same thing XD

Yes, if you have few minutes would you mind try to remove that

  library(DESeq2)
  library(string)

and see if tidybulk installs without DESeq2?

[chilampoon]

Sure! btw is this script dev/dplyr-master-methods.R still being used? can I delete it? cz it causes some errors in checks

[stemangiola]

The way is to add it to .Rbuildignore

The whole dev directory should already been added to .Rbuildignore. I don't see why it complains.

[chilampoon]

I don't know the examples there are being tested

* checking examples ... ERROR
Running examples in ‘tidybulk-Ex.R’ failed
The error most likely occurred in:

> base::assign(".ptime", proc.time(), pos = "CheckExEnv")
> ### Name: arrange
> ### Title: Order rows using column values
> ### Aliases: arrange arrange.tidybulk
> 
> ### ** Examples
> 
> arrange(mtcars, cyl, disp)
Error in arrange(mtcars, cyl, disp) : could not find function "arrange"
Execution halted

the error is strange as well..

[stemangiola]

Remember that would need to add the zzz.R script and its linked scripts as well. Because now you are not exporting anything anymore. Take example from tidySCE package.

[stemangiola]

OK I think I found the problem. It was a dataset in the data/ directory. I created a PR that fixes this.

[stemangiola]

@chilampoon I believe is done.